1. Introduction

Trivalent praseodymium (Pr³⁺) -doped materials hold potential for efficient, visible solid-state lasers because of their intense radiative transitions in the visible spectral range. Furthermore, they can easily be extended to ultraviolet (UV) lasers by a single step of second-harmonic generation (SHG), while UV lasers based on near infrared (NIR) lasers require at least two steps of frequency conversion. Moreover, Pr³⁺-based visible lasers have been attracting much attention because of the recent development of high-power indium gallium nitride (InGaN)-based blue laser diodes (LDs) that emit around ∼450 nm, which corresponds to the absorption peak wavelength of Pr³⁺-doped materials. Laser operations have already achieved with Pr³⁺-doped fluoride crystals [1–8], oxide crystals [9,10], and fluoride glass fibers [11]. In particular, Pr³⁺:LiYF₄ (Pr³⁺:YLF) laser operations have been obtained in the cyan (491 nm), green (523 and 546 nm), orange (604 and 607 nm), red (640, 670, and 698 nm), and deep red (721 nm) regions [1–7]. Output power of ∼5 W is currently available from a single-emitter InGaN blue LD [12]. We recently obtained the highest continuous wave (CW) output powers of 3.7 W at 607 nm and 6.7 W at 640 nm using four ∼5-W single-emitter LDs for pumping [7].

In addition to CW operation, investigations on visible pulse lasers are necessary to initiate new applications. In particular, passively Q-switching with saturable absorbers can generate ns laser pulses with a compact system. Therefore, many investigations on saturable absorbers in the visible region have been carried out. After Abe et al. [13] in 2013 first demonstrated passively Q-switched Pr³⁺:YLF lasers with a tetravalent chromium-doped Y₃Al₅O₁₂ (Cr⁴⁺:YAG), such 2D topological insulators as WS₂, MoS₂, and MoSe₂ [14], semiconductor saturable absorber mirrors [15,16], or graphene [17] were used as a saturable absorber in the visible region. The highest pulse energy of 11.3 µJ with a pulse width of 49 ns was achieved in a Pr³⁺:YLF laser at 640 nm with a Cr⁴⁺:YAG crystal, which has a high damage threshold and low saturation intensity [18]. One attractive function of such pulse operations in the visible region is efficient UV laser generation. UV pulses at 320 nm with a pulse energy of 1.5 µJ and a pulse width of 50 ns were obtained by the intracavity frequency doubling of a passively Q-switched Pr³⁺:YLF laser [19].

A divalent cobalt-doped MgAl₂O₄ spinel (Co²⁺:MALO) crystal has been recognized as a saturable absorber in the 1.5-µm region and used for the Q-switching of Er³⁺ lasers [20]. In 2017, Demesh et al. first utilized a Co²⁺:MgAl₂O₄ (MALO) as a saturable absorber in the visible region and demonstrated a passively Q-switched Pr³⁺:YLF laser at 523, 607, and 640 nm [21]. Note that Co²⁺:MALO can even be used as a saturable absorber for a green laser for which a Cr⁴⁺:YAG does not show saturable absorption. In our experiments, we demonstrated the power scaling of a passively Q-switched Pr³⁺:YLF laser at 523, 607, and 640 nm with a Co²⁺:MALO saturable absorber utilizing four ∼5-W blue InGaN LDs as pump sources. We also generated UV pulses at 261 and 320 nm by intracavity frequency doubling. To the best of our knowledge, the 261-nm pulse is the shortest wavelength laser directly generated from the intracavity second-harmonic generation of passively Q-switched lasers. Furthermore, we compared the output performance of the passively Q-switched pulse by a Co²⁺:MALO with that obtained by a Cr⁴⁺:YAG.

2. Passively Q-switched Pr³⁺:YLF laser at 523, 607, and 640 nm

We first performed passively Q-switched Pr³⁺:YLF laser operations at 523, 607, and 640 nm with a Co²⁺:MALO saturable absorber. We utilized two (111)-cut Co²⁺:MALO crystals (EKSMA Optics) whose absorption spectra were measured by a dual-beam absorption spectrometer (UV3600 Plus, Shimadzu). We experimentally estimated ground-state absorption (GSA) cross section ${\sigma _{gs}}$ and excited-state absorption (ESA) cross section ${\sigma _{es}}$ of the Co²⁺:MALO at 607 and 640 nm by Z-scan measurements [22] and calculated the saturated transmission of the Co²⁺:MALO. The calculated saturation parameters in the absorption at 523, 607, and 640 nm are summarized in Table 1. “Measured transmission” was obtained by the transmission of the saturable absorber measured by a spectrophotometer. The actual transmission was corrected with the Fresnel reflection. “Saturated transmission” was obtained by the ratio ${\sigma _{gs}}$/ ${\sigma _{es}}$.

Table 1. Saturation characteristics of Co²⁺:MALO at 523, 607, and 640 nm

View Table | View all tables in this article

The laser oscillator setup for a diode-pumped Pr³⁺:YLF laser is depicted in Fig. 1. The V-shaped cavity consisted of three mirrors: a plane dichroic mirror (DM1), a concave dichroic mirror of a 75-mm curvature radius (DM2), and a plane output coupler (OC). The OC transmissions were 2.7% at 523 nm, 11.4% at 607 nm, and 9.1% and 21.4% at 640 nm. The two dichroic mirrors have high transmission (T > 98.0%) at 444 nm and high reflectivity (R > 99.9%) from 520 to 650 nm. Only for the 607-nm operation, we used a different plane dichroic mirror with a lower transmission at the pump wavelength (T = 87.1%) for suppressing the lasing at 640 nm. As a gain medium, we employed a 12-mm-long 0.3 at.% Pr³⁺-doped YLF crystal rod (5-mm diameter) that was cut perpendicular to the crystal’s a-axis (AC Materials). The facets of the Pr³⁺:YLF crystal were both polished but uncoated. The crystal was mounted on a water-cooled copper heat sink.

 

Fig. 1. Experimental setup of a diode-pumped Pr³⁺:YLF laser Q-switched by a Co²⁺:MALO saturable absorber.

Download Full Size | PDF

As pump sources, we utilized four InGaN blue LDs (Nichia) at 444 and 442 nm, which correspond to an absorption peak of Pr³⁺:YLF for the π- and σ-polarization. The pump beams of two LD beams were combined by a polarizing beam splitter (PBS) and focused into a Pr³⁺:YLF crystal from both faces. The maximum incident and absorbed pump power were 14.9 and 10.9 W (for 607-nm operation, 14.7 and 10.2 W). The beam quality, ${M^2}$, of the pump beams was measured to be ∼2.0 × 20-25 (fast × slow axis). For the 442-nm LDs, we used optical isolators for blocking the unabsorbed pump beam from the other side because the absorption coefficient of Pr³⁺:YLF at 442 nm for σ-polarization is less than half of that for the π-polarization. In fact, ∼1.5 W of the transmitted pump power was directed to the LD of the other side.

As shown in Fig. 1, the Co²⁺:MALO was placed in the vicinity of the OC. The cavity lengths were 144, 131, and 131 mm at 523, 607, and 640 nm when they were optimized for output power. The cavity mode size for each operating wavelength was calculated by a ABCD matrix method for the fundamental spatial mode, and the calculated mode radius in the Pr³⁺:YLF was ∼60 × 60 µm², ∼95 × 95 µm², and ∼100 × 100 µm² (vertical × horizontal), and the mode radius in the Co²⁺:MALO was ∼50 × 50 µm², ∼60 × 60 µm², and ∼60 × 60 µm² at 523, 607, and 640 nm. Since Pr³⁺:YLF has a smaller stimulated emission cross section at 523 nm (0.3 × 10⁻¹⁹ cm²) than 607 nm (1.4 × 10⁻¹⁹ cm²) and 640 nm (2.2 × 10⁻¹⁹ cm²) [1], the output performance at the wavelength was greatly sensitive to a slight change in the cavity loss induced by aberrated thermal lensing (fourth- and sixth- (and even higher) order wavefront distortion due to thermally induced refractive index profile) in the crystals [23]. Furthermore, the Pr³⁺:YLF crystal exhibits strong negative thermal lensing for π-polarized 523-nm laser emission. Consequently, the laser output power was maximized when the mode radius in the Pr³⁺:YLF at 523 nm was smaller than at 607 and 640 nm, due to the requirement for thermal lens suppression [7].

Before the Q-switching experiment, we performed a CW operation by removing the Co²⁺:MALO from the cavity. At 640 nm, the threshold pump power was 619 mW and 768 mW, the slope efficiency was 50.9% and 40.0%, and the maximum output power was 5.3 W and 4.1 W at T_OC = 9.1% and T_OC = 21.4%, respectively. In the other laser transitions, the threshold pump power was 871 mW and 826 mW, the slope efficiency was 34.1% and 23.8%, and the maximum output power was 3.1 W and 1.1 W at 607 nm and 523 nm, respectively.

Next, we inserted Co²⁺:MALO into the cavity and performed passively Q-switching. The measured average power, pulse width, and repetition frequency as a function of absorbed pump power are shown in Figs. 2–4. Using the OC of T_OC = 11.4 (21.4)%, the highest peak power of 0.76 and 1.08 kW was obtained at the maximum absorbed pump power with an SA1 (see in Table 1) at 607 and 640 nm. The width of the passively Q-switched pulses can generally be shortened by employing lower initial transmissions of the saturable absorber, resulting in larger modulation depths. The pulse width can also be shortened by adopting higher transmission OCs, since for most Q-switched lasers the pulse width and the pulse energy are not significantly changed from those around the lasing threshold because the pulse width becomes shorter due to the larger modulation depth of the SA1 transmission. In fact, when we used SA1 as a saturable absorber and an OC whose transmission is 9.1% at 640 nm for an absorbed pump power of >10 W, a pulse with a much higher peak power was generated; unfortunately, it damaged the plane dichroic mirror (DM1).

 

Fig. 2. (a) Average power, (b) pulse width, and (c) repetition frequency of passively Q-switched Pr³⁺:YLF laser with Co²⁺:MALO (SA2) and output coupler of 9.1% at 640 nm as a function of absorbed pump power. Dashed lines are results of numerical simulation.

Download Full Size | PDF

 

Fig. 3. (a) Average power, (b) pulse width, and (c) repetition frequency of passively Q-switched Pr³⁺:YLF laser with Co²⁺:MALO (SA2) and output coupler of 11.4% at 607 nm as a function of absorbed pump power. Dashed lines are results of numerical simulations with two different ESA cross sections.

Download Full Size | PDF

 

Fig. 4. (a) Average power, (b) pulse width, and (c) repetition frequency of passively Q-switched Pr³⁺:YLF laser with Co²⁺:MALO (SA2) and output coupler of 2.7% at 523 nm as a function of absorbed pump power.

Download Full Size | PDF

We also performed passively Q-switching at 523 nm. Co²⁺:MALO is advantageous since it enables Q-switched green lasers and Cr⁴⁺:YAG does not. At this wavelength, we failed to reach the lasing threshold with SA1 because of its relatively low initial transmission and the small stimulated emission cross section of Pr³⁺:YLF compared to those at 607 and 640 nm. Using the SA2 with higher initial transmission, we successfully generated green Q-switched pulses. However, the plane dichromic mirror (DM1) was damaged by the intracavity green pulses at an absorbed pump power of ∼5.5 W. Figure 4 shows the performance of the green Q-switched laser up to a 5.3-W absorbed pump power that is slightly below the damage threshold of the DM1. The maximum pulse energy of 6.6 µJ and the maximum peak power of 27 W obtained at 523 nm was mainly much lower than those of the other two wavelengths for three reasons. First, the small mode size in the Pr³⁺:YLF crystal led to a relatively low mode matching efficiency of ∼50%. Second, the insertion loss of the Co²⁺:MALO significantly reduced the output performance due to the small stimulated emission cross section at 523 nm. Third, the absorption of the Co²⁺:MALO could not be quickly bleached due to low intracavity laser intensity. The pulse width was inevitably longer at 523 nm because of the small modulation depth of the transmission in the Co²⁺:MALO at 523 nm. Even though we have not yet fully examined the Co²⁺:MALO transmission at 523 nm when it was fully bleached, the measured initial transmission of 97.5% indicates that the maximum obtainable modulation depth is not higher than 2.5%.

Table 2 shows the summary of the passively Q-switched Pr³⁺:YLF laser with Co²⁺:MALO at 523, 607, and 640 nm. In our experiment, we measured ${M^2}$ only at the highest pump laser power of 11 W. In our cavity design, the fundamental cavity mode size (∼100 µm in radius) is always larger than that of the pump laser beam size. Therefore, the thermal aberration induced by the pump laser always degrades the ${M^2}$ value and the degradation increases as the pump laser power increases. In the worst case, the higher cavity mode may start to oscillate and increase the laser mode size in the laser crystal. We did not observe such change in the beam size in our experiment up to the pump laser power of 11 W since the heat-induced lens effect of an YLF crystal is small. The beam quality for the 640-nm laser is better than other wavelengths since we used a higher T_OC mirror (>20%) and the saturable absorber with a lower initial transmission, where the laser threshold became higher also for higher-order cavity modes.

Table 2. Summary of characteristics of passively Q-switched Pr³⁺:YLF laser with Co²⁺:MALO at 523, 607, and 640 nm

View Table | View all tables in this article

We further scrutinized the passively Q-switched lasers by comparing the experimental results with numerical simulations based on a rate equation model. Passively Q-switched lasers can be expressed by the following three equations for a spatially zero-dimensional model:

These are the rate equations of the three parameters: intracavity photon number density $\phi $, the population inversion density in gain medium $\Delta N$, and the ground-state population density of saturable absorber ${n_{gs}}$. The other parameters and values for the calculation are summarized in Table 3. Since the ground-state population in gain medium varies only within 2%, we treated the pumping rate ${P_{abs}}$ as a constant value.

For reproducing the experiments by a numerical model, we first determined an unknown laser parameter: the intracavity loss at the zero-dimensional gain medium model so that the calculated results of the CW laser showed good agreement with those of our experiments. At an output coupling T_OC of 21.4%, we increased the intracavity loss, which is presumably increased by the loss due to a higher upper state density during lasing [2,7], to 6.5% from 0.3% at T_OC= 9.1%, as shown in the Table 3. We need more detailed study to clarify the increase in the residual absorption at high T_OC. Anyway, with these adjusted intracavity losses, we reproduced the output laser performance for the CW laser both in the lasing threshold and the slope efficiency by the numerical mode.

Table 3. Parameters and values used in rate equations of passively Q-switched Pr³⁺:YLF laser with Co²⁺:MALO saturable absorber

View Table | View all tables in this article

The results of the numerical simulation are shown as dashed lines in Figs. 2 and 3 for the Q-switched lasers. In this numerical analysis, with the adjusting parameter selections described above, we obtained fairly good agreement in average power and pulse width using the GSA and ESA cross-section values (Table 1). However, the calculated repetition frequencies were always significantly different from those obtained in the experiments. In the experiment, the repetition frequency saturated when the absorbed power exceeded ∼5.5 W, while the rate equation model always exhibited a linear increase of the repetition frequency with respect to the absorbed pump power. This trend was also obvious in the previous report on a passively Q-switched Pr³⁺:YLF laser pumped by a frequency-doubled optically pumped semiconductor laser (2ω-OPSL) [21]. To the best of our knowledge, Nd³⁺:YAG/Cr⁴⁺:YAG Q-switched lasers do not show such a ceiling in repetition frequency. Note also that the Pr³⁺:YLF lasers Q-switched by Cr⁴⁺:YAG also exhibited a similar tendency [24]. We speculate that the charge transfer in transition metal ions by intense Q-switched pulse at the laser wavelength of 639 nm causes some nonlinear process which our rate-equation model does not consider. Further detailed investigations are necessary to fully understand the laser performance.

At 607-nm operation, the ground-state reabsorption (³H₄ →¹D₂) is involved, resulting in a high intracavity loss. Therefore, the maximum pulse energy at 607 nm decreases and the optimum T_OC must be higher than that at 640 nm. The predicted pulse width by the model using the ESA cross section estimated by a Z-scan measurement [22] in Table 1 was ∼25 ns, which is much shorter than in the experiments. This discrepancy indicates that the experimentally estimated ESA cross section was overestimated. The calculated pulse width showed good agreement with the experimental results when the ESA cross section was set in a range of 9.5–10.5×10⁻¹⁹ cm².

3. UV pulse generation at 261 and 320 nm by intracavity frequency doubling of passively Q-switched Pr³⁺:YLF laser

We extended the Q-switched visible lasers to UV-pulsed lasers at 261 and 320 nm by intracavity frequency doubling. The experimental setup is depicted in Fig. 5. The Z-shaped cavity consisted of four mirrors: a plane dichroic mirror (DM1), a concave dichroic mirror with a 75-mm curvature radius (DM2), a concave output coupler with a 75-mm curvature radius (OC), and a plane high reflection mirror (HR). The dichroic mirrors were identical as those shown in Fig. 1. The OC had a transmission of 78.1% at 261 nm and 83.7% at 320 nm as well as a reflectivity of >99.8% at fundamental wavelengths. The HR had a high reflectivity of >99.9% at both the fundamental and second-harmonic wavelengths. The Pr³⁺:YLF crystal and the pump system are the same as those in the experiments at the fundamental wavelength, and we used the Co²⁺:MALO (SA2) as a saturable absorber. As SHG nonlinear crystals (NLCs), we used a 7.0-mm-long Type-I β-BaB₂O₄ (BBO) crystal (θ = 48.9°) to convert the 523-nm light into 261 nm and an 8.0-mm-long Type-I LiB₃O₅ (LBO) crystal (θ = 90°, φ = 53.6°) to convert the 640-nm light into 320 nm. The Pr³⁺:YLF, the Co²⁺:MALO, and the NLC were placed at the cavity’s three focal points. For the lasers at 523 and 640 nm, the calculated mode radii were ∼60 × 55 µm² (horizontal × vertical) and ∼115 × 95 µm² in the Pr³⁺:YLF, ∼45 × 40 µm² and ∼50 × 50 µm² in the Co²⁺:MALO, and ∼110 × 85 µm² and ∼65 × 65 µm² in the NLC. The cavity length was ∼300 mm for both wavelengths.

 

Fig. 5. Experimental setup of intracavity frequency doubling a diode-pumped Pr³⁺:YLF laser passively Q-switched by a Co²⁺:MALO saturable absorber.

Download Full Size | PDF

We established a rate equation model to evaluate the laser performance at 320 nm. We only applied it to the laser at 640 nm since the absorption cross sections of Co²⁺:MALO at 523 nm were unknown. The time evolution of intracavity photon number density $\phi $ is described by the following equation [19]:

Table 4. Parameters and values used in rate equations of intracavity frequency doubling of passively Q-switched Pr³⁺:YLF laser with Co²⁺:MALO saturable absorber

View Table | View all tables in this article

The measured average power, pulse width, and repetition frequency as a function of the absorbed pump power are shown in Figs. 6 and 7. The characteristics of the Q-switched pulses are summarized in Table 5. At 320 nm, the average power showed a linear increase with respect to the absorbed pump power. The pulse width was longer than the experimental result in a previous work [19] that used a Cr⁴⁺:YAG saturable absorber. The output pulse width was 50 ns, even though the transmission modulation depth of the saturable absorber was almost the same because we used a Z-cavity in this experiment and its cavity length was longer than that of a V-cavity in a previous experiment [19]. In the numerical analysis, we adjusted the intracavity loss to obtain better agreement of the SHG performance. We assumed intracavity loss of 6.0%, which corresponds to the transmission loss at HR and OC for the 640-nm light.

 

Fig. 6. (a) Average power, (b) pulse width, and (c) repetition frequency of UV pulse at 320 nm as a function of absorbed pump power. Dashed lines are results of numerical simulation.

Download Full Size | PDF

 

Fig. 7. (a) Average power, (b) pulse width, and (c) repetition frequency of UV pulse at 261 nm as a function of absorbed pump power.

Download Full Size | PDF

Table 5. Summary of characteristics of UV pulses at 261 and 320 nm

View Table | View all tables in this article

Figure 7 depicts the output characteristics of the Q-switched pulse at 261 nm. To the best of our knowledge, this is the first pulsed DUV laser at 261 nm from a passively Q-switched Pr³⁺:YLF laser at 261 nm. As shown in Fig. 7 and Table 5, both the pulse energy and the peak power of the 261-nm pulse were much lower than those of 320 nm since the Q-switched pulses at 523 nm before frequency doubling (Fig. 4) had a much lower pulse energy and a lower peak power, which significantly lowered the conversion efficiency.

4. Comparison of passively Q-switched Pr³⁺:YLF laser with Cr⁴⁺:YAG

Finally, we compared the characteristics of the Co²⁺:MALO and Cr⁴⁺:YAG saturable absorbers in the visible region. Q-switching operations with Co²⁺:MALO and Cr⁴⁺:YAG was compared again with the setup in Fig. 1 at 640 nm. In this experiment, we utilized a 1.3-mm long Cr⁴⁺:YAG crystal with the initial transmission of 94.6% and the saturated transmission of 96.3%. By utilizing OC of 21.4%, we obtained 60.9-ns pulses with a highest pulse energy of 14.3 µJ and a repetition rate of 126.8 kHz. Therefore, within the various conditions we used for the experiments, Pr³⁺:YLF lasers Q-switched by the Co²⁺:MALO produced higher pulse energy, higher peak power, and higher average power than those obtained by the Cr⁴⁺:YAG Q-switched Pr³⁺:YLF laser. However, output performance is always varied by the doping concentration and the length of a saturable absorber, and the cavity design such as a spot size at the saturable absorber, while the saturation characteristics are determined by GSA cross section and recovery time of the saturable absorber.

To compare the saturation characteristics between two saturable absorbers, we observed a change in the Q-switched pulse train by altering the position of the saturable absorber in the cavity, which accordingly changed the intensity (or fluence) at the saturable absorbers. To achieve passive Q-switching, a saturable absorber needs to be saturated before the laser gain; otherwise the laser operates in CW. The saturation intensity of Pr³⁺:YLF at 640 nm is ∼37 kW/cm², and that of Co²⁺:MALO is estimated to be ∼560 kW/cm² using a GSA cross section of 11×10⁻¹⁹ cm² and a recovery time of 500 ns [22]. Based on these values, the mode radius in Co²⁺:MALO (${w_{SA}}$) needs to be four-times smaller than that in Pr³⁺:YLF (∼100 × 100 µm²). Thus, ${w_{SA}}$ may need to be smaller than ∼25 µm for stable Q-switching. Experimentally obtained pulse trains for different conditions are presented in Fig. 8. We confirmed that the Q-switch operation can be stabilized by adjusting the mode diameter in the saturable absorber. Passively Q-switching was practically obtained even though ${w_{SA}}$ was ∼60 µm. If the Co²⁺:MALO was shifted from the focal point, the Q-switching became unstable. Based on the calculated cavity mode radius, we conclude that the mode radius in Co²⁺:MALO needs to be smaller than approximately 0.6 times of that in Pr³⁺:YLF to obtain stable Q-switching.

 

Fig. 8. Pulse train of passively Q-switched Pr³⁺:YLF laser with Co²⁺:MALO (Sample 2) (a) OC-SA : 3 mm, ${w_{SA}}$: 63×60 µm² (b) OC-SA : 6 mm, ${w_{SA}}$: 65×61 µm² (c) OC-SA : 9 mm, ${w_{SA}}$: 69×64 µm²

Download Full Size | PDF

Cr⁴⁺:YAG’s saturation intensity was estimated to be ∼10 kW/cm² using a GSA cross section of 7.5×10⁻¹⁸ cm² and a recovery time of 4 µs [22]. Therefore, Cr⁴⁺:YAG has a lower saturation intensity than Pr³⁺:YLF gain, and Q-switching can theoretically be obtained even if ${w_{SA}}$ is twice as large as that in Pr³⁺:YLF. In the experiment, Q-switching became unstable when ${w_{SA}}$ (∼120 µm) almost equaled that in Pr³⁺:YLF (Fig. 9(b)). Thus, the mode radius in Cr⁴⁺:YAG needs to be smaller than that in Pr³⁺:YLF for stable Q-switching.

 

Fig. 9. Pulse train of passively Q-switched Pr³⁺:YLF laser with Cr⁴⁺:YAG (a) OC-SA : 3 mm, ${w_{SA}}$: 63×06 µm² (b) OC-SA : 30 mm, ${w_{SA}}$: 115×112 µm² (c) OC-SA : 40 mm, ${w_{SA}}$: 143×142 µm²

Download Full Size | PDF

These results indicate that using a Cr⁴⁺:YAG saturable absorber with a relatively low saturation intensity is suitable for a microchip laser, in which we cannot focus the beam into the saturable absorber. However, in the Cr⁴⁺:YAG crystal, the remaining Cr³⁺ in the octahedral site absorbs light in the orange-red region and does not contribute to the saturable absorption. This issue only takes place for the visible absorption, while the absorption around 1µm is purely due to the Cr⁴⁺ in the tetrahedral site. Since the amount of such a parasitic absorption center depends on the crystal growth conditions, the absorption spectra of Cr⁴⁺:YAG differ from sample to sample [22]. The optimization of growth processes, including the post-annealing process for controlling the valence of Cr ions, is required for improving the Cr⁴⁺:YAG crystal as a saturable absorber in the visible region. On the other hand, since Co²⁺:MALO has a higher saturation intensity, we have to place it in the focal point to obtain Q-switching, which is not useful to achieve a microchip laser with Co²⁺:MALO. However, Co²⁺:MALO can work as a saturable absorber in a wider spectral range than Cr⁴⁺:YAG. In particular, Q-switching in the green range enables DUV pulse generation by a single step of frequency conversion. More detailed scrutiny of the saturation characteristics in this range is required to estimate the performance of the Q-switched pulses, which can be applied to a demonstration of Q-switching with other gain mediums, such as terbium (Tb³⁺)-doped materials [25].

5. Conclusion

We reported the power scaling of passively Q-switched Pr³⁺:YLF lasers at 523, 607, and 640 nm with a Co²⁺:MALO saturable absorber pumped by four ∼5-W blue InGaN laser diodes. At 640 nm, we obtained the highest pulse energy of 33.5 µJ and the highest peak power of 1.08 kW with a pulse width of 30.9 ns and a repetition frequency of 64.0 kHz. In the other spectral ranges, the maximum pulse energies were 6.6 and 30.7 µJ at 523 and 607 nm. We also demonstrated UV pulse generation by intracavity frequency doubling. The maximum pulse energies were 0.2 and 7.0 µJ at 261 and 320 nm. Moreover, we demonstrated numerical analysis based on the rate equation and confirmed good agreement with experimental results of average power and pulse width at 320, 607, and 640 nm. Further power scaling can be achieved by increasing the pump power with strategies to overcome thermal fracture by employing a pump system that leads to lower pump density or a Pr³⁺-doped oxide. However, unlike such other gain media as Nd³⁺-doped crystal, a shorter fluorescence lifetime of Pr³⁺-doped crystal is the main issue for the peak power scaling of Q-switched pulses.