Open Access
Add to BibSonomy
Abstract
High-performance buried heterostructure (BH) C-band InAs/InP quantum dot (QD) and L-band InGaAsP/InP quantum well (QW) two-section passively mode-locked lasers (MLLs) are investigated. From the irregularity of the longitudinal mode spacing in the comb spectra, we confirm that under stable passive mode locking, both devices have strong group velocity dispersion (GVD) and corresponding GVD-induced pulse width broadening. After compensation with anomalous dispersion fibers (SMF-28), short pulse trains with sub-ps pulse widths are achieved for both devices. This observation demonstrates our ability to generate high peak power sub-ps pulses using QD MLLs and QW MLLs for many photonic applications of optical communications.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Increasing the capacity of optical networks with minimal cost and energy consumption has become a priority for supporting Internet traffic. Wavelength division multiplexing (WDM) technology and comb sources are effective ways for terabit optical networks. Monolithic semiconductor mode-locked lasers (MLLs) with high repetition rates and sub-ps pulse durations are of great interests for many optical communications systems in clock sources [1], frequency comb generation [2], and high-capacity transmission networks [3]. At 1550 nm, MLLs based on quantum wells (QW) are a well-established technique to generate ps or sub-ps optical pulses with high repetition rates [4]. On the other hand, zero-dimensional quantum dot (QD) materials have a three-dimensional limiting effect of carriers due to their special energy level structure, which allows the fabrication of lasers with properties such as realization of low threshold currents, fast carrier dynamics, inhomogeneous broadened spectra, and low amplified spontaneous emission (ASE) [5–8]. The performance of GaAs-based QD MLLs in the 1310 nm band has been shown to be greatly improved [9]. In 1550 nm, single-section QD MLL lasers have been reported to generate pulses of sub-ps width [10–12]. In contrast, two-section mode-locked lasers with a saturable absorber allow for larger operating tolerances, and also provide design flexibility for improved performance. Two-section QD MLLs by MOVPE epitaxial growth have been reported rarely [13–15], and the devices that have been reported have strongly chirped pulses. These works show that, comparing to the reported 1310 nm QD MLLs, the advantages of the 1550 nm two-section QD MLLs have not been fully realized. Due to the strong chirp of the semiconductor MLLs, the group velocity dispersion (GVD) influence on the pulse train is required to analyse [16]. Pulse width compression with several techniques of dispersion shifting has been reported, for example, the pulse width was reduce from 14 ps to 970 fs by a dual grating dispersion compression technique [17]. Pulse width of 770 fs was also obtained by using 1200 m single-mode compensated fiber [18]. In [19], we have reported buried heterostructure (BH) C- and L-band coherent combs and pulsed lasers for >1 Tb/s optical networks. The output of these devices was strongly chirped, leading to elongated pulses, which indicates that pulse train is strongly influenced by the GVD in the laser cavity.
In this letter, we demonstrate MOVPE grown two-section passive QD and QW MLLs with high mode-locking performance and compare their mode-locking properties. Furthermore, we analyze the GVD in these two devices. Through theoretical calculations and experimental analysis, we confirm that the GVD induced chirp is so linear that it can be compensated by an appropriate anomalous dispersion SMF-28 fiber length. Finally, combined with the calculation of the measured optical spectra of the devices, we obtain accurate sub-ps pulses both with the QD and QW MLLs.
2. Devices and experimental setup
The devices used in this experiment were two-section passively MLLs using 7x stacked layers of InAs QDs and 4x InGaAsP QWs laser structure, respectively. Buried heterostructure (BH) laser structure with conventional pn-InP blocking layers were processed to achieve a strong carrier and light confinement. More details on material growth and characterization can be found in [20]. Figure 1(a) shows a schematic diagram of the two-section BH QD and QW lasers with a stripe width of 1.4 µm and 20 µm isolation gap in the top p-type contact layers. Both devices have a total length of 840 µm, and the saturable absorber length is 50 µm, corresponding to a repetition frequencies (RF) of ∼50 GHz, with 10% / 90% facet coating on the front facet and back facet, respectively. The experimental configuration for the passive mode locking measurements is shown in Fig. 1(b). The temperature controller ensures a constant temperature of 20 °C. The optical signal of the laser chip is collimated by a lensed fiber and transmitted to the optical switcher after passing through a 60 dB isolator and polarization controller. The mode-locked performance of the device is characterized by an optical spectrometer (Advantest Q8384) with a resolution of 0.01 nm (∼ 1.3 GHz), a high-frequency u2t IR photoelectric detector, an electrical amplifier, an electrical signal analyzer (Keysight N9030A), SMF-28 fibers with different lengths and a second harmonic generation autocorrelator (SHG, pulseCheck).
Fig. 1. (a) Schematics of the processed two-section BH MLLs (Insert figures on the right are active layers of QD and QW devices, respectively); (b) experimental setup for MLLs characterization.
3. Results and discussion
3.1 Mode locking characteristics
We first determine the mode locking behavior of QD MLLs and QW MLLs. We measured the optical spectra and RF spectra of both devices at Igain = 300 mA and absorber voltage Vab = −0.2 V. As shown in Fig. 2, both devices have excellent spectral characteristics with wide spectral bandwidth and a maximum side-mode suppression ratio close to 60 dB in both case. Figure 3 depicts the beat frequency signal in the RF domain. Stable passive mode locking for the QD and QW devices at 50.169 GHz and 51.405 GHz with the −3dB RF linewidth for the two devices are about 125 kHz and 50 kHz, respectively. The stable RF signal confirms the phase correlation between the modes, so that both devices are under stable passive mode-locking.
Fig. 2. Optical spectrum with Igain = 300 mA and Vab = −0.2 V of the (a) 7x QD device at C-band and (b) 4x QW device at L-band.
Fig. 3. RF spectrum of the two-section (a) QD and (b) QW mode-locked laser at Igain = 300 mA and Vab = −0.2 V.
Figure 2 shows that QD devices have better mode uniformity and RF signal stability than QW devices [19]. The wider spectral bandwidth in the QW MLLs leads to more modes involved in mode locking, however, the non-uniform intensity distribution among the modes results in a narrower RF linewidth but a higher noise level. In QD MLLs, the laser mode extracts electrons and holes only from the QD that resonates with the wavelength of that mode, i.e., the QD with the corresponding dot size. Due to the spatial isolation and the interaction only through the wetting layer properties of the QDs, a large number of carriers will stay around the QDs, allowing ultra-fast gain recovery and thus suppressing gain fluctuations. The fast recovery, ultra-broadband, and inhomogeneous gain of the QDs will produce a multi-wavelength laser with better intensity uniformity.
3.2 Group velocity dispersion analysis in MLLs
The electric field of a pulse in the time domain can be described by
The φ0(ω) is spectral phase, which is the relationship of frequency components in the pulse. The refractive index n(ω) in the laser cavity is frequency dependent due to dispersion. Then the dispersion relation for propagation constant k(ω) and refractive index n(ω) is considered. To be able to directly understand the effect of dispersion on pulse characteristics, the Taylor expansion of φ(ω) was carried out [21]
Recalling the definition of group velocity vg and GVD
The second term in (3) adds a delay to pulse, which does not affect the pulse shape. The third term in (3) is proportional to the GVD and is called group delay dispersion (GDD)
The GDD introduces a frequency-dependent delay for the different frequency components of the pulse, and the pulse acquires a quadratic spectral phase from GDD. Substitute into Eq. (2), the pulsed electric field passing through the dispersive laser cavity is composed of the phase multiplication of the input electric field and the conversion function of the dispersive laser cavity. Therefore, the spectral phase of the pulse in the frequency domain is simply summed. After the Fourier transform, new pulse duration is obtained in the time domain.
3.3 Group velocity dispersion evaluation in MLLs
Equation (5) contain terms proportional to $d{n_g}/d\lambdabar $, which suggests a simple way to get the spectral dispersion of GVD. We use an optical spectrometer to record the optical spectrum so that the longitudinal mode spacing of the laser can be widely measured, but may suffer from an aliasing artefact due to the limited accuracy of the spectrometer. We investigate the spectral dispersion for InAs/InP QD MLLs and InGaAsP/InP QW MLLs under stable passive mode locking. As shown in Fig. 4, we plotted the mode spacing and the group refractive index versus wavelength based on optical spectra measurements in Fig. 2. The group refractive index on wavelength was fitted with a linear equation even under stable passive mode locking. The dispersion properties of group refractive index ng versus $\lambdabar {\; }$ in QD and QW MLLs are calculated as $d{n_g}/d\lambdabar = $−5.5667*10−4 nm−1 and $d{n_g}/d\lambdabar = $−3.105*10−4 nm−1, respectively. Substitute the linear variation value of $d{n_g}/d\lambdabar {\; }$ into the Eq. (5), the GVD can be roughly calculated (due to the limited accuracy of the method) as shown in Fig. 5. Although it is stable mode locking in terms of mode beating frequencies and narrow RF linewidths, there is still a strong positive GVD in the laser cavity. In this case, chirp compensation can be performed by applying a negative chirp using a dispersive medium with anomalous dispersion. Here, we applied dispersion compensation by an SMF-28 single-mode fiber.
Fig. 4. (a) and (b) is frequency spacing versus wavelength, (c) and (d) is derived group index dispersion at 20 °C of the two-section QD and QW laser at Igain = 300 mA and Vab = −0.2 V. Red line is a linear fit.
For chirp compensation and pulse compression, it is important to find the appropriate fiber length and the estimation can be calculated with the help of the following equation
Dfiber is the dispersion coefficient of fiber. Based on Eqs. (5) and (8), we expect optimal fiber lengths for the QD and QW devices of 92.8 m and 51.7 m, respectively. Electric field intensity autocorrelation of QD MLLs and QW MLLs are then measured after traveling through different lengths of the fiber by means of a SHG autocorrelator under the bias conditions of Igain = 370 mA and VSA = −0.2 V, and Igain = 335 mA and VSA = −0.4 V, respectively. We chose these specific conditions because the −3 dB RF linewidth is minimal at these working conditions. As shown in Fig. 6(a) and 6(b), we set the measured SHG autocorrelation pulse traces at different lengths of anomalously dispersive SMF-28 fibers to the same scale for comparison. Figure 6(c) and 6(d) show the sharp pulse traces with FHWM of 689 fs and 626 fs after 71 m and 58 m optimum fiber length for QD and QW devices, respectively. The deviation of the determined optimum fiber lengths compared with the values derived from the group index dispersion can be attributed to some dispersion existent in the fibers used for the measurements and due to the uncertainty of the mode spacing measurements.
Fig. 6. Optical intensity autocorrelation pulse trains with the periodic time of 20 ps of the two-section (a) and (c) QD laser at Igain = 370 mA and Vab = −0.2 V; (b) and (d) QW laser at Igain = 335 mA and Vab = −0.4 V after 71 m and 58 m in two types devices respectively.
To calculate the pulse width from the width of the intensity autocorrelation, we first need to calculate the deconvolution factors. We first use the Fourier transform of the Lorentzian pulse shape fitting, which is proved not in common with the measured optical spectra. According to Ref. [22], a better approach is to calculate the inverse Fourier transform function and the autocorrelation function from the optical spectrum, and calculate the deconvolution factor by the ratio of these two functions. To this end, we calculate the theoretical optical pulse from the measured optical spectra. The instantaneous optical power I(t) in the time domain is proportional to the magnitude of the square of the complex electrical field ${|{\textrm{E}(\textrm{t} )} |^2}$, which can be written as
This notation is similar to that found in [22,23]. And then we determine the full-width-at-half-maximum (FWHM) tpulse,calc of the calculated optical pulse.
Here, we calculate the theoretical intensity autocorrelation Rpp(τ) from the measured complex optical spectrum, which can be written as
The amplitude factor Am of each spectral component of the detected photocurrent is determined by a complex optical spectrum, which can be found more details in [22]. We then determine the FWHM tACF,calc of the calculated Rpp(τ). The ratio of the calculated optical pulse width tpulse,calc and the calculated intensity autocorrelation width tACF,calc is the deconvolution factor
The approximate width of the real optical pulse can be received by multiplying the measured autocorrelation width with the deconvolution factor
With the measured optical spectra, we calculated the deconvolution factors for the QD MLLs and QW MLLs: Deconvolution factor for the QD device ≈ 0.73, and for the QW device ≈ 0.72. Applying these to our measured autocorrelation widths gives optical pulse widths of QD and QW MLL of 503 fs and 451 fs, respectively. This short pulse width proves that the second order term of the propagation constant is small at the fiber end after the fiber compensation, allowing for good pulse generation. We also note that higher order chirp components prevent full pulse compression to the transform limit.
4. Conclusion
In conclusion, we demonstrate ultra-broadband two-section passively mode-locked lasers using MOVPE-grown 7x stacked layers of InAs QDs and 4x InGaAsP QWs structure. The spectra of both devices under same conditions consist of 26 and 31 comb lines (>−3 dBm), which is suitable for applications in optical networks > 1 Tb/s and DWDM networks. Compared with QW MLLs, QD MLLs exhibit better mode power uniformity in the spectra and a wider stable mode-locking range compared to the QW MLLs. We also analyzed the irregularities of the longitudinal mode spacing and pulse broadening caused by the cavity dispersion. The linear pulse chirp of the devices can be compensated by a suitable SMF-28 fiber length, and ultrashort pulses with pulse durations of 503 fs and 451 fs were achieved in both devices, respectively. Our QD and QW comb lasers are high-performance, monolithic and low cost emitters ideally suited for a variety of pulsed source applications.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. J. Renaudier, B. Lavigne, F. Lelarge, M. Jourdran, B. Dagens, O. Legouezigou, P. Gallion, and G. Duan, “Standard-compliant jitter transfer function of all-optical clock recovery at 40 GHz based on a quantum-dot self-pulsating semiconductor laser,” IEEE Photonics Technol. Lett. 18(11), 1249–1251 (2006). [CrossRef]
2. K. W. Holman, D. J. Jones, J. Ye, and E. P. Ippen, “Orthogonal control of the frequency comb dynamics of a mode-locked laser diode,” Opt. Lett. 28(23), 2405–2407 (2003). [CrossRef]
3. E. A. Avrutin, J. H. Marsh, and E. L. Portnoi, “Monolithic and multi-GigaHertz mode-locked semiconductor lasers: Constructions, experiments, models and applications,” in Proc. Inst. Elect. Eng.-Optoelectron., (2000), pp. 251–278.
4. L. Hou, M. Haji, J. Akbar, B. Qiu, and A. C. Bryce, “Low divergence angle and low jitter 40 GHz AlGaInAs/InP 1.55 µm mode-locked lasers,” Opt. Lett. 36(6), 966–968 (2011). [CrossRef]
5. O. B. Shchekin and D. G. Deppe, “1.3 µm InAs quantum dot laser with To=161 K from 0 to 80 °C,” Appl. Phys. Lett. 80(18), 3277–3279 (2002). [CrossRef]
6. R. L. Sellin, C. Ribbat, M. Grundmann, N. N. Ledentsov, and D. Bimberg, “Close-to-ideal device characteristics of high-power InGaAs/GaAs quantum dot lasers,” Appl. Phys. Lett. 78(9), 1207–1209 (2001). [CrossRef]
7. O. Karni, K. J. Kuchar, A. Capua, V. Mikhelashvili, G. Sęk, J. Misiewicz, V. Ivanov, J. P. Reithmaier, and G. Eisenstein, “Carrier dynamics in inhomogeneously broadened InAs/AlGaInAs/InP quantum-dot semiconductor optical amplifiers,” Appl. Phys. Lett. 104(12), 121104 (2014). [CrossRef]
8. J. Gomis-Bresco, S. Dommers-Völkel, O. Schöps, Y. Kaptan, O. Dyatlova, D. Bimberg, and U. Woggon, “Time-resolved amplified spontaneous emission in quantum dots,” Appl. Phys. Lett. 97(25), 251106 (2010). [CrossRef]
9. M. Kuntz, G. Fiol, M. Lämmlin, D. Bimberg, M. G. Thompson, K. T. Tan, C. Marinelli, R. V. Penty, I. H. White, V. M. Ustinov, A. E. Zhukov, Y. M. Shernyakov, and A. R. Kovsh, “35 GHz mode-locking of 1.3µm quantum dot lasers,” Appl. Phys. Lett. 85(5), 843–845 (2004). [CrossRef]
10. Z. G. Lu, J. R. Liu, S. Raymond, P. J. Poole, P. J. Barrios, and D. Poitras, “312-fs pulse generation from a passive C-band InAs/InP quantum dot mode-locked laser,” Opt. Express 16(14), 10835–10840 (2008). [CrossRef]
11. Z. G. Lu, J. R. Liu, P. J. Poole, S. Raymond, P. J. Barrios, D. Poitras, G. Pakulski, P. Grant, and D. Roy-Guay, “An L-band monolithic InAs/InP quantum dot mode-locked laser with femtosecond pulses,” Opt. Express 17(16), 13609–13614 (2009). [CrossRef]
12. F. Gao, S. Luo, H. M. Ji, S. T. Liu, F. Xu, Z. R. Lv, D. Lu, C. Ji, and T. Yang, “Ultrashort Pulse and High Power Mode-Locked Laser With Chirped InAs/InP Quantum Dot Active Layers,” IEEE Photonics Technol. Lett. 28(13), 1481–1484 (2016). [CrossRef]
13. T. Sadeev, D. Arsenijević, D. Franke, J. Kreissl, H. Künzel, and D. Bimberg, “1.55-µm mode-locked quantum-dot lasers with 300 MHz frequency tuning range,” Appl. Phys. Lett. 106(3), 031114 (2015). [CrossRef]
14. M. J. R. Heck, E. A. J. M. Bente, B. Smalbrugge, Y.-S. Oei, M. K. Smit, S. Anantathanasarn, and R. Nötzel, “Observation of Q-switching and mode-locking in two-section InAs/InP (100) quantum dot lasers around 1.55 µm,” Opt. Express 15(25), 16292–16301 (2007). [CrossRef]
15. M. J. R. Heck, A. Renault, E. A. J. M. Bente, Y. S. Oei, M. K. Smit, K. S. E. Eikema, W. Ubachs, S. Anantathanasarn, and R. NÖtzel, “Passively Mode-Locked 4.6 and 10.5 GHz Quantum Dot Laser Diodes Around 1.55 µm With Large Operating Regime,” IEEE J. Sel. Top. Quantum Electron. 15(3), 634–643 (2009). [CrossRef]
16. Z. Jiao, R. Zhang, X. Zhang, J. Liu, and Z. Lu, “Modeling of Single-Section Quantum Dot Mode-Locked Lasers: Impact of Group Velocity Dispersion and Self Phase Modulation,” IEEE J. Quantum Electron. 49(12), 1008–1015 (2013). [CrossRef]
17. J. Kim, M.-T. Choi, and P. J. Delfyett, “Pulse generation and compression via ground and excited states from a grating coupled passively mode-locked quantum dot two-section diode laser,” Appl. Phys. Lett. 89(26), 261106 (2006). [CrossRef]
18. R. Rosales, K. Merghem, C. Calo, G. Bouwmans, I. Krestnikov, A. Martinez, and A. Ramdane, “Optical pulse generation in single section InAs/GaAs quantum dot edge emitting lasers under continuous wave operation,” Appl. Phys. Lett. 101(22), 221113 (2012). [CrossRef]
19. M. Zander, W. Rehbein, M. Moehrle, S. Breuer, D. Franke, M. Schell, K. Kolpatzeck, and J. C. Balzer, “High performance BH InAs/InP QD and InGaAsP/InP QW mode-locked lasers as comb and pulse sources,” in Optical Fiber Communication Conference (OFC) 2020, OSA Technical Digest (Optica Publishing Group, 2020), T3C.4.
20. D. Franke, J. Kreissl, W. Rehbein, F. Wenning, H. Kuenzel, U. W. Pohl, and D. Bimberg, “Effect of the Shape of InAs Nanostructures on the Characteristics of InP-Based Buried Heterostructure Semiconductor Optical Amplifiers,” Appl. Phys. Express 4(1), 014101 (2011). [CrossRef]
21. “Preface to the First Edition,” in Ultrashort Laser Pulse Phenomena (2nd Edition), J.-C. Diels and W. Rudolph, eds. (Academic Press, Burlington, 2006), pp. xvii–xxi.
22. K. Kolpatzeck, X. Liu, K.-H. Tybussek, L. Häring, M. Zander, W. Rehbein, M. Moehrle, A. Czylwik, and J. C. Balzer, “System-theoretical modeling of terahertz time-domain spectroscopy with ultra-high repetition rate mode-locked lasers,” Opt. Express 28(11), 16935–16950 (2020). [CrossRef]
23. R. Rosales, S. G. Murdoch, R. T. Watts, K. Merghem, A. Martinez, F. Lelarge, A. Accard, L. P. Barry, and A. Ramdane, “High performance mode locking characteristics of single section quantum dash lasers,” Opt. Express 20(8), 8649–8657 (2012). [CrossRef]
