Open Access
Add to BibSonomy
Abstract
For optical interconnect applications, multi-wavelength comb sources require uniform comb spacings and high reliability at high operating temperature. Here, the high-temperature reliability measurements of a InAs quantum dot colliding pulse mode-locked (QD-CPML) laser with 100 GHz comb spacing are systematically investigated. Laser lifetime measurements are performed for over 1600 hours at 80 °C under constant stress current of 150 mA. The mean time to failure (MTTF) of the laser is approximately 38 years (336,203 hours), extracted from the threshold currents extrapolation method. The optical spectral revolutions are also monitored during the aging process, while the grids of comb laser are remarkably stable. The outstanding reliability and spectrum stability make this 100 GHz QD-CPML a promising candidate as a multi-wavelength laser source for datacom and optical I/O applications.
© 2024 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Quantum dot (QD) lasers [1–3] have been extensively investigated for low-power consumption applications, due to their unique properties, such as low threshold current [4–6], high-temperature stability [4,7] and small linewidth enhancement factor [4,8]. In addition, the dislocation insensitive properties of QD lasers enable the potential capability of integrating them on silicon substrates, which can provide cost-effective manufacturing and high-density integration [9]. As research of QD lasers have significantly advanced over past decades, it is anticipated that these novel light sources will play an increasingly important role in fields of telecommunication, sensing and optical interconnects.
Very recently, colliding pulse mode-locked lasers (CPML) have attracted significant attention as multi-wavelength lasers [9–12], which utilize multiple intracavity saturable absorbers to generate flat-top optical frequency comb [13]. The unique ultra-fast carrier and broadband properties of QDs allows ultra-stable mode locking with large channel spacing (>100 GHz) and short-pulse generation [14,15], thereby providing a great candidate for broadband comb generation. Furthermore, QD-CPMLs are capable of achieving high repetition rates (corresponds to large comb spacings), ranging from tens to hundreds of gigahertz. For optical interconnect applications, multiwavelength lasers with channel spacing of 100 GHz, 200 GHz and 400 GHz are required according to CW-WDM MSA [16], which leads to ultra-short cavity length for conventional two-section mode-locked lasers. For instance, 200 GHz two-section MLL requires cavity length as short as 204 µm, which leads to very low optical gain. Another important feature of multi-wavelength laser source is the bandwidth of flat-top optical frequency combs, which relates to number of comb channels within 3 dB optical bandwidth. Owing to the broadband gain spectrum of QD gain media, QD-CPMLs exhibit ultra-broad 3 dB bandwidth in comparison to conventional quantum well lasers. Most importantly, QD-CPMLs’ insensitivity to temperature [17,18] and reflection [19–21], as well as its direct growth capability on silicon [22–30], make it a promising on-chip light source for various applications.
For optical I/O applications, the longevity of CPML must be thoroughly evaluated, as it directly influences their performance and reliability. It is essential to test their durability to ensure sustained performance over extended periods, including power, threshold current and wavelength stability. The foremost parameter is the number of comb lines within 3 dB below the spectrum’s maximum power, which correlates to the CPML’s available channels. We expect to obtain as many comb lines as possible under specific operating conditions, meanwhile, the number of comb lines will not decrease and the position of the peaks of the comb lines will not drift as the operating time increases. From previous research, the operating stability of a passively mode-locked InAs QD laser with 25.5 GHz spacing over temperature range of 20-120 °C has been demonstrated [31], but the systematical studies evaluating the stability of CPML over accelerated aging time is still absent in the field. In this paper, we present a 4th-order QD-CPML with an estimated lifetime over 38 years at high operating temperature of 80 °C. The QD-CPML’s overall performance as multi-wavelength source, including optical spectral mappings and wavelength stability of individual comb line, indicates great stability over the entire aging process.
2. Design, material growth and fabrication
The material epitaxy of QD-CPMLs is performed by molecular beam epitaxy on GaAs substrate as shown in Fig. 1(a). This QD-CPML contains eight tracks of InAs QD layers with p-doped in the active region of the dots in well structure. The upper and lower cladding layers are 1500 nm thick p-type and n-type Al0.4Ga0.6As, respectively. Details of the growth method can be found in our previous results [26]. The top-view microscope image of CPML is shown in Fig. 1(b). The laser ridge width is defined as 4 µm for single transverse mode. The laser cavity (total length of 1580 µm) divided equally into four gain sections by three saturable absorbers (SAs) with identical length of 50 µm, which each gain section exhibits identical length of 395 µm. The width of the electrical isolation gap between SAs and gain section is approximately 10 µm. The measured isolation resistance is approximately 30 kΩ, affirming adequate electrical isolation between the SAs and gain sections. For reliability measurements, the QD-CPML is mounted onto chip-on-carrier (CoC) with butterfly packaging, as shown in Fig. 1(c). Packaged butterfly includes the laser chip thermistor, thermoelectric cooler (TEC) and lensed fiber. Thorlabs’ CLD1015 laser controller is implemented for lifetime measurements.
Fig. 1. (a) Schematic of the laser structure. (b) Microscope image of 4th-order QD-CPML chip. (c) Butterfly packaged QD-CPML chip mounted on CoC.
3. Measurements and results
To investigate the long-term stability of the CPML, comprehensive aging measurements are carried out. The laser device is subjected to a constant injection current of 150 mA (2×Ith) and maintained at fixed temperature of 80 °C for over 1600 hours. To analyze the overall performance consistency of CPML as the aging time progressed, evaluation methods are conducted as below. The conventional light current voltage (LIV) curve is measured every 48 hours at 80 °C after initial 336 hours of pre-aging. The optical spectral mapping is performed every 200 hours, starting after 400 hours of aging, in order to monitor the spectral evolution. For the spectral mapping measurements, the device is set to three different operating conditions for each testing point: room temperature (23 °C), free-running temperature (55 °C), and the aging temperature (80 °C). The typical optical comb spectrums at three individual temperatures are shown in Fig. 2. To note, the QDCPML exhibit the maximum 16 channels of flat-top comb spectrum at 55 °C as indicated in Fig. 2(b).
Fig. 2. Typical optical spectrums of CPML at three different temperatures of (a) 23 °C (Iinj = 130 mA, VSA = 2.4 V), (b) 55 °C (Iinj = 250 mA, VSA = 4.2 V) and (c) 80 °C (Iinj = 160 mA, VSA = 2.6 V).
3.1 Lifetime analysis
The measurement of the light current voltage characteristics (LIV) is primarily intended to investigate the changes in the threshold current and output optical power of the CPML throughout the aging test. L-I curves are measured at about 48-hour intervals, starting from 336th hour and continuing for a total duration of 1728 hours.
As shown in Fig. 3(a), the L-I curve evolution of the CPML is analyzed over 1700 hours of accelerated aging, including total 27 times of L-I measurements conducted between the 336th hour and the 1728th hour, under consistent temperature of 80 °C. The threshold current of this device is 69.6 mA at the beginning of aging test. After the accelerated aging process, the threshold current increased to 72.74 mA at 1728th hour, representing threshold current change of only 4.5% as shown in Fig. 3(b). The corresponding output power of the laser at injection current of 150 mA is reduced by 7% over the entire period.
Fig. 3. (a) L-I curves evolution during the aging process at 80 °C. (b) Normalized threshold current variation in aging and fitted curves. Coordinates are in linear form. (c) Variation of threshold current with aging time. Axes are in exponential form. MTTF = 336,203 h.
The L-I data collected from Fig. 3(a) can be processed to obtain the variation of the threshold current ($I_{{\rm th}}$) against time, as presented in Fig. 3(b). The threshold current variation can be exponentially fitted as the below [32]:
where $I_{{\rm th}}$(0) is the initial threshold current, which is 69.6 mA. t is aging time in hours. a and m are constants that need to be determined through numerical fitting.Laser failure is usually defined as the increase of the threshold current of the laser by one hundred percent [7,32,33], from the fitting curve shown in Fig. 3(b) we can calculate the mean time to failure (MTTF) of the CPML.
MTTF can be calculated by the following formula [32]:
After substituting the extracted values of “a” and “m” into the above equation, estimated MTTF value of 336,203 hours is obtained, which is more than 38 years. This means that this device will not double its threshold current until 38 years at an injection current of 150 mA and temperature of 80 °C, which potentially leads to significantly longer lifetime at room temperature operating condition.
Figure 3(c) can show more visually the change of threshold current with time. The use of exponential coordinates in this representation allows for trivial understanding of the relationship between the change rate of the threshold current and aging time “t”. The intersection of the fitted line with the upper lateral axis is 38 years, which is consistent with the value of MTTF.
3.2 Optical spectral mappings
Different from conventional distributed feedback (DFB) lasers, the reliability of multi-wavelength laser source focus rather than just optical power and threshold current degradation, but also the optical spectral revolution. Therefore, optical spectral mappings of CPML under different operating temperature during aging process are performed here as shown in Fig. 4. These mappings are created by measuring the mapping data starting from the 400th hour of aging and every 200 hours thereafter until 1600 hours, at three individual operating temperature (23 °C, 55 °C, and 80 °C) for each time interval. The injection current (Iinj) ranges from 70 mA to 250 mA with 10-mA intervals and the reverse biased voltage of saturable absorber (VSA) ranges from 0 V to 5 V with 0.2 V intervals. Optical spectra are measured at each operating condition with calculated 3 dB optical bandwidth. In the mapping diagram, each cell represents single point operating condition at fixed temperature. The colors of these cells represent the number of comb lines within 3 dB optical bandwidth. The dark red color indicates the maximum comb line number of 16, in comparison, the dark blue cells represent un-locked region of CPML.
Fig. 4. (a)-(c) Optical spectral mapping of CPML after 400 hours and (d)-(f) after 1600 hours of accelerated aging process at three different operating temperature of 23 °C, 55 °C and 80 °C, respectively.
Figure 4 shows the spectral evolution mapping conducted at accelerated aging time of 400 hours and 1600 hours, respectively, to investigate the changes in the operational condition of the device. For both interrupted aging point, three different spectral mapping temperature are selected at 23 °C, 55 °C and 80 °C for systematic comparison. As observed from Fig. 4, the spectral mappings remain relatively stable at each operating temperature. The most essential condition, which requires the optical spectra to be highly stable, is the laser free-running temperature at 55 °C (Figs. 4(b) and 4(e)). Moreover, the spectral mapping at 55 °C represents the temperature condition in which the maximum numbers of comb lines within 3 dB can be acquired, while the TEC operates at the lowest power consumption. Both the 400-hour and 1600-hour optical spectral mappings show that the comb operation of largest optical bandwidth is achieved in the lower right corner of the spectral mapping, where both high injection current and high VSA conditions are present. Unstable conditions appear in the upper right corner of the low current high VSA area. The stability is equally high for room temperature and high temperature operation. Shapes of mappings at either 23 °C or 80 °C remains unchanged. There are no significant drifts in the current or VSA direction as the aging test time increases. This demonstrates that the laser’s operation conditions are highly stable at all three temperatures through the aging process. To note, the maximum number of comb lines did not change at 23 °C and 80 °C, which are 11 channels and 12 channels, respectively. At 55 °C, the maximum number of comb lines within 3 dB optical bandwidth decreases by one channel, from 16 channels to 15 channels. Overall, the operation conditions of the maximum comb numbers are maintained relatively consistent at all three different temperatures.
To demonstrate the reliability of the side-mode suppression ratio (SMSR) for varied injection current and VSA condition during the aging experiments, SMSR mappings are plotted and shown in Fig. 5. At the beginning of the aging test, under most of the operating conditions, the CPML exhibit SMSR mostly at 40 dB, which is sufficient for external modulations. Most importantly, such good performance does not deteriorate with increased aging time. Until the end of the aging test after 1600 hours, the SMSR mapping of the device at three temperatures remains similar. Comparing SMSR mappings before and after the aging test at 23 °C (Figs. 5(a) and 5(d)), 55 °C (Figs. 5(b) and 5(e)) and 80 °C (Figs. 5(c) and 5(f)), the mappings remain unchanged for either the shapes of mappings or the highest values of SMSR. This proves the suppressed modes did not creep upward over time, reconfirming the high reliability of this QD-CPML structure.
Fig. 5. (a)-(c) SMSR mapping of CPML after 400 hours and (d)-(f) after 1600 hours of accelerated aging process at three different operating temperature of 23 °C, 55 °C and 80 °C, respectively.
3.3 Wavelength stability
Beside the optical bandwidth stability, the precision of the comb peaks at different aging point under the identical injection current and VSA is also the point of interest in wavelength stability analysis. The wavelength stability plots shown in Fig. 6 offer visualized representation of the optical spectral information. The horizontal axis displays the wavelength data, while the vertical axis indicates the progression of the accelerated aging time. The intensity of the spectrogram is depicted by the warmth of the color. The plots are designed in this manner to facilitate the observation of the position of the comb line peaks at different aging time.
Fig. 6. The plots obtained from the wavelength stability analysis at (a) 23 °C 120 mA 5 V, (b) 55 °C 200 mA 1 V and (c) 80 °C 190 mA 0.8 V after 1600 hours of aging test. (d) The plot of wavelength stability from 400 h to 1600 h at 200 mA of Iinj, 0.6 V of VSA at 80 °C. (e) and (f) are two peak position plots against aging time at identical working conditions to (c) and (d).
Here, we selected two typical operating conditions for wavelength stability analysis at 23 °C, 55 °C, and 80 °C, respectively. As observed from Figs. 6(a), 6(b) and 6(c), the positions of comb peaks are relatively stable from 400 hours to 1600 hours. Under specific operating conditions (80 °C, Iinj = 200 mA, VSA =0.6 V), the wavelength of the comb peaks may drift by integer number of fundamental frequency (n × 25 GHz) due to mode competitions during the aging process, as shown in Fig. 6(d).
To achieve a better visualization of comb peak stability, the peak position plots against aging time are analyzed in Figs. 6(e) and 6(f). These diagrams present the exact wavelength of the peaks of the comb lines over time, with the horizontal coordinate being the time elapsed to perform the aging test and the vertical value being the peak wavelength. By studying the peak position plots, intriguing patterns have been discovered. Under operating current of 190 mA and VSA of 0.8 V at 80 °C, the positions of comb peaks, obtained from the 400th hour to the 1600th hour of the aging test, remained relatively stable and did not exhibit significant deviations in their peak wavelengths. But the 3 dB optical bandwidth may vary slightly as shown in Fig. 6(f). Under operating current of 200 mA and VSA of 0.6 V at 80 °C, there are wavelength drifts in the peak positions at multiple integers of 25 GHz, as previously described due to mode competitions.
Overall speaking, this is still highly encouraging outcomes. The observation that the peak wavelength only shifts in orders of 25 GHz reveals that the CPML is still in strong locking modes. With further development in future works, we are confident to suppress the mode competitions, achieving perfectly stable peak wavelength.
4. Conclusion
In this work, we demonstrated an ultra-reliable multi-wavelength laser source by implementing colliding pulse mode-locking. The calculated extrapolated lifetime is more than 38 years at operating temperature of 80 °C under constant stress current of 150 mA, with no obvious degradation. This is the first reported lifetime measurements with spectral stability analysis for multi-wavelength laser source to the best of our knowledge. The results systematically investigate that the distribution of optical spectral bandwidth, wavelength stability after 1600 hours of accelerated aging time. Under most operating conditions, the peak wavelength position of the comb lines remain extremely stable, with no wavelength drift during the entire aging process. Reliability assessment of mode-locked lasers over long periods of time has been added to the temperature reliability [31] of these lasers in related fields. Such low-power-consumption multi-wavelength lasers with long lifetime and great stability exhibit great potential to be widely implemented in large bandwidth optical communication and optical interconnects.
Funding
National Key Research and Development Program of China (2021YFB2800403); National Natural Science Foundation of China (62334013, 62225407, 62005308, 12274449); Innovation Program for Quantum Science and Technology (2021ZD0302300); Youth Innovation Promotion Association of the Chinese Academy of Sciences (Y2022005).
Acknowledgments
The authors would like to express their appreciation to Mingchen Guo for his assistance with the fabrication process. The authors are also grateful to Mr. Xianbiao Hu and Mr. Yixin Chu for their technical support and valuable and constructive discussions.
Disclosures
The authors declare no conflicts of interest.
Data availability
The data that support the findings of this study are available from the first author upon reasonable request.
References
1. D. Bimberg, M. Grundmann, and N. N. Ledentsov, Quantum Dot Heterostructures (Wiley, 1999).
2. N. N. Ledentsov, M. Grundmann, F. Heinrichsdorff, et al., “Quantum-dot heterostructure lasers,” IEEE J. Select. Topics Quantum Electron. 6(3), 439–451 (2000). [CrossRef]
3. A. Y. Egorov, N. A. Maleev, V. M. Ustinov, et al., Quantum Dot Lasers (Oxford University Press, 2003).
4. Y. Wan, J. Norman, S. Liu, et al., “Quantum dot lasers and amplifiers on silicon: recent advances and future developments,” IEEE Nanotechnology Mag. 15(2), 8–22 (2021). [CrossRef]
5. M. Grundmann and D. Bimberg, “Gain and threshold of quantum dot lasers: theory and comparison to experiments,” Jpn. J. Appl. Phys. 36(6S), 4181 (1997). [CrossRef]
6. L. V. Asryan and R. A. Suris, “Theory of threshold characteristics of semiconductor quantum dot lasers,” Semiconductors 38(1), 1–22 (2004). [CrossRef]
7. C. Shang, E. Hughes, Y. Wan, et al., “High-temperature reliable quantum-dot lasers on Si with misfit and threading dislocation filters,” Optica 8(5), 749–754 (2021). [CrossRef]
8. A. Y. Nevsky, U. Bressel, I. Ernsting, et al., “A narrow-linewidth external cavity quantum dot laser for high-resolution spectroscopy in the near-infrared and yellow spectral ranges,” Appl. Phys. B 92(4), 501–507 (2008). [CrossRef]
9. C. Shang, Y. Wan, J. Selvidge, et al., “Perspectives on advances in quantum dot lasers and integration with Si photonic integrated circuits,” ACS Photonics 8(9), 2555–2566 (2021). [CrossRef]
10. E. U. Rafailov, M. A. Cataluna, and W. Sibbett, “Mode-locked quantum-dot lasers,” Nat. Photonics 1(7), 395–401 (2007). [CrossRef]
11. X. Huang, A. Stintz, H. Li, et al., “Passive mode-locking in 1.3 µm two-section InAs quantum dot lasers,” Appl. Phys. Lett. 78(19), 2825–2827 (2001). [CrossRef]
12. E. A. Viktorov, P. Mandel, A. Vladimirov, et al., “Model for mode locking in quantum dot lasers,” Appl. Phys. Lett. 88(20), 201102 (2006). [CrossRef]
13. Y. K. Chen, M. Wu, T. Tanbun-Ek, et al., “Subpicosecond monolithic colliding-pulse mode-locked multiple quantum well lasers,” Appl. Phys. Lett. 58(12), 1253–1255 (1991). [CrossRef]
14. K. Wundke, S. Pötting, J. Auxier, et al., “Pbs quantum-dot-doped glasses for ultrashort-pulse generation,” Appl. Phys. Lett. 76(1), 10–12 (2000). [CrossRef]
15. A. Vagov, V. M. Axt, and T. Kuhn, “Electron-phonon dynamics in optically excited quantum dots: exact solution for multiple ultrashort laser pulses,” Phys. Rev. B 66(16), 165312 (2002). [CrossRef]
16. M. Sysak, J. Johnson, and D. Lewis, “CW-WDM MSA technical specifications Rev 1.0,” https://cw-wdm.org/?wpdmdl=2092.
17. J. Kwoen, B. Jang, K. Watanabe, et al., “High-temperature continuous-wave operation of directly grown InAs/GaAs quantum dot lasers on on-axis Si (001),” Opt. Express 27(3), 2681–2688 (2019). [CrossRef]
18. J. C. Norman, D. Jung, Z. Zhang, et al., “A review of high-performance quantum dot lasers on silicon,” IEEE J. Quantum Electron. 55(2), 1–11 (2019). [CrossRef]
19. J. Duan, H. Huang, B. Dong, et al., “Dynamic and nonlinear properties of epitaxial quantum dot lasers on silicon for isolator-free integration,” Photon. Res. 7(11), 1222–1228 (2019). [CrossRef]
20. J. J. Chen, Z. H. Wang, W. Q. Wei, et al., “Sole excited-state InAs quantum dot laser on silicon with strong feedback resistance,” Front. Mater. 8, 648049 (2021). [CrossRef]
21. J. J. Chen, W. Q. Wei, J. L. Qin, et al., “Multi-wavelength injection locked semiconductor comb laser,” Photon. Res. 10(8), 1840–1847 (2022). [CrossRef]
22. Z. H. Wang, W. Q. Wei, Q. Feng, et al., “InAs/GaAs quantum dot single-section mode-locked lasers on Si (001) with optical self-injection feedback,” Opt. Express 29(2), 674–683 (2021). [CrossRef]
23. W. Q. Wei, J. Y. Zhang, J. H. Wang, et al., “Phosphorus-free 1.5 µm InAs quantum-dot microdisk lasers on metamorphic InGaAs/SOI platform,” Opt. Lett. 45(7), 2042–2045 (2020). [CrossRef]
24. W. Q. Wei, Q. Feng, J. J. Guo, et al., “InAs/GaAs quantum dot narrow ridge lasers epitaxially grown on SOI substrates for silicon photonic integration,” Opt. Express 28(18), 26555–26563 (2020). [CrossRef]
25. M. Heydari, A. R. Zali, R. E. Gildeh, et al., “Fully integrated, 80 GHz bandwidth, 1.3 µm InAs/InGaAs CW-PW quantum dot passively colliding-pulse mode-locked (CPM) lasers for IR sensing application,” IEEE Sensors J. 22(7), 6528–6535 (2022). [CrossRef]
26. J. Z. Huang, Z. T. Ji, J. J. Chen, et al., “Ultra-broadband flat-top quantum dot comb source,” Photon. Res. 10(5), 1308–1316 (2022). [CrossRef]
27. S. Chen, W. Li, J. Wu, et al., “Electrically pumped continuous-wave III-V quantum dot lasers on silicon,” Nat. Photonics 10(5), 307–311 (2016). [CrossRef]
28. Q. Feng, W. Q. Wei, B. Zhang, et al., “O-band and C/L-band III-V quantum dot lasers monolithically grown on Ge and Si substrate,” Appl. Sci. 9(3), 385 (2019). [CrossRef]
29. W. Q. Wei, A. He, B. Yang, et al., “Monolithic integration of embedded III-V lasers on SOI,” Light Sci. Appl. 12(1), 84 (2023). [CrossRef]
30. J. Z. Huang, W. Q. Wei, B. Yang, et al., “Monolithic quantum dot discrete mode laser on SOI,” ACS Photonics 10(6), 1813–1820 (2023). [CrossRef]
31. S. Pan, J. Huang, Z. Zhou, et al., “Quantum dot mode-locked frequency comb with ultra-stable 25.5 GHz spacing between 20 °C and 120 °C,” Photon. Res. 8(12), 1937–1942 (2020). [CrossRef]
32. S. Srinivasan, N. Julian, J. Peters, et al., “Reliability of hybrid silicon distributed feedback lasers,” IEEE J. Select. Topics Quantum Electron. 19(4), 1501305 (2013). [CrossRef]
33. W. Wei, J. Huang, Z. Ji, et al., “Reliable InAs quantum dot lasers grown on patterned Si (001) substrate with embedded hollow structures assisted thermal stress relaxation,” J. Phys. D: Appl. Phys. 55(40), 405105 (2022). [CrossRef]
