Ultrashort superradiant pulse generation from a 1580 nm AlGaInAs multiple quantum-well (MQW) semiconductor structure has been experimentally demonstrated for the first time. Superradiance is confirmed by analyzing the evolution of the optical temporal waveforms and spectra. Superradiant trends and regimes are studied as a function of driving condition. An optical pulse train is obtained at 1580 nm wavelength, with pulse durations as short as 390 fs and pulse peak powers of 7.2 W.
© 2012 OSA
Superradiance (SR) is a cooperative radiation arising from spontaneous emission . Owing to its potential for the generation of ultra-short high power optical pulses, SR has been investigated both theoretically and experimentally [2,3] since the concept was first proposed in 1954 . SR emission has been of particular interest because of its unique properties namely, spatial coherency and large field amplitude. Such high power femtosecond pulses show great potential for many applications such as medical systems, imaging systems, sensing devices and bio-photonics . Until recently, both theoretical and experimental studies have been focused on atomic and molecular systems with external optical pumping. The main difficulty in obtaining SR emission from semiconductor structures is the very short polarization relaxation time (<100 fs at room temperature) and the need to attain the extremely high concentrations of the electron-hole (e-h) pairs. Although superradiant emission from semiconductor material systems has been previously studied at helium/cryogenic temperatures [6–8], until recently SR, has only been experimentally reported in bulk semiconductor (GaAs/AlGaAs heterostructures) at room temperature at wavelengths around 880 nm . Cooperative emission from an e-h ensemble in QWs placed in a magnetic field was previous reported . In addition, femtosecond SR pulses have been recently observed in violet InGaN quantum-well laser structures at room temperature .
In this letter, we report for the first time the experimental demonstration of femtosecond SR emission from a long-wavelength MQW device. SR is confirmed by analyzing the evolution of the optical spectra, and superradiant trends and regimes are studied as a function of driving condition. A 5 MHz optical pulse train has been obtained in the 1.55 μm wavelength window, with pulse durations as short as 390 fs and pulse peak powers of 7.2 W.
The epitaxial structure of the QW device under investigation is illustrated in Fig. 1 . The structure was grown on an InP substrate using molecular beam epitaxy (MBE). The active region incorporates five AlGaInAs QW active lasers consisting of 6 nm thick Al0.24GaIn0.71As quantum-well (refractive index n = 3.4845) and 10 nm thick Al0.44GaIn0.49As barrier layers (n = 3.3988). The active region is sandwiched between two Al0.9GaIn0.53As confinement layers (with a wider bandgap and n = 3.238). The top confinement layer is p-doped with zinc and the bottom is n-doped with silicon. Such materials with wider band-gaps offer confinement of the carriers in the active region. The cladding layers have a lower refractive index compared with the confinement and active layers and the lateral guide of the optical signal is formed within the cladding layers. The top p+ cladding layer is p-doped with zinc while the bottom n– cladding layer is n-doped with silicon. The metal layer at the top forms the contact pads and a heavily p-doped layer is implemented between the metal layer and the top cladding layer to facilitate a fast carrier transfer to the active region.
Standard photolithography and ICP dry etch processes were used to fabricate the 3 μm wide ridge waveguide structures. In order to achieve SR, it was necessary to split the p-type metal contact to define the gain sections and the absorber (as shown in Fig. 1). The contact separation was realized through post-processing of a two-section device using a focused ion beam (FIB) etching technique. A trench was etched parallel to the device facets, forming a 415 μm long gain section and a 200 μm long absorber section.
There are two critical criteria required to achieve SR emission in semiconductor structures at room temperature: (i) high current pumping giving large internal carrier densities and high optical gain, namely the realization of the condition αL ≫ 1, (ii) resonant emission at the bandgap photon energies . Here α is the small-signal gain and L is the length of the structure. These criteria can be experimentally implemented in multiple section devices by applying a strong current pulse to the gain section. With the other reverse biased section in a strongly absorptive state, high optical gain can build up and suppress the intraband relaxation processes at the early stage of evolution of SR pulses. Therefore, a pulsed drive current of up to 400 mA is used in the gain sections of the investigated QW. At the same time, a dc reverse bias from 0 to 8 V is applied to the absorber section in order to prevent lasing and achieve higher levels of e-h densities in the structure. The evolution of the output emission is shown in Fig. 2 .
The drive pulse amplitude is fixed at 400 mA while the reverse bias voltage is varying from 0 to 8 V to control the optical round trip gain profile. Initially, stimulated emission can be observed at lower bias voltage (i.e. low e-h density) in the form of pulse modulation and Q-switching. As the bias voltage changing from 0 to 2.5 V, the absorption effect is almost negligible under the strong pumping (400 mA). The device is therefore operated in the form of pulse modulation, where the laser output is controlled by the on and off of the electrical gain driving pulses and the pulse width of the output optical pulses of a few nanoseconds is comparable to that of the electrical driving pulses (9 ns). When the saturable absorber starts to function strongly as the bias voltage increases from 2.5 V up to 5.7 V, Q-switched pulses can then be observed from the device. However, at this stage, the intraband phase relaxation dominates over the accumulation of coherent collectively paired oscillators in the QW structure, and the build-up of SR optical field therefore cannot be observed under such driving conditions as lasing prevents the growth of the inversion at the threshold. There is a transition (from 5.7 V to 6.4 V) between the Q-switching and SR regimes, and the optical spectrum drifts dramatically (>20 nm) toward longer wavelength during this transition. This red shift corresponds to the band gap shrinkage as the e-h density increases with increasing reverse bias . At higher reverse voltages SR emission from a coherent e-h ensemble can be observed when mutual phasing of the individual e-h pairs occurs and a coherent cooperative state forms right at the band gap.
Ultrashort SR pulses are observed in the experiment from a narrow spectral range at the lowest possible photon energy. This is a result of the depletion of the electrons in the conduction band, cooperative e-h pairing and their condensation at the bottom of the band . In contrast to the pulse modulation and Q-switching spectra, which exhibit longitudinal modes, the SR spectrum is continuous without any mode structure and hence SR is considered as the collective emission of the correlated e-h pairs (rather than lasing). The shape of the continuous spectra of the cooperative recombination is asymmetric with the long wavelength edge being steeper than the short wavelength. Including the large red shift, this is exactly the same spectral behavior as has been previously observed in both GaAs/AlGaAs  and GaN/InGaN  structures.
Figure 3(a) shows the regimes of SR emission as a function of forward pulsed current to the gain section and reverse bias voltage to the absorber section. The device can also work in different operating regimes under different driving conditions: pulse modulation and Q-switching at low bias voltages and SR at high bias voltages. A train of single SR pulses with femtosecond pulse duration can be observed for drive currents from 275 to 325 mA while bursts of multiple SR pulses are detected at higher drive currents. Multiple SR pulses can also be observed for electrical drive pulse widths greater than 9 ns at lower drive currents. The individual pulse width however always lies in the sub-picosecond range. The average output power as a function of driving conditions is shown in Fig. 3(b). It should be noted that different operating regimes can also be reflected by the average output powers. The average power decreases rapidly during the transition along with the associated red shift of the optical spectrum.
For optimized SR operation, the gain section is driven at 325 mA, whilst a −5.95 V bias voltage is applied to the absorber section. As shown in Fig. 4(a) , a SR pulse is obtained with an optical pulse duration of 390 fs at a pulse repetition rate of 5 MHz. The estimated pulsepeak power is 7.2 W with a pulse energy of around 3 pJ. Figure 4(b) shows the optical spectrum under such operation. The central emission wavelength is 1583.6 nm, with a spectral width of 3.67 nm. The time-bandwidth product of the SR pulses is approximately 0.2, which is very close to the theoretical value. Therefore, in contrast to gain- and Q-switching the superradiant pulse generation produces bandwidth-limited pulses. The small spectral peak observed from the shorter wavelengths (~1515 nm) (see Fig. 4b) corresponds to the nanosecond spontaneous emission background pulse with much higher photon energies in the QW structure. Femtosecond SR pulses are always situated on top of such background pulses [2, 9].
It should be noted that one of the most interesting features of SR emission is its macroscopically large timing jitter and amplitude fluctuation. The SR noise performance differs from that of lasing. The reason for the presence of the large timing jitter and amplitude fluctuation is that the formation of the SR pulses is initially triggered by a random spontaneous noise inside the cavity and this leads to an uncertainty of the SR pulse start time . The fundamental physical phenomenon behind this is quantum-mechanical fluctuations of initial e-h polarization in the structure just before the pulse emission. It is interesting to note that there are generally two main reasons for the observation of a noisy autocorrelation trace: (1) large timing jitter and (2) instability of the pulse energy. Therefore, the noisy autocorrelation trace shown in Fig. 4(a) is as a result of the presence of the large timing jitter and amplitude fluctuation of SR pulses.
SR emission from QW structures can serve as a physical process that allows us to study microscopic quantum fluctuations of e-h ensembles through their macroscopic manifestations. Significant pulse-to-pulse variations in both amplitude and time domains have been predicted in many theories starting from early studies in 1970s . It was calculated that intensity variations could be as large as up to 50-60% of the average intensity of SR pulses. At the same time, timing jitter may be larger than 60-70% of the mean delay time of the SR pulses. The latter value exceeds a few picoseconds in case of a semiconductor medium. These facts make the SHG autocorrelation traces of SR pulses much less pure than in the case of mode locking.
To conclude, we have experimentally demonstrated SR emission from an AlGaInAs MQW semiconductor structure at room temperature for the first time. SR pulse evolution and different operating regimes have been measured as a function of driving conditions varying both the forward pulsed current to the gain section and the reverse bias voltage to the absorber section. A 5 MHz SR pulse train has been obtained at an emission wavelength of 1.58 μm, with a pulse width as short as 390 fs and pulse peak power as high as 7.2 W. Similarly to the SR pulse generation in bulk GaAs/AlGaAs structures, enhanced timing jitter and pulse envelopes variations of the generated ultrashort pulses have been observed.
The authors would like to thank the UK Engineering and Physical Science Research Council for support of this work. The fabrication of the laser structures under test was carried out by CST Global.
References and links
1. A. V. Andreev, V. I. Emel'yanov, and Y. A. Il'inskii, Cooperative Effects in Optics: Superradiance and Phase Transitions (IOP, 1993).
2. P. P. Vasil'ev, “Femtosecond superradiant emission in inorganic semiconductors,” Rep. Prog. Phys. 72(7), 076501 (2009). [CrossRef]
4. R. H. Dicke, “Coherence in spontaneous radiation processes,” Phys. Rev. 93(1), 99–110 (1954). [CrossRef]
5. C. M. Eigenwillig, B. R. Biedermann, G. Palte, and R. Huber, “K-space linear Fourier domain mode locked laser and applications for optical coherence tomography,” Opt. Express 16(12), 8916–8937 (2008). [CrossRef] [PubMed]
6. V. I. Yukalov and E. P. Yukalova, “Dynamics of quantum dot superradiance,” Phys. Rev. B 81(7), 075308 (2010). [CrossRef]
8. Y. D. Jho, X. Wang, J. Kono, D. H. Reitze, X. Wei, A. A. Belyanin, V. V. Kocharovsky, V. V. Kocharovsky, and G. S. Solomon, “Cooperative recombination of a quantized high-density electron-hole plasma in semiconductor quantum wells,” Phys. Rev. Lett. 96(23), 237401 (2006). [CrossRef] [PubMed]
9. V. F. Olle, P. P. Vasil’ev, A. Wonfor, R. V. Penty, and I. H. White, “Ultrashort superradiant pulse generation from GaN/InGaN heterostructures,” Opt. Express 20(7), 7035–7039 (2012). [CrossRef]
10. F. Haake, H. King, G. Schroder, J. Haus, and R. Glauber, “Fluctuations in superfluorescence,” Phys. Rev. A 20(5), 2047–2063 (1979). [CrossRef]