1. Introduction

Recently, polycrystalline ceramics as a new type of laser gain host have attracted considerable attention [1–4]. Comparing with the single crystal laser gain media, ceramic gain media have several advantages: high doping concentration and large size ceramic samples can be easily obtained; multiplayer and multifunctional ceramic laser materials are possible due to the polycrystalline nature of ceramics. Potentially the cost of the ceramic laser materials could be much lower than that of the single crystals because of the short period of fabrication process and mass production. In particular, no complicated facilities and critical techniques are required to grow ceramics. Driven by these advantages, excellent quality Nd³⁺-doped ceramic laser materials have been developed, which are a good alternative to the widely used Nd:YAG single crystals [5]. Besides the Nd³⁺-doped ceramic laser materials, the Yb³⁺-doped Y₂O₃ ceramic laser gain medium has also been intensively investigated as a novel diode-pumped solid-state laser material. Both the continuous wave (CW) emission and passively mode-locked emission of the laser material have been demonstrated recently [6–10]. With 11-W incident pump power, 0.7-W CW laser emission at 1078 nm has firstly achieved [6]. With improvement on the optical quality of the sample, in particular, through reducing the surface reflection losses and efficiently removing the heat generated in the ceramic, an output power up to 4.2 W on the low-order transverse mode emission was further achieved under a pump power of 19 W [7]. By reducing the doping concentration and sample length, laser emission around 1030 nm has also been demonstrated [8]. These results demonstrate that the Yb:Y₂O₃ ceramic has excellent optical and thermal properties, which is suitable for high power diode-pumped laser systems. Furthermore, benefiting from its wide emission spectral bandwidths around 1030 nm and 1076 nm, mode-locked operations with the pulse widths of 615-fs and 1-ps, mode-locked by a semiconductor saturable mirror (SESAM) and a GaAs single crystal wafer respectively, have been obtained [9–10]. In addition, it was noted that comparing with the Nd³⁺-doped laser media, the Yb:Y₂O₃ ceramic has an even longer upper-level lifetime and smaller emission cross section [9]. Therefore, the Yb:Y₂O₃ ceramic should also be suitable to be used in passively Q-switched laser systems.

The GaAs single crystal had been widely used as the saturable absorber for passively Q-switching or mode locking 1-µm lasers [10–12]. As GaAs has a band gap of 1.42 eV, which is much higher than the energy of the 1-µm photons, it is well accepted that the Q-switching operation in the laser is caused by the saturable single photon absorption (SPA) due to the EL2 defect located in the band gap, which yields a deep level 0.82 eV below the GaAs band edge [13]. So far the mechanism for the GaAs single crystal mode locking is not very clear, however, it is believed that the combined action of the saturable single-photon-absorption and the gain saturation could contribute to the process. In addition, the self-diffraction of light pulses due to the transient free-carrier grating in GaAs could play an important role in suppressing the pulse width [13]. For both Q-switching operation and mode-locking operation, two-photon absorption (TPA) in the semiconductor materials, corresponding to a transition of the carriers between the valence and conduction band, and free carrier absorption (FCA) contribute losses if sufficiently high intensities are presented. Recent investigations also showed that two-photon absorption in the semiconductor materials could limit the peak intensity of the mode-locked pulses, and lead to a continuous-wave mode locking (CWML) [14]. We emphasize that in the Yb:Y₂O₃ ceramic laser with a GaAs wafer simultaneously as saturable absorber and output coupler, both pure Q-switching and CW mode-locking can be obtained provided that appropriate cavity configuration is selected.

In this letter we report on an experimental realization of the passive Q-switching in a Yb:Y₂O₃ ceramic laser. By using a single crystal GaAs wafer as the saturable absorber as well as the output coupler, Q-switched pulses with a pulse energy of 7.7 µJ have been obtained. The emission wavelength in our laser is around 1076.5 nm. The average output power of the Q-switched pulses was measured to be 0.51 W with a 17.7-W incident pump power. The minimum pulse of 50 ns has been achieved with a repetition rate of 52.6 KHz.

2. Experimental setup

The laser setup used in our experiment is schematically shown in Fig. 1. The pump light at 940 nm from a fiber coupled laser diode bar was focused into the ceramic by two coupling lenses of 2 cm focal length. The focused pump beam in the laser medium had a diameter of about 450 µm. The Yb:Y₂O₃ ceramic sample has a Yb³⁺-doping concentration of 8 at. % and a dimension of 3×3 mm in cross-section and 3 mm in length. Both sides of the ceramic sample were AR coated in a wide band from 1030 nm to 1100 nm to decrease the optical loss. To efficiently remove the generated heat during the experiment, the sample was wrapped with indium foil and tightly mounted in a water-cooled copper holder. The temperature of the ceramic sample was set at as low as 7°C. A simple two-mirror-cavity configuration was employed in our laser. The input mirror (M1) was HR coated in a broad band from 1030 nm to at 1100 nm and AR coated at 940nm, it had a 1000 mm radius of curvature (ROC). The output coupler was a high purity GaAs wafer, which was <100> cut and had a cross section of 10 mm×20 mm. The thickness of GaAs was 450 µm. One side (close to the gain medium) of the GaAs wafer was anti-reflection coated at 1064 nm; the other side was coated with almost linearly variable transmission from 1.6% to 81.7% along the 20 mm direction around 1064 nm (standard dielectric coating). By properly translating the GaAs wafer along the 20 mm edge direction, different transmission of the output coupler could be obtained. To achieve short pulses, a short cavity length is favorable for passively Q-switched laser. In our experiment, the cavity length was kept as short as 20 mm, which was limited by the thickness of the crystal holder.

 

Fig. 1. Schematic of the laser setup.

Download Full Size | PDF

3. Experimental results

The average output power of the laser with different output couplings was firstly investigated. Figure 2 shows the results of the average output power as a function of incident pump power with two different output couplings. With a 5% output coupling, the average output power increases almost linearly with the incident pump power and no power saturation is observed. However, with 2% output coupling, the average output power almost keeps a constant when the incident pump power is above 14 W. The power saturation under a low output coupling could be caused by the increased losses due to two-photon absorption (TPA) and free carrier absorption (FCA) inside the GaAs as observed in reference [15]. At a maximum incident pump power of 17.7 W, an average output power of 510 mW was achieved in the fundamental transverse mode under the 5% output coupling. The beam quality was measured to be 1.6. The beam radius in GaAs wafer was estimated to be 220 µm. It’s worth to mention that in previous experiment we have achieved the CW mode locking operation in a Yb:Y₂O₃ ceramic laser with GaAs single crystal saturable absorber [10]. The main difference between these two experiments is that a folded cavity was used in the mode-locking experiment. Therefore, very small beam spot was achieved for efficient mode locking: the beam radius in GaAs was measured to be 60 µm, which significantly reduced the minimum intracavity pulse energy [16] required for CW mode locking operation. For mode-locked laser, with the pump power of 17.7 W, the maximum CW mode-locked output power obtained was 1.14 W with an output reflectivity R=94.0%, which is much higher than the average output power of the Q-switched pulses at the same pump power level. This result is expected because a much larger beam radius in GaAs in the Q-switching experiment had greatly enhanced the cavity loss of the laser.

 

Fig. 2. Average output power versus incident pump power.

Download Full Size | PDF

Figure 3 shows the pulse width as a function of the incident pump power. The pulse widths are around 200 ns at threshold pump power and decrease to 50 ns for the 2% output coupling and 75 ns for the 5% output coupling at maximum pump power, respectively. The corresponding peak powers at maximum pump power are 99 W and 102 W. It is also shown in Fig. 3 that a small peak exists in the pulse width variation at the incident pump power of 14 W. The mechanism for the slightly increase pulse width at this pump power level is not very clear yet. One possible explanation is the thermal lens effect of the gain medium causes a reduction of the beam radius in the saturable absorber. If the light intensity is still not strong enough and the losses caused by TPA and FCA are not visible, the total effect of the reduced beam radius in saturable absorber should be a reduction in the saturable loss, which broadens the pulse duration. We also found that by further increasing the pump power higher than 18 W, even shorter Q-switching pulses could be obtained. However, the pulse train became unstable and sub-pulses also appeared in this case. The instability of the Q-switched pulses at a high pump power level should be induced by the intrinsic nonlinear dynamics of the system such as the deterministic chaos [17]. When the incident pump power was further increased to 20 W, unstable Q-switched mode-locking pulses could be observed. We believe that stable mode-locking pulse could be achieved if the pump power is further increased. However, the pump power of 20 W has been close to the damage threshold of the sample. To avoid the pump-induced damage, we didn’t further increase the pump power. Nevertheless, it demonstrates that GaAs can be used for both Q-switching and mode locking in our laser if appropriate parameters are provided by selected cavity configurations.

 

Fig. 3. Pulse width versus incident pump power.

Download Full Size | PDF

A typical single pulse profile is shown in Fig. 4 with pulse duration of 50 ns. The repetition rate as a function of pump power is shown in Fig. 5. The repetition rate increases with the pump power when the incident pump power is slightly larger than the threshold. When the incident pump power is further increased, the repetition rate tends to slightly decrease both in the case of low and high output coupling. This phenomenon has been observed in several Q-switched laser systems with GaAs saturable absorber [18–19]. Li et al. had gave a preliminary explanation for this behavior of repetition rate, which could be attribute as a result of the increased threshold of pulse generation caused by high intracavity light intensity [18]. Under high laser intensity the increased losses due to TPA and FCA slow down the build up of the population inversion, which therefore decreased the repetition rate of the Q-switched pulse. For output coupler with low output coupling, a higher intracavity light intensity is easier to be achieved than that with high output coupling. Therefore, for the low output coupling, the turning point of the repetition rate should be firstly reached at a relative lower incident pump power as shown in Fig. 5. However, it’s surprising that under the high pump power level the repetition rate with low output coupling is lower than that with high output coupling. Whether it is caused by increased loss due to high light intensity needs to be further investigated. Figure 6 shows the pulse energy as a function of the incident pump power. The pulse energy keeps relatively stable at a relative lower pump level and slowly increases with the incident pump power at a high pump power level. The pulse energy with 5% output coupling is higher than that with 2% output coupling at the same pump level. With 5% output coupling the pulse energy reaches to the maximum value of 7.7 µJ at the maximum pump power.

 

Fig. 4. Oscilloscope trace of a typical single pulse profile.

Download Full Size | PDF

 

Fig. 5. Repetition rate versus incident pump power.

Download Full Size | PDF

 

Fig. 6. Pulse energy versus incident pump power.

Download Full Size | PDF

Figure 7 shows a typical oscilloscope trace of the Q-switched pulse trains. Small jitters in the peak power and repletion rate are observable in Fig. 7. The peak-to-peak intensity fluctuations and the inter-pulse time jittering are estimated to be <10% and <5% respectively. The instability of the laser may be attributed as a deterministic chaos of the laser as previously discussed. Another possible reason is related with the ceramic structure of the gain media. The uniformity of the dopant concentration may become poorer with the increase of doping concentration. Therefore, local heating of the active channel could induce visible instability of the Q-switching pulses.

 

Fig. 7. Oscilloscope trace of the Q-switched pulse train

Download Full Size | PDF

4. Conclusion

In conclusion, we have demonstrated a diode-end-pumped passively Q-switched Yb:Y₂O₃ ceramic laser with a GaAs wafer simultaneously as saturable absorber and output coupler. At a 17.7-W incident pump power, a 510-mW average output power has been achieved. The pulses width was measured to be about 50 ns with a repetition rate of 52.6 KHz. The maximum pulse energy was measured to be 7.7 µJ. To the best of our knowledge, this is the first demonstration on a passively Q-switched Yb:Y₂O₃ ceramic laser.