Open Access
Add to BibSonomy
Abstract
We experimentally demonstrated DFB lasers containing an active distributed reflector that has the same waveguide core as the active section. Although without current injection, the distributed reflector will be optically pumped to near transparency by the laser itself, and therefore can provide relatively high reflection to the laser. The laser, fabricated with processing steps similar to standard DFB lasers, has achieved 10-mA threshold current, 0.38-mW/mA slope efficiency, above 55-dB side mode suppression ratio, and 24-GHz modulation bandwidth at 60-mA current injection. 28-Gb/s transmission over 10-km single-mode fibers with a power penalty of less-than 0.5 dB has been demonstrated as well.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
With continuous growth of data traffic on the Ethernet and data center, the demands of large capacity optical communication system rapidly increase. The capacity growth in data traffic has made it necessary to increase data rates of the optical network, thus generates large demands on high speed and cost-effective transceivers. In order to meet the needs, 100-Gbit Ethernet (100 GbE) that consists of a multichannel configuration by 4 × 25 Gb/s in the 1.3-μm wavelength band were standardized, and 400 Gb Ethernet standard have been placed on the agenda as well [1]. The multilane configuration with directly modulated lasers (DMLs) is a proven and cost-efffective solution for large capacity Ethernet [1–4]. Because DMLs have the advantages of lower power consumption, lower cost, less footprint and higher output power compared to electro-absorption modulator integrated lasers. In order to make the multilane configuration with DMLs stable, efficient and cost-effective, the DMLs need to have higher rate, higher reliability, higher yield and lower cost. Therefore, it’s very important to improve the modulation bandwidth of DMLs with simple and reliable fabrication.
In order to meet the requirement on high transmission rate of DMLs, it is essential to make its modulation bandwidth as high as possible. The modulation bandwidth of DMLs can been shown as the following formula.
where is the relaxation oscillation frequency; is the damping coefficient; R and C is the total chip resistance and parasitic capacitance, respectively. The R, C and has less effect on the modulation bandwidth because they are normally small enough. In fact, the 3-dB modulation bandwidth is approximately equal to 1.55 if the effect of C, R and can be neglected [9]. Therefore, it is a standard choice to improve the modulation bandwidth by increasing the relaxation oscillation frequency of the DMLs. The relaxation oscillation frequency can be expressed through the following formula [10].Where is the optical confinement factor; is the material differential gain; and is the number and thickness of the quantum wells; L and W is the length and width of the active section; I and is the injection current and threshold current, respectively. According to the formula (2), there are two methods to increase the relaxation oscillation frequency fr. One is to decrease the volume of the active-section, the other one is to increase the differential gain of the gain medium. Highly compressively strained InGaAlAs quantum wells have been proven to have high differential gain and are thus widely used in high speed DMLs currently. Reducing the length and width of active section is also an effective method to improve the relaxation oscillation frequency. There are many reported work on achieving high modulation bandwidth of the DMLs through decreasing the length and width of the active section [3–10]. To reduce width of the active section effectively, generally buried hetero-structure has to be used, which is challenging when InGaAlAs quantum wells are used [4,8].Generally, it is widely believed that reducing the cavity length L and using InGaAlAs quantum wells is the easiest effective way to increase the relaxation oscillation frequency so as to achieve high modulation bandwidth. However, it is also quite challenging to make DFB lasers with short active sections (< 150 μm) because a short active section also means small equivalent reflection, which results in high threshold gain and finally deteriorates the 3-dB modulation bandwidth. Furthermore, lasers with short cavity lengths (< 150 μm) are difficult to fabricate through conventional cleavage process. Many methods have been demonstrated to obtain short active section lengths with reasonable threshold gain, such as DFB lasers integrated with passive waveguides and DFB lasers integrated with passive DBR mirrors [4–8]. DFB lasers integrated with passive waveguides have one cleaved facet high reflection (HR) coated in order to effectively reduce the threshold gain when a short active section length being used. However, these lasers usually suffer from the single mode yield issue due to the reflection phase at the HR coated facet cannot be accurately controlled. The DFB lasers integrated with passive DBR mirrors can obtain additional feedback from the DBR mirrors with accurately controlled reflection phase, therefore they have 100% single mode yield in theory [4].
These DFB lasers integrated with passive waveguides all need butt-joint regrowth, which can significantly increase the fabrication complexity especially when aluminum-containing quantum wells are used for high differential gain.
In this paper, we experimentally demonstrated a two-section DFB laser with an active distributed reflector (ADR-DFB) for shortening the active section length so as to achieve high modulation bandwidth. The ADR-DFB laser have an unbiased reflection section acting as a distributed reflector, which has the same waveguide core and grating pitch as the active section. The reflection section without current injection can be pumped to near transparency by the emission from the laser itself and can provide additional feedback with well controlled phase to the laser, which improves the output slope efficiency and reduces the threshold current. It’s worth mentioning that the fabrication of the ADR-DFB laser is similar to standard DFB lasers. Thus, the ADR-DFB laser with simple and mature fabrication is a more efficient and cost-effective choice for shortening the active section length and for achieving high direct modulation bandwidth.
The paper is organized as follows: in section 2 the device structure and fabrication process of the ADR-DFB laser is explained; section 3 shows the experimental results of the fabricated laser which verify that the reflection section can provide reflectivity as anticipated; the static characteristics of the ADR-DFB laser such as the threshold current and the slope efficiency, the dynamic response and the transmission characteristic are introduced as well; the conclusion is give in the final section.
2. Device structure and fabrication
The schematic view of the ADR-DFB laser is shown in Fig. 1. The ADR-DFB laser contains two sections: the active section and the reflection section. The reflection section acting as a distributed reflector, is integrated right behind the active section. Two sections contain the same grating period and the phase is continuous across the interface between these two sections. Both sections also have the same wafer structure, such as the same multi-quantum wells (MQWs) and the same upper and lower optical confinement layers. The difference between these two sections is that the reflection section has no current injection when the DFB laser is normally working. Although without electrical pumping, the reflection section will be optically pumped to near transparency, therefore the loss is mainly the waveguide internal loss, not the loss due to absorption. The internal waveguide loss is small compared with the grating coupling coefficient. Thus, a significant reflection can still be generated.
We designed the device structure according to the schematic shown above, as Fig. 2(a) shows. The wafer is based on an n-InP substrate with highly S doping. The MQWs consist of 8 compressively strained quantum wells and 9 tensile strained barriers, and two graded InGaAlAs separate optical confinement layers on both sides of the quantum wells. A p-InP cladding layer and a p+-InGaAs contact layer were grown on top of the grating layer to finish the wafer structure.
The active section has a λ/4 shifted grating with the phase shift positioned closer to the reflection section [11,12]. The reflection section has a uniform grating with the same pitch as the active section. An electrical isolation between the active section and the reflection section was achieved through partially etching off the highly p doped layers. Both the active section and the reflection section are 150 μm long. The surface ridge waveguide of the laser has a width of about 2 μm. The front and rear cleaved facets were both anti-reflection (AR) coated. We formed P-electrode of the active section on benzocyclobutene (BCB) to make its stray capacitance small. Although the reflection section has the electrode fabricated, it is generally under the floating status when the laser is normally working. In order to evaluate the contribution from the reflection section, we designed an experiment in which the reflection section was biased under different status, as section 3.2 shown.
The microscope image of the fabricated device is shown in Fig. 2(b). The grating was patterned by electron beam lithography. Figure 3(a) shows the surface roughness and morphology of grating structure before and after the regrowth process. The grating tooth has a triangular shape finally. The laser used a surface ridge waveguide structure with a vertical and smooth sidewall, as shown in Fig. 3(b). The electrical pad of the laser was isolated from the highly p doped layer through BCB. Thus, the only difference of fabricating the ADR-DFB laser from a standard DFB laser is the additional process for achieving electrical isolation between the two sections of the ADR-DFB laser. The isolation resistance about 18 kΩ was obtained through etching off the highly p doped layers. However, this is a standard process to make electro-absorption modulated lasers for example. Therefore the whole fabrication process for the ADR-DFB laser is simple and reliable.
3. Characteristic of the ADR-DFB laser
3.1 Extraction of parameters from Fabry-Pérot lasers and DFB lasers with uniform gratings
Parameters of the wafer structure can be extracted from the below-threshold amplified spontaneous emission (ASE) spectrum of Fabry-Pérot (FP) lasers fabricated with the DFB lasers, such as the gain parameters, internal loss, etc. These parameters could play very important roles in the DFB laser simulation especially if we want to compare the simulation results with measured results from practically fabricated lasers. Various methods have been developed to extract these parameters such as the Fourier series expansion (FSE) method [13] which has been used in this work. 400 μm long FP lasers were used in the experiment. Gain parameters, internal loss, injection efficiency and linewidth enhancement factors were extracted according to the procedures introduced in the paper [13]. The below-threshold ASE spectrum of the FP laser is shown in the Fig. 4(a). We can obtain the net modal gain and internal loss based on the FSE method working on individual longitudinal modes from the ASE spectrum. The results are shown in the Fig. 4(b). The internal loss of the waveguide structure is about 32 cm−1 which is indicated by an arrow in Fig. 4(b). Based on this internal loss, the injection efficiency can also be estimated from the output slope efficiency, which is about 0.85. Then, the material gain can be calculated based on the following relationship
where Γ is the optical confinement factor and is the internal loss; gn and gm is the net modal gain and the material gain, respectively. A theoretical value of 0.12 is used for Γ. In general, the material gain gm varies with carrier density following a logarithmic functionality, as the following formula.where and is the carrier density and the transparency carrier density, go is the gain coefficient. Below threshold the stimulated emission is weak and can be neglected, therefore the relationship between the injection current and the carrier density can be expressed aswhere is the thickness of the active layer; A is the linear recombination coefficient; B is the spontaneous emission recombination coefficient and C is the Auger recombination coefficient. The value of N as calculated from (5), is shown in the Fig. 4 (c), where s−1, cm3s−1, cm6s−1, = 0.85, L = 150 μm, and W = 2.0 μm have been used in the calculation. According to the formula of (3-4), we can get the relationship between the gain coefficient and the transparency carrier density versus wavelength, as shown in Fig. 4(c).
Fig. 4 (a) The measured below-threshold ASE spectra of the FP laser; (b) the net modal gain spectrum and the internal loss; (c) the extracted gain parameters N0 and g0 varying with wavelength; (d) the ASE spectrum of the DFB laser with uniform grating at threshold.
In order to estimate the grating coupling coefficient, we also fabricated the DFB laser with a uniform grating. For the DFB laser with a uniform grating at threshold current, the contribution from the grating is dominative and the influence from gain or loss can be neglected. The ASE spectrum from the DFB laser at threshold shows the stop band and two modes on the edge of the stop band, the spacing between the two modes is . The grating coupling coefficient is estimated from by using the relation [14,15].
where is the grating coupling coefficient; ng is the group index; is the Bragg wavelength of the grating; and Lg is the length of the grating. Figure 4(d) shows the ASE spectrum from a 550 μm long DFB laser at threshold current. It is seen that the stop band width is about 1.9 nm and the Bragg wavelength is about 1326 nm. So, the coupling coefficient of the grating we fabricated is estimated to be about 108 cm−1 as calculated from the formula (6). The DFB lasers with short active sections (150 μm) generally need to work at high temperatures and at high modulation speeds. To meet this critical requirement, regularly κL takes a relatively large value for example 2 or even higher, just to make the threshold gain lower. We designed the κL value of the laser to be 2.25 and the grating coupling coefficient κ was thus supposed to be about 150 cm−1. The corresponding simulation results can be found in reference [11], where the major improvements include lower threshold current and higher modulation bandwidth. The triangular shape of the grating we have fabricated has resulted in a smaller κ less than what we have designed.3.2 Characteristics of the ADR-DFB laser
The fabricated ADR-DFB laser was characterized. The active section has a series resistance about 17 Ω as seen from Fig. 5(a). It's also worth noting that the thermal effect becomes a critical influence factor for a short-cavity DFB laser, because the series resistance increases and the thermal dissipation becomes less effective. Figure 5(b) shows the measured optical spectrum coupled by a lensed fiber from an ADR-DFB laser at 1315 nm. It’s seen that the side-mode suppression ratio (SMSR) of the ADR-DFB laser is about 60 dB. The spectrum shows distinct characteristics of a standard λ/4-shifted DFB laser.
In order to evaluate the contribution from the reflection section, we designed an experiment in which the reflection section was biased under different status, such as floating, 1 mA current injection and reverse bias at 1 V. For different biases of the reflection section, we measured light-current (LI) curve of the ADR-DFB laser with results shown in Fig. 6. With the reflection section floating as shown by the blue solid line in Fig. 6, the threshold current (Ith) of the laser is about 10 mA and the slope efficiency is up to 0.38 mW/mA. When reversely biased, the slope efficiency dramatically decreased to 0.25 mW/mA as shown by the red dotted line in Fig. 6. When the reflection section is reversely biased, it works like a photodiode and the photon-generated carriers are effectively removed so the loss keeps high. The feedback contribution from the reflection section in this case is low and the laser works like a standard DFB laser with two antireflection coated cleaved facets. Thus, it is estimated that the reflection section has increased the slope efficiency of the ADR-DFB laser by about 34% compared with a standard DFB laser with two antireflection coated cleaved facets.
Fig. 6 Light-current characteristics of the ADR-DFB laser under different bias status of the reflection section; the simulation results of the AR/AR coated, AR/HR coated standard DFB laser and the ADR-DFB laser.
It’s essential to understand the difference between the ADR-DFB laser we fabricated and a standard DFB laser. The standard DFB laser can be divided into two types according to the coating situation of the two cleaved facets: AR/AR or AR/HR coating. The short cavity (150 μm) standard DFB laser is hard to fabricate due to the difficult to cleave out such a short cavity length with high quality cleaved facets. Thus, the comparison is made theoretically. To make the simulation more realistic, parameters extracted from real fabricated lasers from last section were used in the simulation, such as the gain parameters, internal loss, injection efficiency, and the grating coupling coefficient. The power reflectivity of the HR and AR coated facet is assumed to be 0.99 and 0, respectively, in the simulation. For the HR coated facet, it is assumed to be at the right position providing an in-phase reflection (real reflection). The simulation results of the AR/HR coated, AR/AR coated standard DFB laser and the ADR-DFB laser are shown by magenta circle, black triangle and violet square line in Fig. 6, respectively. The experiment and simulation results of the ADR-DFB laser agree with each other very well before the thermal effect starts to dominate. This proves that the experimental parameters we extracted is very close to practical parameters of the wafer. It is also seen that the HR/AR coated standard DFB laser has the lowest threshold current of 8 mA, slightly lower than the ADR-DFB laser (10 mA). The AR/AR coated standard DFB laser has the highest threshold current of 15 mA. Therefore it is seen that the ADR-DFB laser has decreased the threshold current by about 34%, compared with the AR/AR coated standard DFB laser. It is also seen that at high injection current, without thermal saturation the ADR-DFB laser should have the output power close to the AR/HR coated standard DFB laser. When the reflection section is reversely biased, the ADR-DFB laser has the output power close to the AR/AR coated standard DFB laser. From simulations we expect that the reflection from the reflection section increases with the output power of the laser increasing [11]. Comparing the LI curve of the ADR-DFB laser with the HR/AR coated standard DFB laser, we can clearly see this trend. Just at threshold current, the reflection is low so the two curves show the biggest difference. As the output power increases, the ADR-DFB laser catches up, showing the reflection increases and at 50 mA injection current, the output power difference is only 0.5 mW. Therefore, it is proved that the reflection section acting as a distributed reflector can effectively provide additional feedback to reduce the threshold current and increase the output slope efficiency of the ADR-DFB laser.
The Fig. 7 shows the measured optical spectra coupled by a lensed fiber from ten adjacent ADR-DFB lasers. The microscope image of ten adjacent ADR-DFB lasers is shown in Fig. 7 as an insert. The SMSRs of ten adjacent ADR-DFB lasers are all above 55 dB from 1321 nm to 1339 nm. This high SMSR is achieved thanks to the fact that the λ/4-phase shift is placed at proper positions of the active section as seen from the simulations in [11]. This also proves that the reflection generated by the reflection section has a well-controlled phase. Thus, the yield of the ADR-DFB laser exceeds standard DFB lasers with HR coated facets. The maximum wavelength deviation from design values as shown in Fig. 7, is below 0.42 nm. This indicates that the single mode yield and wavelength controlling capability of the ADR-DFB laser has been improved when compared with standard DFB lasers with HR and AR coated facets.
To evaluate high speed performance of the ADR-DFB laser, we soldered the laser onto an AlN carrier with coplanar waveguides and a terminator resistor. The coplanar waveguide was carefully designed to reduce the electrical reflection and transmission loss in a wide frequency range. In order to reduce any parasitic effects, the length of the bonding wire that connects the P-electrode of the laser and the coplanar waveguide needs to be minimized. The termination resistor was used to reduce the electrical reflection. Considering the series resistance of the laser and impedance matching, we chose a resistor of 35 Ω integrated on the AlN carrier.
We measured the small signal electrical to optical (E/O) response of the ADR-DFB laser at different injection currents through a 40-G vector network analyzer, with the results shown in Fig. 8. It is seen that the ADR-DFB laser has a of 24 GHz at 60-mA current injection at 20 °C. There is an interesting phenomenon that the amplitude of the relaxation oscillation peak increases under large current injections. The reason is due to the fact that the detuned-loading effect of the ADR-DFB laser enhances the effective differential gain [17–19]. The Bragg wavelength of the active section is red shifted due to self-heating under large current injections. This makes the mode of the laser located on the long wavelength side of the reflection peak from the reflection section, which creates the so-called detuned-loading effect and helps to increase the effective differential gain. Thus, the dynamic of the reflection from the reflection section enhances the amplitude of the relaxation oscillation peak when the ADR-DFB laser is under high current injections.
To accurately evaluate the relaxation oscillation frequency of the ADR-DFB laser, we also measured relative intensity noise (RIN) [9,16]. The RIN spectra at different current injections are shown in Fig. 9(a), from which the relaxation oscillation frequency can be extracted by fitting to the measured curves. Figure 9(b) shows as a function of (I - Ith)0.5, the slope, i.e. the D factor is about 2.87 GHz/mA0.5. The bending dashed line in Fig. 9(b) is due to thermal saturation which is more severe for a short-cavity laser. The determined by the relaxation oscillation frequency which is 1.55 , is slightly higher than measured, which is reasonable considering that the laser modulation bandwidth is also influenced by the damping factor and parasitic parameters.
Fig. 9 (a) RIN spectra of the ADR-DFB laser; (b) relaxation oscillation frequency fr versus (I - Ith)0.5.
The Fig. 10(a) shows the 28-Gb/s eye diagrams for a back-to-back (BTB) configuration with a non-return-to-zero (NRZ) pseudorandom binary sequence of 231–1. Figure 10(b) is the 28-Gb/s eye diagrams after transmission over 10-km single-mode fibers (SMFs) without amplification. The measured wavelength of the laser was 1320 nm. The center bias current was 45 mA, and the modulation current is about 50 mA peak-to-peak. We obtained a dynamic extinction ratio over 6 dB, and the mean output power coupled into a lensed fiber was about 4 dBm. The signal to noise ratio (SNR) of these eye diagrams is slightly worse than expected due to the influence of reflection coming from the tip of the coupling lensed fiber.
Finally, we also measured the bit-error-rate (BER) performance for 28-Gb/s operation by using the ADR-DFB laser. The laser was under 45 mA current injection and the modulation current was about 50 mA peak-to-peak. Figure 11 shows the BER characteristics for the BTB configuration and after 10-km SMFs transmission. Error-free operation was obtained for 10-km SMFs transmission. From the eye diagrams shown in Fig. 10, eye opening after 10-km SMFs transmission is very clear. These results suggest the possibility of transmission over even longer distances with error-free operations. The power penalty after 10-km SMFs transmission is less than 0.5 dB. These results show that the ADR-DFB laser with simple fabrication is an attractive candidate for high capacity optical communications.
4. Conclusions
We have experimentally demonstrated that the ADR-DFB laser containing an active distributed reflector with the same waveguide core as the active section can shorten the active section length so as to improve the modulation bandwidth. We fabricated the ADR-DFB laser with processing steps similar to a standard DFB laser, and extracted key parameters of the wafer and gratings experimentally. We also designed an experiment to demonstrate the functionality of the reflection section. We found that the reflection section increased the slope efficiency by about 34% and decreased the threshold current by about 34% compared with the AR/AR coated standard DFB laser.
We carefully measured the characteristics of the ADR-DFB laser. A slope efficiency of about 0.38 mW/mA and a threshold current of about 10 mA have been achieved. The SMSRs of ten adjacent ADR-DFB lasers are all above 55 dB from 1321 nm to 1339 nm. The ADR-DFB laser has also achieved the 3-dB modulation bandwidth about 24 GHz at 60 mA current injection (5Ith) at 20 °C. The DFB lasers integrated with passive DBR mirrors have reported that the 3-dB modulation bandwidth about 30 GHz at 80 mA current injection (16Ith) at 25 °C [4]. We performed 28-Gb/s transmission experiments using the ADR-DFB laser with clear eye openings for BTB configuration and after 10-km SMFs transmission. The ADR-DFB laser was under 45 mA current injection and the modulation current was about 50 mA peak-to-peak. We also obtained a mean output power about 4 dBm through lensed fiber coupling, and a dynamic extinction ratio over 6 dB. Finally, we measured the BER characteristics for BTB configuration and 10-km SMFs transmission. We obtained error free operation after 10-km SMFs transmission and a power penalty of less than 0.5 dB.
Therefore, the ADR-DFB laser with simple and reliable fabrication is an attractive candidate of DMLs for shortening the active section length and for achieving high direct modulation bandwidth.
Funding
National High-tech R & D Program of China (2015AA017101).
References
1. IEEE, Standard, 802.3, http://standards.ieee.org/about/get/802/802.3.html.
2. K. Adachi, K. Shinoda, T. Kitatani, T. Fukamachi, Y. Matsuoka, T. Sugawara, and S. Tsuji, “25-Gb/s multichannel 1.3-μm surface-emitting lens-integrated DFB laser arrays,” J. Lightwave Technol. 29(19), 2899–2905 (2011). [CrossRef]
3. S. Kanazawa, W. Kobayashi, Y. Ueda, T. Fujisawa, K. Takahata, T. Ohno, T. Yoshimatsu, H. Ishii, and H. Sanjoh, “30-km error-free transmission of directly modulated DFB laser array transmitter optical sub-assembly for 100-Gbit/s application,” J. Lightwave Technol. 34(15), 3646–3652 (2016). [CrossRef]
4. M. Matsuda, A. Uetake, T. Simoyama, S. Okumura, K. Takabayashi, M. Ekawa, and T. Yamamoto, “1.3-μm wavelength AlGaInAs multiple-quantum-well semi-insulating buried-hetero structure distributed-reflector laser arrays on semi-insulating InP substrate,” IEEE J. Sel. Top. Quantum Electron. 21(6), 1–7 (2015). [CrossRef]
5. K. Nakahara, T. Tsuchiya, T. Kitatani, K. Shinoda, T. Taniguchi, T. Kikawa, M. Aoki, and M. Mukaikubo, “40-Gb/s direct modulation with high extinction ratio operation of 1.3-μm InGaAlAs multi quantum well ridge waveguide distributed feedback lasers,” IEEE Photonics Technol. Lett. 19(19), 1436–1438 (2007). [CrossRef]
6. T. Tadokoro, W. Kobayashi, T. Fujisawa, T. Yamanaka, and F. Kano, “43 Gb/s 1.3 μm DFB laser for 40 km transmission,” J. Lightwave Technol. 30(15), 2520–2524 (2012). [CrossRef]
7. K. Nakahara, Y. Wakayama, T. Kitatani, T. Taniguchi, T. Fukamachi, Y. Sakuma, and S. Tanaka, “56-Gb/s direct modulation in InGaAlAs BH-DFB lasers at 55 °C,” in Optical Fiber Communication Conference, OSA Technical Digest (online) (Optical Society of America, 2014), paper Th3A.1.
8. Y. Matsui, T. Pham, T. Sudo, G. Carey, B. Young, J. Xu, C. Cole, and C. Roxlo, “28-Gbaud PAM4 and 56-Gb/s NRZ performance comparison using 1310-nm Al-BH DFB laser,” J. Lightwave Technol. 34(11), 2677–2683 (2016). [CrossRef]
9. W. Kobayashi, T. Ito, T. Yamanaka, T. Fujisawa, Y. Shibata, T. Kurosaki, M. Kohtoku, T. Tadokoro, and H. Sanjoh, “50-Gb/s direct modulation of a 1.3-μm InGaAlAs-based DFB laser with a ridge waveguide structure,” IEEE J. Sel. Top. Quantum Electron. 19(4), 1500908 (2013). [CrossRef]
10. T. Yamamoto, “High-Speed Directly Modulated Lasers,” in Optical Fiber Communication Conference, OSA Technical Digest (online) (Optical Society of America, 2012), paper OTh3F.5. [CrossRef]
11. G. Zhao, J. Sun, Y. Xi, D. Gao, Q. Lu, and W. Guo, “Design and simulation of two-section DFB lasers with short active-section lengths,” Opt. Express 24(10), 10590–10598 (2016). [CrossRef] [PubMed]
12. G. Liu, G. Zhao, Q. Lu, and W. Guo, “Demonstration of a novel two-section DFB laser,” in CLEO: Applications & Technology (2018), paper JTu2A.22.
13. W. H. Guo, D. C. Byrne, Q. Lu, B. Corbett, and J. F. Donegan, “Fabry–Pérot laser characterization based on the amplified spontaneous emission spectrum and the Fourier series expansion method,” IEEE J. Sel. Top. Quantum Electron. 5(17), 1356–1363 (2011). [CrossRef]
14. M. Razeghi, R. Blondeau, M. Krakowski, J. C. Bouley, M. Papuchon, B. Cremoux, and J. Duchemin, “Low-threshold distributed feedback lasers fabricated on material grown completely by LP-MOCVD,” IEEE J. Quantum Electron. 6(6), 507–511 (1985). [CrossRef]
15. N. Matuschek, F. X. Kartner, and U. Keller, “Exact coupled-mode theories for multilayer interference coatings with arbitrary strong index modulations,” IEEE J. Quantum Electron. 33(3), 295–302 (1997). [CrossRef]
16. R. Olshansky, P. Hill, V. Lanzisera, and W. Powazinik, “Frequency response of 1.3 μm InGaAsP high speed semiconductor laser,” IEEE J. Quantum Electron. 23(9), 1410–1418 (1987). [CrossRef]
17. U. Feiste, “Optimization of modulation bandwidth in DBR lasers with detuned Bragg reflectors,” IEEE J. Quantum Electron. 34(12), 2371–2379 (1998). [CrossRef]
18. M. Chacinski and R. Schatz, “Impact of losses in the Bragg section on the dynamics of detuned loaded DBR lasers,” IEEE J. Quantum Electron. 46(9), 1360–1367 (2010). [CrossRef]
19. Y. Matsui, R. Schatz, T. Pham, W. A. Ling, G. Carey, H. M. Daghighian, D. Adams, T. Sudo, and C. Roxlo, “55 GHz bandwidth distributed reflector laser,” J. Lightwave Technol. 35(3), 397–403 (2017). [CrossRef]
