1. Introduction

Diode-pumped passively Q-switched all-solid-state lasers are widely used in material processing, micromachining, laser ranging and medical applications due to their compactness, low cost, simplicity and high efficiency. Possible saturable absorber materials to achieve passive Q-switching at 1 μm include color centers [1], Cr⁴⁺:YAG crystals [2] and semiconductors [3]. Among the semiconductor saturable absorbers, GaAs is desirable because of its ease of fabrication, high damage threshold and large optical nonlinearity. Passive Q-switching with GaAs has been investigated with many Nd or Yb doped gain media, such as Nd:YAG [4], Nd:YAG ceramic [5], Nd:GGG [6], Nd:YVO₄ [7], Yb:Y₂O₃ ceramic [8] and Yb:YAG [9]. GaAs has a band gap of 1.42eV, which is much higher than the 1 μm wavelength photon energy for such lasers. It is believed that saturation of the single photon absorption (SPA) of the EL2 defect, which forms a deep donor level at about 0.82 eV, is mainly responsible for the Q-switching effect. At high intra-cavity intensity, the nonlinear absorption due to two-photon absorption (TPA) and free-carrier absorption (FCA) could lead to a pulse width reduction effect with a steeper trailing edge observed [4,5]. Such passively Q-switched lasers with GaAs often show irregular variations of pulse repetition rate with pump power and exhibit relatively low pulse stability due to the relative complex transmission behaviors [6–9].

Yb doped gain media have the advantages of low quantum defect and long upper level life time, which are beneficial for high power Q-switched operation. As a disordered crystal, Yb³⁺:NaY(WO₄)₂ (Yb:NYW) has a large gain bandwidth due to the random distribution of the Na and Yb cations in the crystal space group, and has attracted significant interest recently. Spectroscopic properties and mode-locked operation of Yb:NYW have been investigated [10,11] and passively Q-switched operation has been realized with Cr⁴⁺:YAG saturable absorber [12]. In this study, we report the first passively Q-switched Yb:NYW lasers using a GaAs saturable absorber. In the present experimental configuration the GaAs was placed close to Yb:NYW crystal, the temperature of the GaAs was significantly affected by absorption of the residual pump light. For continuous pump conditions, stable Q-switched output with pulse duration as short as 5 ns has been obtained. However under these conditions the pulse shape indicated rapid truncation of the pulse tail indicative of a cut off process in the GaAs crystal. Under pulsed pump mode, longer pulse durations with more symmetrical pulse shape were observed with decreased pulse stability. This difference in behavior may indicate that GaAs might display different saturable absorption properties at different temperatures.

2. Experimental setup

The schematic diagram of the laser setup is shown in Fig. 1 . It is a plano-concave resonator, the pump source is a fiber-coupled (200 μm diameter, NA = 0.22) laser diode with a central wavelength around 976 nm. The radiation of the pump source was coupled into the laser crystal by two collimating lenses (11:8 demagnification) with a spot diameter of approximately 145 μm. M1 is a plane dichroic mirror, one face is antireflection-coated (AR) at 976 nm, the other face is high-reflection (HR) coated at 1010 nm to 1200 nm and high-transmission (HT) coated at 976 nm. The output coupler M2 is a curved mirror with a radius of curvature of 50 mm and the transmission of T = 3% at 1030 nm. The laser material was a 4.8%-Yb-doped Yb:NYW crystal with a dimension of 3 × 3 × 3 mm³; both end faces of the crystal have single layer MgF₂ AR coatings (R<1% from 1010 nm to 1100 nm). In order to remove the heat generated in the laser crystal, the Yb:NYW crystal was wrapped with indium foil and held in a water-cooled copper block. The water temperature was controlled to be 9 °C throughout the experiment. The saturable absorber was a (100) cut semi-insulating GaAs wafer with a thickness of 380 μm and an initial transmission of 92%, both faces are AR coated (R<1% from 1020 nm to 1080 nm). The GaAs wafer was held in an aluminum block without extra cooling. To get a good overlap between the laser cavity mode and the pump beam, the laser crystal was put close to M1. The GaAs crystal was placed close to the laser crystal to get a small beam radius (w ~90μm) on the saturable absorber. A thermistor was used to measure the temperature of the GaAs wafer at a position 5 mm from the laser spot.

 

Fig. 1 Experimental Setup: LD, pump diode laser at 976 nm; M1, dichroic mirror; M2, output coupler.

Download Full Size | PDF

3. Results and discussions

At first, the CW output characteristics of the laser were investigated without the GaAs saturable absorber. Maximum output power was obtained when the cavity length was ~35 mm. As can be seen in Fig. 2(a) , at an absorbed pump power of 12 W, 3.5 W output power at 1031 nm was obtained, corresponding to an optical-to-optical efficiency of 29% and a slope efficiency of 35%. The output power scaled linearly with the increase of pump power, and no saturation was found within the pump power range.

 

Fig. 2 (a) CW output power and (b) Q-switched average output power versus absorbed pump power.

Download Full Size | PDF

Passively Q-switched operation was obtained by placing the GaAs wafer as close as possible to the laser crystal. As can be seen in Fig. 2(b), the output started to saturate when the output power reached 120 mW. The observed saturation of the output power is an indication of an increase of losses in the cavity. Since no saturation was observed at this pump power in CW operation, this might be due to the thermal effects within the GaAs crystal. The thermistor temperature of the GaAs crystal was measured to be approximately 35°C and 55°C at absorbed pump powers of 4.7 W and 12 W, respectively. The temperature of the central 200 μm region of the GaAs crystal was estimated to be 42°C and 74°C using a steady state cylindrical heat flow model driven by absorption of residual pump light. The comparatively larger variation of the output power versus pump power at high pump power is possibly a sign of the onset of nonlinear absorption mechanisms of TPA and FCA [4]. With the increasing of the pump power, the output laser wavelength shifted from 1026 nm to 1019 nm. As a quasi-three level gain medium, the peak gain wavelength will move to shorter wavelength at higher population inversion levels due to reduced re-absorption from the ground level, as described in references [11,13]. Thus the shifted wavelength is indicative of higher cavity losses at high pump levels which result in the saturation of output versus pump power.

The variation of pulse repetition rate, pulse width, pulse energy and the peak power with pump power are shown in Fig. 3(a) and 3(b). The pulse repetition rate increased monotonically until the pump power reached ~11 W. However average output power increased more slowly with pump power, as shown in Fig. 2(b), which result in decreasing pulse energy with the increase of pump power as shown in Fig. 3(b). The highest pulse energy of 4.5 μJ was obtained at a low pump power of 4.7 W. The pulse width decreased monotonically with increasing pump power. At a pump power of 12 W, Q-switched output has been obtained with a pulse repetition rate of 83 kHz, pulse energy of 2.2 μJ and pulse duration of 5 ns, corresponding to a peak power of 0.44 kW. The typical pulse train and pulse shape are shown in Fig. 4(a) and 4(b). The pulse period is stable with a rms timing jitter less than 2%; the measured rms variations of pulse intensities was less than 3%, which is partly due to the limitation of sampling rate of digital oscilloscope used in experiment. The output stability shows considerably improvement compared to other lasers Q-switched with GaAs saturable absorbers [6–8] which displayed 4-15% rms jitter. The pulse duration of 5 ns is much shorter than that reported of 26 ns for a passively Q-switched Yb:NYW laser with Cr⁴⁺:YAG saturable absorber [12]. Compared with other lasers Q-switched with GaAs, the pulse shows a more asymmetric shape with a much steeper trailing edge than leading edge, as shown in Fig. 4(b), which indicates a quick turn-off of transmission generating a shorter truncated pulse.

 

Fig. 3 (a) Pulse repetition rate and pulse width and (b) pulse energy and peak power versus absorbed pump power.

Download Full Size | PDF

 

Fig. 4 (a) Trace of the Q-switched pulse train and (b) pulse shape at the maximum absorbed pump power of 12 W.

Download Full Size | PDF

Figure 5 shows the typical pulse train and pulse shape for an absorbed pump power of 4.7 W. Compared with Fig. 4, we can see at low pump level, the pulse train has apparently poor timing stability. The single pulse has a long tail, not like the truncated pulse in Fig. 4(b) obtained under high pump power. We suspect this is related to the GaAs wafer temperature increasing at high pump power. In particular, in the present geometry a significant fraction of pump laser power is directly absorbed by the GaAs crystal. In previous lasers Q-switched with a GaAs saturable absorber, usually a curved dichroic mirror was used for pump coupling and the GaAs wafer was placed near the output coupler or used as an output coupler to obtain good focusing on it. The effects of the unabsorbed pump light on GaAs would be much smaller in these cases. In our experiment, a plane dichroic mirror was used to couple the pump light and the GaAs sample was placed very close to the laser crystal. A significant temperature increase of the GaAs at high pump power can cause an increase of TPA coefficient due to band gap shrinking with temperature [14], which can induce an increase of the absorption and lead to the rapid cut off of the pulse trailing edge. However further study will be required to better understand the detailed physicals mechanisms associated with the phenomena.

 

Fig. 5 (a) Trace of the Q-switched pulse train and (b) typical pulse shape at an absorbed pump power of 4.7 W.

Download Full Size | PDF

To further investigate the temperature effects of GaAs for Q-switched operation, we switched the pump laser to a pulsed mode without any change of the cavity. The pump period was 1 ms and duty cycle was 35%; the rise and fall times of the pump pulses were both 50 μs. Therefore the average pump power was ~35% of that of the CW pump at the same pump current. The GaAs has around 600 μs for cooling between pulses. With the same pump current as that CW pump of 12 W, an output power of 155 mW has been obtained, corresponding to an improvement of laser efficiency of 246% compared with that of CW pump. Under this condition, the thermistor temperature of the GaAs was measured to be approximately 35 °C and the central temperature was estimated to be an average of 41°C due to average absorbed pump power at 35% duty cycle with some additional heating within each pump pulse period. This indicates that the additional heating from the full CW pumping significantly reduced the laser efficiency. When the pump current was just above threshold for Q-switching, a single output pulse of 5 μJ energy and 3 μs duration was obtained for each pump period with an rms timing jitter less than 0.5%, as shown in Fig. 6(a) . With further increase of the pump current, a pulse train was obtained for each pump period. Figure 6(b) and 6(c) show the typical pulse train and pulse shape for pulsed pumping with a peak absorbed pump power of 12 W. Within the pulse train the repetition rate is approximately 200 kHz, which is much higher than the CW case. The pulse to pulse timing stability within the pulse train burst now had a measured rms jitter of ~11%. The pulse shortening effect did not appear and the typical pulse duration was ~70 ns as seen in Fig. 6(c).

 

Fig. 6 (a) Trace of the Q-switched pulse train under pulsed pump conditions with an absorbed peak pump power of 5.6 W; (b) Trace of the Q-switched pulse train during one duty cycle with an absorbed peak pump power of 12 W; (c) Typical pulse shape under pulsed pump conditions with an absorbed peak pump power of 12 W.

Download Full Size | PDF

Relaxation oscillation is a property of laser systems when the upper-state lifetime of the gain medium is much longer than the cavity damping time and the pump power is turned on abruptly. Removing the GaAs saturable absorber from the cavity, self-pulsing was observed with relaxation oscillations under pulsed pump mode. When the pump current was just above the threshold, single pulse output could be achieved for each pump period. Figure 7(a) and 7(b) show the pulse train and pulse shape for a pump at a 4 kHz repetition rate and duty cycle of 60%. The pulse duration is ~1.1 μs and the pulse energy is 3 μJ. Higher pulse intensities can be expected if the cavity is modified, such as by using a higher transmission output coupler. Although the output peak power is limited, the output has very good timing stability with a pulse to pulse timing jitter of less than 0.5%. This configuration may find use in applications that require high pulse repetition rate stability but not high peak powers.

 

Fig. 7 Output with pulsed pumping and no GaAs saturable absorber in the cavity for a peak absorbed pump power of 4 W: (a) the output pulse train (below) accompanied with pulsed pump train (above) and (b) the single pulse shape.

Download Full Size | PDF

4. Conclusion

In conclusion, we have demonstrated both GaAs passively Q-switched and gain Q-switched Yb:NYW lasers. For CW pumping, significant pulse shortening and better pulse stability were obtained at high pump power yielding pulses with pulse duration of 5 ns, pulse repetition rate of 83 kHz, peak power of 0.44 kW and pulse to pulse rms timing jitter of less than 2%. Much longer pulse durations and poorer timing jitter were observed at lower pump powers where the temperature of the GaAs crystal was significantly lower, indicating that temperature could be an important factor when GaAs is used as a saturable absorber. With pulsed pumping at a 1 kHz repetition rate, better stability but longer pulses were achieved when pumping near threshold. With pulsed pumping at high pump powers a pulse train burst was generated with pulse durations down to 70 ns. Gain switched lasing without a saturable absorber with 4 kHz pulsed pumping near threshold gave reproducible 1.1 μs pulses with 3 μJ energy. The combination of a GaAs saturable absorber and pulsed pumping allows for the generation of microjoule pulses with a range of output pulse durations from 5 ns to a few μs from a Yb:NYW crystal laser. Such pulses should be suitable for a variety of applications.