Open Access
Add to BibSonomyAbstract
We studied the transient behaviors of current-injection quantum-dot microdisk lasers at room temperature. Unique optical responses were observed, including the suppression of relaxation oscillations and fast turn-on. With the help of rate-equation modeling, the suppressed relaxation oscillations are attributed to the enhanced spontaneous emission factor in microdisk lasers. Short turn-on time, around 1 ns without pre-bias, results from the reduced carrier lifetime caused by the Purcell effect and increased nonradiative recombination rate due to higher surface/volume ratio. With short turn-on time, a large-signal direct modulation experiment at 1 Gbps is demonstrated. Modal transient behavior was also investigated under various temperatures from 100 to 300 K. Both of the transient lasing and steady-state lasing from side modes are suppressed at temperatures higher than 250K. Therefore, the quantum-dot microdisk lasers show the potential of single-mode operation under high-speed modulation at room temperature.
©2012 Optical Society of America
1. Introduction
Microdisk laser cavities with high quality factor Q and small mode volume whispering gallery mode (WGM) resonances can achieve lasing with ultralow threshold currents. Due to the low power consumption, the microdisk lasers are the promising candidate as light sources in integrated photonic circuits. The single-mode operation is potentially possible in such small cavities with diameters only a few microns owing to the large cavity mode separations. Room-temperature lasing of current-injection microdisk lasers with GaInAsP/InP compressive-strained multiple-quantum-well (MQW) as active material has been demonstrated [1,2]. The delta-function-like density of states and narrow gain profile of quantum dots (QDs) can potentially reduce the laser threshold further [3]. GaAs-based QD lasers can cover 1.3 μm and longer wavelength, serving as candidates for fiber communication applications. Recently, we have reported the room-temperature lasing of QD current-injection microdisk lasers [4].
The compact size and in-plane emission of microdisk lasers make it an efficient light source for photonic integrated circuits. The integration of current-injection InP-based microdisk lasers and silicon-on-insulator (SOI) waveguide has also been demonstrated [5]. It provides the possibility of the microdisk laser cavities to be exploited for all optical wavelength conversion [6] and electro-optical modulation [7]. The study of transient behaviors of microdisk lasers is crucial for these miniaturized lasers to be put into practical applications. The temporal responses of microdisk lasers have been measured by ultrafast optical pulse excitations [8,9]. The small-signal modulation for optically-pumped InGaAs/InGaAsP MQW has also been demonstrated [10]. In this work, we study the transient behaviors of current-injection QD microdisk lasers under large-signal direct modulation at room temperature. Relaxation oscillations and turn-on behavior will be investigated. Modal behavior under various temperatures will be also studied in both steady-state and transient regimes.
2. Device fabrication and experimental setup
The QD sample in this work consists of five-stacked In0.5Ga0.5As QDs layers. One of the important features of our microdisk devices is that the diameter of the pedestals which serve as the current injection channel is properly controlled. The etching depth near the microdisk perimeter along the radial direction is about 1 μm. The thick pedestals not only help to suppress the WGMs with higher radial numbers but also benefit the carrier injection efficiency by shortening the carrier diffusion path from the injection channel to the disk edge, where the WGM resonances take place. After the microdisk formation, the devices are planarized by benzocyclobutene (BCB) [11]. The schematic drawing of the facricated device is illustrated in Fig. 1 . For sample details and fabrication description, please refer to Ref [4]. The fabricated microdisk laser diameter is 6.5 μm. The corresponding separation of WGMs with fundamental radial number is 15 nm.
Fig. 1 Schematic drawing of a 6.5-μm-diameter current-injection QD microdisk laser with BCB planarization layer.
We measure the optical response of the QD microdisk lasers by collecting the scattering light from microdisk edges by an objective lens. The microdisk lasers are excited by electrical pulses. The spectra are analyzed by a monochromator with a 600-groove/mm grating and a single photon counting avalanche diode (APD). The APD can effectively record the weak signals scattered from the high-Q laser cavity. After the APD signals analyzed by a time-resolved photon detection system, we can also extract the transient behaviors of the microdisk devices. The time resolution of the measurement setup is better than 200 ps. In temperature-dependent experiments, the device temperature is controlled in a cryostat, cooled by liquid nitrogen.
3. Transient behaviors at room temperature
First, we measure the device at room temperature. The microdisk is injected by electric pulses to avoid the sample heating. The L-I curve of a 6.5-μm-diameter microdisk laser is shown in Fig. 2 with a threshold current of 0.47 mA. This microdisk laser is of the same size as that used previously for static measurements [4], but not identical. Figure 2 also shows the WGM spectrum measured above lasing threshold. Only single lasing WGM is observed in the spectrum. The lasing WGM wavelength is centered around 1059 nm. The wavelength corresponds to the TE(1,53) mode, the TE-polarized WGM with fundamental radial number and the azimuthal number equal to 53. The WGM linewidth corresponds to a quality factor of 2300 which is limited by the measurement setup, and the actual quality factor should be higher.
Fig. 2 The L-I curve of the 6.5-μm-diameter microdisk laser with a threshold current of 0.47 mA. Inset: the spectrum showing single lasing WGM near 1059 nm.
The temporal optical responses of the microdisk laser are shown in Fig. 3(a) . The turn-on times of this device are short, around or less than 1 ns without pre-bias. Note that the relaxation oscillation peaks were not observed in these temporal responses, similar to the experimental results from optically-pumped QW microdisk lasers [12]. As a comparison, the step responses of a 2 mm long and 12 μm wide conventional QD edge-emitting laser fabricated from the same wafer are measured. The results are shown in Fig. 3(b). The turn-on times of the conventional edge-emitting laser are much longer than those in Fig. 3(a). The relaxation oscillations exist clearly in Fig. 3(b). The differences of transient responses between the QD microdisk laser and the edge-emitting laser can be explained by a rate-equation model incorporating the Purcell effect, homogeneous broadening, and nonradiative recombination. The rate equations for carrier density (N) and photon density (S) can be expressed as:
J and d are the injection current density and the thickness of the active layer, respectively. We separate the carriers into two groups, uN and (1-u)N [13]. u is the ratio of homogeneous broadening to the spectral width of spontaneous emission. u is estimated to be 0.12 for our sample. By the rectangle spectrum approximation, uN represents the carriers within the homogeneous broadening spectrum, and therefore can contribute to the lasing mode. In microdisks, the spontaneous emission lifetime τsp of resonant wavelengths may be shortened to τ’sp by Purcell effect [14,15]. Γ is the confinement factor, set to be 0.03. β is the spontaneous emission factor, defined by the percentage of spontaneous emission energy entering into the lasing mode. vg is the group velocity of light in the material. G(N) is the optical gain approximated by a logarithmic function G(N) = gln[(N+Ns)/(Ntr+Ns)] [16] with three parameters, g = 1300 cm−1, Ns = 0.9×1018 cm−3, and Ntr = 1.8×1018 cm−3. τnr and τp are the nonradiative carrier lifetime and photon lifetime, respectively.Fig. 3 The temporal optical responses measured from (a) a QD microdisk laser and (b) a conventional QD edge-emitting laser fabricated from the same wafer. The origins of time axis are defined by the rising edge of the electric pulses. The normalized injection currents are indicated in the graphs.
For the edge-emitting laser, we choose β = 8×10−5. There is no enhancement of spontaneous emission in this case such that τ’sp = τsp = 5.8 ns. The obviously long carrier lifetime from Fig. 3(b) is longer than those obtained from common InAs quantum dot samples. The reason can be due to two factors. First, this sample, with spacer thickness of 10 nm, consists of five layers of vertically coupled QDs. It has been reported that vertically coupled QDs show longer radiative lifetime [17]. Second, from a temperature-dependent carrier lifetime measurement of this wafer not shown here, the common decreasing carrier lifetime at higher temperature due to nonradiative recombination is not observed in this sample. The nonradiative recombination is not the dominant recombination mechanism in this wafer at room temperature. Therefore, the carrier lifetime is attributed to the radiative process and the spontaneous emission lifetime τsp is chosen as 5.8 ns. The nonradiative recombination term is neglected. The photon lifetime τp can be determined by τp = 1/(vgα), where α describes the total cavity loss. We choose α = 12 cm−1 for the edge-emitting laser. For the microdisk lasers, in order to account for the enhancement of spontaneous emission factor β in microcavity lasers, we set β = 3.5×10−3, which is of the same order of magnitude as that reported in the literature [2]. The spontaneous emission lifetime is shortened by a factor of three to τ’sp = τsp/3 = 1.93 ns. The assumed enhancement factor is similar to those in the literature [14,15] when taking into account the mode volume difference.
In microdisk lasers, the nonradiative recombination is an important issue [18]. Since the carriers are injected from the microdisk pedestal and diffuse to the periphery via the wetting layer, the carriers populating in the wetting layer before captured by QDs can be dissipated by the nonradiative recombination at the surface. The increased surface recombination rate in microdisks can be demonstrated in the time-resolved micro-photoluminescence (μ-PL) experiments. The microdisks for the time-resolved μ-PL measurement are fabricated from the same wafer, with the upper cladding layer removed. After the two-step wet-etching [4], three different cavity sizes, 6.5, 8.5 and 10.5 μm, is prepared for optical pumping. The samples are excited at room temperature by 790 nm ultrafast pulses with the repetition rate at 76 MHz and 130 fs pulse width. The laser excitation is focused by an objective lens onto the center of the microdisk cavities. The time-resolved PL signals are analyzed by the same setup as introduced previously for the transient measurements. The signals of the QD ground states are chosen by the monochromator. The purpose of performing the time-resolved optical pumping measurement is to determine the carrier lifetime and to study its size-dependence. The time-resolved μ-PL results are shown in Fig. 4 . An unprocessed sample is also measured for comparison. All time-resolved PL signals show a two-component decay profile. The fast decay at the beginning depends on the disk size with the shortest carrier lifetime from the smallest 6.5 μm microdisk. On the other hand, all four samples show a similar slow decay after the fast components die away. From experimental data, the fast decay time is size-dependent and can be attributed in general to either the reduced radiative lifetime by Purcell enhancement, or the reduced nonradiative lifetime by increasing the surface to volume ratio. For the 6.5 μm microdisk, which is also the cavity size of our current-injection microdisk device, the measured decay time is 0.5 ns. This value will be too short for this device if it is the radiative lifetime reduced by Purcell enhancement. Therefore, we conclude that the dominant mechanism is the nonradiative recombination for the 6.5 μm microdisk. To exclude the influence of Purcell effect and to extract the nonradiative lifetime for the rate-equation model, we choose the optical pumping intensity and excite the microdisk at the center by an objective lens so that we do not observe WGM peaks in the spectrum. Under this condition, the carrier lifetime measurements will not be influenced by the Purcell effect which will shorten the carrier lifetime. For the reasons stated above, we attribute the fast decay profiles to the nonradiative recombination at the surface. The smallest microdisk will have the fastest decay since its surface to volume ratio is the highest. The slow decay signals come from the recombination of the carriers apart from any surface. For the microdisk with diameter of 6.5 μm, this carrier lifetime of 0.5 ns is similar to the surface recombination lifetime of 0.3 ns estimated in other InAs QD microdisk lasers [18]. We choose τnr = 0.5 ns in our simulation.
The simulation results of temporal responses for both lasers are shown in Fig. 5 . It shows that the rate-equation model can describe the measured transient behaviors in Fig. 3 quite well. The relaxation oscillations of microdisk lasers are suppressed mainly due to the increased spontaneous emission factor β. This is an advantage that the optical signal will not suffer from the overshoot caused by relaxation oscillations. The reason that the turn-on of microdisk lasers is very fast when compared with edge-emitting lasers is the shortened carrier lifetime. In this case, the fast surface recombination plays an important role. From the rate-equation model, the fast carrier recombination rate may not only result from τnr, but also can be introduced by the enhanced spontaneous emission lifetime τ’sp. It is theoretically predicted that the Purcell effect can improve the modulation responses in vertical-cavity surface-emitting lasers [19]. Therefore, microdisk lasers can potentially exhibit fast turn-on, even when the surface recombination is largely reduced.
Fig. 5 The temporal optical responses calculated from the rate-equation model for (a) a QD microdisk laser and (b) a conventional QD edge-emitting laser. The injection starts at t = 0 ns. The normalized injection currents are indicated in the graphs.
The short turn-on time implies that the modulation speed of the microdisk laser under large-signal direct modulation can be fast. We demonstrate the large-signal direct modulation by driving the device with non-return-to-zero (NRZ) signals patterned at 1 gigabit per second (Gbps). In order to obtain the fast operation condition, we shorten the turn-on time further by increasing the injection current to Ipeak 0.97 mA. Also, we applied a small dc bias current Ibais below threshold. The laser is modulated between the two injection levels, Ipeak and Ibais. As shown in Fig. 6 , a bias current of 0.19 mA was applied to the device in this experiment. One can see that the pulse shape is successfully preserved in the optical signals. With pre-bias, the turn-on delay becomes shorter and the square-shaped signals can be represented more faithfully. We expect that it is possible for the device to show the modulation speed over 1 Gbps when optimally biased.
Fig. 6 The ideal data stream at 1 Gbps, the real electric pulse shape, and the optical responses of the QD microdisk laser with bias current of 0.19 mA. The peak current is 0.97 mA.
4. Modal transient behaviors under various temperatures
Next, we study the temporal optical response of the microdisk laser at different temperatures. Spectrally-resolved tranients at five temperatures have been measured. The threshold currents are 0.19, 0.16, 0.20, 0.32, and 0.47 mA for 100, 150, 200, 250, and 300 K, respectively. Among these temperatures, the minimum threshold current is observed at 150 K. This phenomenon is described as the negative characteristic temperature of QD microdisk lasers [4]. Figure 7(a) shows the observed WGM wavelengths at each temperature, measured at the injection level of 1.1 Ith. The electric pulse duration is 20 ns. At 100 K, there are two WGMs near 1014 and 1029 nm achieving steady-state lasing with almost identical threshold currents. The WGMs at 1014 and 1029 nm correspond to TE(1,55) and TE (1,54) mode. As temperature rises, the TE(1,54) mode shifts to near 1032 (150 K), 1035 (200 K) and 1039 nm (250 K). The wavelength redshift with increasing temperature is owing to the temperature-dependent cavity refractive index. All of the observed modes correspond to the QD ground states. A large wavelength jump to 1059 nm at 300 K is due to the QD energy shift, and the TE(1,53) mode is excited at room temperature.
Fig. 7 (a) The observed WGM wavelengths of a 6.5-μm-diamenter microdisk laser at different temperatures. (b)-(d): The temporal optical responses measured at (b) 100 K, (c) 150 K, and (d) 200 K. For 250 K and 300 K, there is only one WGM lasing with the temporal response similar to those in Fig. 3(a).
As shown in our previous temperature-dependent study [4], the number of lasing WGMs decreases at higher temperatures. This phenomenon is also observed in the transient measurements here. It can be explained by the carrier redistribution in QD ensemble. The temporal optical responses measured at 100-200 K are plotted in Fig. 7(b)-7(d). At 100 K, as shown in Fig. 7(b), two WGMs can achieve steady-state lasing. At 150 K, in Fig. 7(c), the WGM at 1017 nm only shows a transient lasing behavior with pulse duration of 0.8 ns. This implies that the quasi-equilibrium among carriers is not yet fully established during the turn-on delay (~1 ns) and the transient lasing duration 0.8 ns of high-energy 1017 nm WGM. Similar time scale for carrier redistribution among QDs at 150 K has been reported by a time-resolved PL study [20]. Only the WGM at 1032 nm remains lasing after the carrier redistribution process is complete. A similar phenomenon can be observed at 200 K, in Fig. 7(d), with a very small transient peak of 0.5 ns at 1020 nm. The WGM with lower energy (1035 nm) can remain steady-state lasing after the thermal redistribution. The time-resolved PL spectra at 250 and 300 K only show single WGM lasing with the temporal response similar to those previously shown in Fig. 3(a) and no transient peaks are observed at the neighboring WGM wavelengths. It implies that the carrier redistribution is fast established at these temperatures.
5. Conclusion
In this work, we studied the transient behaviors of current-injection QD microdisk lasers under large-signal direct modulation at room temperature. Suppressed relaxation oscillations and fast turn-on behavior are observed. With the help of rate-equation modeling, the suppressed relaxation oscillation is attributed to the enhanced spontaneous emission factor in microdisk lasers. Short turn-on time results from the reduced carrier lifetime caused by the Purcell effect and increased nonradiative recombination rate due to higher surface to volume ratio. Due to these phenomena, a large-signal direct modulation experiment at 1 Gbps is demonstrated. Operation at even higher bit rates is possible when these QD microdisk lasers are optimally biased. In temperature-dependent dynamic study, both of the transient lasing and steady-state lasing from side modes are suppressed at temperatures higher than 250 K. Therefore, the quantum-dot microdisk lasers demonstrate the potential of single-mode operation under high-speed modulation at room temperature.
Acknowledgment
The authors would like to thank Mr. S.-Z. Wu for providing measurement data of the edge-emitting laser and Mr. J. Z. Hong for his assistance in sample preparation. This work was supported by the National Science Council, Taiwan, under the Grant No. NSC-99-2628-E-002-025 and NSC-100-2628-E-002-016.
References and links
1. T. Baba, M. Fujita, A. Sakai, M. Kihara, and R. Watanabe, “Lasing characteristics of GaInAsP-InP strained quantum-well microdisk injection lasers with diameter of 2–10 μm,” IEEE Photon. Technol. Lett. 9(7), 878–880 (1997). [CrossRef]
2. M. Fujita, A. Sakai, and T. Baba, “Ultrasmall and ultralow threshold GaInAsP-InP microdisk injection lasers: design, fabrication, lasing characteristics, and spontaneous emission factor,” IEEE J. Sel. Top. Quantum Electron. 5(3), 673–681 (1999). [CrossRef]
3. D. Bimberg, M. Grundmann, and N. N. Ledentsov, Quantum Dot Heterostructures (John Wiley & Sons, 1999).
4. M.-H. Mao, H. C. Chien, J. Z. Hong, and C. Y. Cheng, “Room-temperature low-threshold current-injection InGaAs quantum-dot microdisk lasers with single-mode emission,” Opt. Express 19(15), 14145–14151 (2011). [CrossRef] [PubMed]
5. J. Van Campenhout, P. Rojo Romeo, P. Regreny, C. Seassal, D. Van Thourhout, S. Verstuyft, L. Di Cioccio, J.-M. Fedeli, C. Lagahe, and R. Baets, “Electrically pumped InP-based microdisk lasers integrated with a nanophotonic silicon-on-insulator waveguide circuit,” Opt. Express 15(11), 6744–6749 (2007). [CrossRef] [PubMed]
6. L. Liu, J. V. Campenhout, G. Roelkens, D. V. Thourhout, P. Rojo-Romeo, P. Regreny, C. Seassal, J. Fédéli, and R. Baets, “Ultralow-power all-optical wavelength conversion in a silicon-on-insulator waveguide based on a heterogeneously integrated III-V microdisk laser,” Appl. Phys. Lett. 93(6), 061107 (2008). [CrossRef]
7. L. Liu, J. Van Campenhout, G. Roelkens, R. A. Soref, D. Van Thourhout, P. Rojo-Romeo, P. Regreny, C. Seassal, J. M. Fédéli, and R. Baets, “Carrier-injection-based electro-optic modulator on silicon-on-insulator with a heterogeneously integrated III-V microdisk cavity,” Opt. Lett. 33(21), 2518–2520 (2008). [CrossRef] [PubMed]
8. K. J. Luo, J. Y. Xu, H. Cao, Y. Ma, S. H. Chang, S. T. Ho, and G. S. Solomon, “Dynamics of GaAs/AlGaAs microdisk lasers,” Appl. Phys. Lett. 77(15), 2304–2306 (2000). [CrossRef]
9. K. J. Luo, J. Y. Xu, H. Cao, Y. Ma, S. H. Chang, S. T. Ho, and G. S. Solomon, “Ultrafast dynamics of InAs/GaAs quantum-dot microdisk lasers,” Appl. Phys. Lett. 78(22), 3397–3399 (2001). [CrossRef]
10. S. M. K. Thiyagarajan and A. F. J. Levi, “Dynamic behavior of scaled microdisk laser,” Solid-State Electron. 45(10), 1821–1826 (2001). [CrossRef]
11. R. Ushigome, M. Fujita, A. Sakai, T. Baba, and Y. Kokubun, “GaInAsP microdisk injection laser with benzocyclobutene polymer cladding and its athermal effect,” Jpn. J. Appl. Phys. 41(Part 1, No. 11A), 6364–6369 (2002). [CrossRef]
12. S. M. K. Thiyagarajan and A. F. J. Levi, “High-speed response of optically-pumped InGaAs/InGaAsP microdisk lasers,” Electron. Lett. 37(3), 175–176 (2001). [CrossRef]
13. T. Baba and D. Sano, “Low-threshold lasing and Purcell effect in microdisk lasers at room temperature,” IEEE J. Sel. Top. Quantum Electron. 9(5), 1340–1346 (2003). [CrossRef]
14. B. Gayral and J. M. Gérard, “Strong Purcell effect for InAs quantum boxes in high-Q wet-etched microdisks,” Physica E 7(3-4), 641–645 (2000). [CrossRef]
15. W. Fang, J. Y. Xu, A. Yamilov, H. Cao, Y. Ma, S. T. Ho, and G. S. Solomon, “Large enhancement of spontaneous emission rates of InAs quantum dots in GaAs microdisks,” Opt. Lett. 27(11), 948–950 (2002). [CrossRef] [PubMed]
16. L. A. Coldren and S. W. Corzine, Diode Lasers and Photonic Integrated Circuits (John Wiley & Sons, 1995), chap. 2.
17. M. Colocci, A. Vinattieri, L. Lippi, F. Bogani, M. Rosa-Clot, S. Taddei, A. Bosacchi, S. Franchi, and P. Frigeri, “Controlled tuning of the radiative lifetime in InAs self-assembled quantum dots through vertical ordering,” Appl. Phys. Lett. 74(4), 564–566 (1999). [CrossRef]
18. T. Ide, T. Baba, J. Tatebayashi, S. Iwamoto, T. Nakaoka, and Y. Arakawa, “Lasing characteristics of InAs quantum-dot microdisk from 3 K to roomtemperature,” Appl. Phys. Lett. 85(8), 1326–1328 (2004). [CrossRef]
19. D. G. Deppe and H. Huang, “Quantum-dot vertical-cavity surface-emitting laser based on the Purcell effect,” Appl. Phys. Lett. 75(22), 3455–3457 (1999). [CrossRef]
20. A. Fiore, P. Borri, W. Langbein, J. M. Hvam, U. Oesterle, R. Houdre, R. P. Stanley, and M. Ilegems, “Time-resolved optical characterization of InAs/InGaAs quantum dots emitting at 1.3 μm,” Appl. Phys. Lett. 76(23), 3430–3432 (2000). [CrossRef]
