Open Access
Add to BibSonomyAbstract
A technique for stabilizing the repetition frequency of a passively Q-switched laser is presented using an optically driven surface-normal semiconductor modulator. A method is capable of significant reduction of the timing jitter in a passively Q-switched laser by optical triggering the saturable absorber semiconductor reflector. The experimental demonstration using passively Q-switched ytterbium-doped fiber laser shows the jitter reduction by factor of 1.66×103 from 50 µs down to 30 ns.
©2008 Optical Society of America
1. Introduction
In different applications of pulsed laser systems, it is necessary to have an accurate control of the temporal characteristics of the pulse train including both the pulse width and the repetition rate. It is known that the duration of the pulses passively Q-switched using saturable absorbers depends critically on the cavity round-trip time, modulation depth of the absorber and the output coupler reflectivity [1–3]. Very recently, a novel mechanism of pulse shortening in a Q-switched laser induced by the gain compression effect under strong pumping conditions has been reported [4]. In passively Q-switched fiber lasers, using high modulation depth saturable absorbers have proved to be an efficient method in short pulse generation. With 70% modulation depth absorber, we have achieved recently a 8-ns pulses, as described in [5]. The modulation depth ΔR of the absorber is determined mainly by the value of the low-intensity reflectivity RL-I assuming low non-saturable losses (αN-S≪1) in the absorber mirror, ΔR=1-RL-I-αN-S. The highest usable modulation depth, therefore, depends largely on the round-trip gain in the cavity, which is essentially high for fiber lasers. Consequently, the usable value of the modulation depth ΔR could be very large particularly for fiber systems providing convenient means of pulse shortening for passive Q-switched operation. The practical fiber system should, however, in addition offer the high stability of the pulse train repetition rate, whereas the timing jitter of passively Q-switched fiber lasers is rather large, typically of few tens of microseconds. It is well documented that a large timing jitter of a passively Q-switched lasers is caused by fluctuations in temperature, pump, loss, etc, and recognized as an intrinsic and inevitable feature because the “first photon” comes to the lasing mode from noise-like spontaneous emission and imposes a fundamental limit on the stability of the repetition rate [6]. This issue is particularly operative in fiber lasers characterized with a high level of intracavity spontaneous emission owing to the waveguiding geometry and relatively long-length cavity. Intensity of the amplified spontaneous emission captured in a fiber cavity could be comparable with the intensity of the laser emission and, therefore, it can significantly affect the start-up and temporal stability of the pulse regime [7]. In this study we demonstrate a novel method for timing jitter reduction in a passively Q-switched laser using an optically-driven saturable absorber reflection modulator (SARM) that provides both passive pulse shaping and active locking of the pulse train to periodic control signal, simultaneously.
2. Experimental details
The linear fiber cavity of the Q-switched laser studied here is terminated by lens-coupled SARM and a narrow bandwidth fiber Bragg grating (FBG) with reflectivity of 65% at 1035 nm, as shown in Fig. 1. The core-pumped 10-cm-long ytterbium doped gain fiber has absorption of 1348 dB/m at 980 nm. The cavity roundtrip time is 11.6 ns with a focused beam diameter at the modulator of 8.4 µm.
Fig. 1. Q-switched fiber laser with saturable absorber mirror acting also as a modulator triggered with a control signal. Fiber Bragg grating (FBG) and the modulator define the linear cavity. The fiber coupled pump diodes are protected by free-space isolators. The absorber modulator was protected against residual (non-absorbed in ytterbium fiber) CW 980-nm pump by second 980/1035-nm dichroic coupler used for launching into the cavity 980-nm control signal.
The modulator used in this study was fabricated by solid-source molecular-beam epitaxy on n-type GaAs (100) substrate. The sample includes a bottom mirror comprising 30.5 pairs of AlAs/GaAs quarter-wave layers forming a distributed Bragg reflector (DBR) with the center wavelength of 1060 nm and a 100-nm bandwidth. The absorber section comprises 13 InGaAs quantum wells with 6.9-nm thickness. To enhance the modulation response of the SARM, the absorbing quantum wells were placed at the antinodes of the optical standing-wave pattern formed in the microcavity. The modulator is built of high-quality lattice-matched quantum-well structure to achieve long recovery time of absorption (>1 ns) and thus to reduce possible loss for Q-switched pulse. Due to this, the non-bleachable loss was also very low (≤0.5%). When this is a case, the resonant dip at the wavelength of 1035 nm in the low-intensity reflectivity spectrum shows the achievable modulation depth of 35% defined as the difference between unity and low intensity reflectivity, as it is defined in the introduction and shown in Fig. 2. In addition to the normal use of the absorber mirror, the laser setup allows for optical modulation of the reflection response of the absorption using a control signal.
Fig. 2. Low-intensity reflectivity spectrum of the modulator structure used in this study. Repetitive passive Q-switching was achieved near 1035 nm resonant wavelength of the monolithic semiconductor microcavity formed by DBR and Fresnel reflection at semiconductor-air interface.
The Bragg wavelength of fiber grating was chosen to be centered at the 1034.5-nm corresponding to the resonant wavelength of the semiconductor modulator and during Q-switched operation it preserved the spectrally narrow-bandwidth performance. In our experiments, the Q-switched single-mode fiber laser was core-pumped with a fiber-coupled, 125 mW, 980-nm diode laser. The optical control signal was generated using another 980-nm diode laser by direct current modulation. The control signal at 980 nm induces reflectivity change at lasing wavelength of 1035 nm by partial saturation of the quantum well absorption. The Q-switched pulse then tends to be temporally trapped by the lower-loss absorption window. The control pulses locked firmly the pulse repetition rate near the frequency of the free-running purely passive Q-switched operation in a range of ~(1÷1.6)×fFREE-RUN, where fFREE-RUN was determined at a given pump power when the control signal was switched off. Robust locking could be achieved with low-energy modulation. The typical average power of the control signal at 5 kHz and 750 ns pulse width was 144 µW. This gives pulse energy of 28.8 nJ and demonstrates that low-energy control signal is suitable for synchronization and jitter reduction.
Optical modulation response was studied in reflection scheme. Low-intensity cw probe signal at resonance wavelength (1035 nm) and control pulse train at 980 nm were focused on the sample similar to the Q-switched laser setup described in Fig. 1. An additional 75/25-tap coupler was used for monitoring purposes. The reflectivity of the modulator was 65 % at 1035 nm, when modulation was turned off. Modulation response was found from the temporal reflectivity change induced by the control signal, as seen from Fig. 3. The amplitude of the modulation signal was identical to that used in the locking experiment. The modulator reflectivity was found to increase up to 81.5 % at the control pulse time slot. It is, however, obvious that high-energy Q-switched pulse provides the dominant saturation of the absorption resulting in modulator reflectivity of 99.5 %.
Fig. 3. Modulation response monitored with the probe signal at 1035 nm. Control signal at 980 nm induces a reflectivity change due to partial saturation of the absorption.
3. Results
Figure 4 shows the frequency locking performance of Q-switched laser for 3 different pump powers corresponding to fFREE-RUN of 2, 5 and 10 kHz. As it can be seen from this plot, with an increase in the modulation frequency fMOD, the Q-switched pulse repetition rate fREPRATE first increases accordingly (fREPRATE=fMOD) until fMOD exceeds the value of ~(1.4–1.6)×fFREE-RUN which denotes the upper level of the locking bandwidth ΔΩlock indicated in Fig. 4 by the length of the arrows. For higher modulation frequencies, the pulse repetition rate switches to the certain subharmonic M of the modulation frequency, fREPRATE=fMOD/M. For instance, for the Q-switching regime with fFREE-RUN=5 kHz, the 11th subharmonic at fMOD=83 kHz resulted in fREPRATE=fMOD/M≈7.55 kHz=1.5×fFREE-RUN.
Fig. 4. Q-switched pulse repetition rate for three different pump powers corresponding to the free-running frequencies fFREE-RUN of 2, 5 and 10 kHz. Modulation frequency was ramping up in this experiment starting from the fFREE-RUN.
Fig. 5. Optical spectra for identical pump power and different repetition rates set by the control signal.
As it is generally expected, the pump efficiency decreases at low pulse repetition rates in a manner consistent with the spontaneous-relaxation time resulting in a reduced inversion density and time-averaged output power of the laser. This feature could be observed as an increase of the background non-lasing radiation in the optical spectrum for low repetition frequencies, as seen from Fig. 5. The typical average output power was of the order of 1 mW.
The oscilloscope trace reveals the Gaussian shape of the pulse with a width of ~80 ns nearly independent on the repetition rate. The pulse-to-pulse timing jitter of the Q-switched fiber laser was measured directly from oscilloscope traces. It can be seen from Fig. 6(a) that without modulation-induced stabilization, the Q-switched pulse builds up with a large timing jitter of ~50 µs. By contrast, with the control signal activated, timing jitter remained below 30 ns, as shown in Fig. 6(b). In Fig. 6(a) the oscilloscope has been triggered on one pulse, and the subsequent pulse has been recorded to a histogram for 2 s. In Fig. 6(b) the oscilloscope has been triggered to the control pulse, and the timing jitter of the first pulse was recorded with similar histogram. The technique used in Fig. 6(a) was not possible in this case, since the jitter was too low to be observed at that scale.
Fig. 6. Timing jitter shown as a histogram recorded for 2 s. With the activated control signal, the timing of the Q-switched pulse is temporally locked to the control pulse resulted in the jitter reduction from 50 µs (a) to 30 ns (b) at the repetition rate of fREPRATE=5 kHz. The timing jitter has been defined as shown in the Figure by arrow gap.
Fig. 7. The relative time location of the control and Q-switched pulses for the locking state. For larger time delay between the pulses, the synchronization of the Q-switched pulse repetition rate to the control signal could not be achieved. (Control and Q-switched pulse amplitudes are not in scale).
Locking performance and timing jitter were further studied by tuning the Q-switched pulse location in the time domain in respect to control pulse for various pump powers. Figure 7 illustrates the range of relative pulse detunings that ensures the pulse synchronization to the locking signal. The control pulse with duration of 750 ns is seen in the figure as a small-amplitude square-shaped pulse superimposed with Gaussian-like Q-switched pulse. The amplitude of the Q-switched pulse in these measurements was strongly attenuated for clarity of the scope traces. The uppermost pulse trace in this figure indicates the largest detuning that still provides the firm pulse locking. For larger shifts, the laser operates in the free-running passive Q-switching mode.
The timing jitter of the locked pulses gradually decreases with the pump power, as seen from Fig. 8. This observation is in agreement with results presented in Fig. 4, where the locking bandwidth ΔΩlock for pulse repetition rate at the fundamental harmonic of modulation frequency (fREPRATE=fMOD) is shown to increase with free-running frequency that is proportional to the pump power. Finally, the measurements confirm that the locking bandwidth scales up with the duration of the control pulse, as seen from Fig. 9.
4. Conclusions
In conclusion, we described a novel technique for controlling Q-switched pulse timing demonstrated experimentally using surface-normal semiconductor modulator in an Yb-doped fiber laser. The dependence of locking bandwidth and pulse timing jitter on the parameters of control pulse and pump power was thoroughly studied. The Q-switched ytterbium fiber laser shows more than by factor of 1.66×103 reduction in timing jitter from 50 µs down to 30 ns with the proposed modulation technique.
Acknowledgments
This work was supported by Emil Aaltonen Foundation, Vilho, Yrjö ja Kalle Väisälä Foundation, Finnish Foundation for Technology Promotion, Jenny and Antti Wihuri Foundation, and National Graduate School in Materials Physics. The authors also wish to thank Mr. Tero Kontkanen for designing the control electronics used for modulation.
References and links
1. J. J. Zayhowski and C. Dill III, “Diode-pumped passively Q-switched picosecond microchip lasers,” Opt. Lett. 19, 1427–1429 (1994). [CrossRef]
2. J. J. Zayhowski and P. L. Kelley, “Optimization of Q-switched Lasers,” IEEE J. Quantum Electron. 27, 2220–2225 (1991). [CrossRef]
3. J. J. Degnan, “Optimization of Passively Q-switched Lasers,” IEEE J. Quantum Electron. 31, 1890–1901 (1995). [CrossRef]
4. R. Herda, S. Kivistö, and O. G. Okhotnikov, “Dynamic gain induced pulse shortening in Q-switched lasers,” Opt. Lett. 33, 1011–1013 (2008). [CrossRef] [PubMed]
5. T. Hakulinen and O. G. Okhotnikov, “8 ns fiber laser Q-switched by the resonant saturable absorber mirror,” Opt. Lett. 32, 2677–2679 (2007). [CrossRef] [PubMed]
6. J. B. Khurgin, F. Jin, G. Solyar, C. Wang, and S. Trivedi, “Cost-effective low timing jitter passively Q-switched diode-pumped solid-state laser with composite pumping pulses,” Appl. Opt. 41, 1095–1097 (2002). [CrossRef] [PubMed]
7. R. Herda and O. G. Okhotnikov, “Effect of amplified spontaneous emission and absorber mirror recovery time on the dynamics of the mode-locked fiber lasers,” Appl. Phys. Lett. 86, 011113 (2005). [CrossRef]
