1. Introduction

During the past decades, diode-pumped solid-state lasers (DPSSLs) have developed greatly, especially for DPSSLs operating in near infrared wavelength range from about 0.9 μm to 2.0 μm. Although Yb³⁺, Er³⁺ and Tm³⁺ doped lasers can provide efficient light sources at 1.0 μm, 1.6 μm and 1.9 μm, respectively, only Nd³⁺ doped lasers can provide all of these laser sources and even beyond. In fact, to date, Nd³⁺ doped lasers have already been demonstrated successfully at 0.9, 1.06, 1.1, 1.3 and 1.4 μm corresponding to emission transitions from upper level of ⁴F_3/2 to lower levels of ⁴I_9/2, ⁴I_11/2 and ⁴I_13/2, respectively [1–6].

Apart from these widely studied lasing wavelengths, Nd³⁺ lasers also provide another possibility, i.e. lasing at about 1.8 μm, corresponding to ⁴F_3/2 to ⁴I_15/2 transition. This emission channel has been far less studied because of its small emission cross section (3 × 10⁻²¹ cm² for Nd:YAG [7]) in comparison to other lines. However, in fact, for Nd:YAG this value is indeed comparable to the emission cross sections of many Tm³⁺-doped materials, even higher. For instance, the emission cross sections of Tm:YAG, Tm:GGG, Tm:LuAG and Tm:YLF are 2.2 × 10⁻²¹ cm² [8], 1.05 × 10⁻²¹ cm² [9], 1.87 × 10⁻²¹ cm² [10] and 3.7 × 10⁻²¹ cm² [8], respectively. As we know, Tm³⁺ lasers are the main provider of 1.9 μm sources. Although the lasing wavelength of Tm³⁺ fiber laser could extend to about 1.7 μm via wavelength selective feedback [11], the lasing wavelengths for solid-state laser with Tm³⁺ crystals and/or ceramics have still been limited to longer than about 1.9 μm because of the reabsorption losses despite broad emission spectra extending to shorter than 1.8 μm for many Tm³⁺ laser materials. The 1.8-μm Nd³⁺ lasers belong to four-level system and therefore no reabsorption effect exists. Moreover, in medicine, choice of laser sources with wavelengths shorter than 1.9 μm could be particularly needed for shallow water absorption of biological tissues [12]. In addition, it is worthwhile to mention that apart from the aforementioned various applications the high-power 1834 nm Nd:YAG laser source could be much more efficient to pump Cr²⁺:ZnSe than currently used 1.9-μm Tm³⁺ lasers [13] for the realizations of tunable and ultrafast middle infrared lasers.

Development towards the realization of high-power 1.8-μm Nd³⁺ laser sources has been far limited in the past years. The first laser oscillation at this line was demonstrated in 1971 by using a Kr flash lamp pumping with Nd:YAG cooling down to −40 °C [14]. In 1992, Kubo and Kane [15] operated a diode-pumped 3-mW 1833-nm Nd:YAG laser and the crystal was cooled to −28°C. Much better laser performance was obtained later by operating Nd:YAG waveguides for improving this laser emission to 0.4 W [16] and 1.4 W [17] by the same group. In 2016, we reported a 1834-nm Nd:YAG bulk laser with a maximum output power of 1.31 W [18]. However, at present, no higher power than 1.4 W has been achieved at this specific laser line and pulsed laser operation at this line has not yet been reported.

For further improving the 1.8-μm laser output power, simply increasing the pump power is a primary consideration. However, quantum defect for 808-nm diode pumped 1.8-μm Nd³⁺ lasers is about 56%, which is a very serious negative factor in restricting the laser output power improvement because of thermal lensing effect. From this point of view, reducing the quantum defect by direct (or in-band) pumping should be very advantageous. In fact, the so-called direct pumping has indeed gained great successes in Nd³⁺ lasers [19–23]. In this work, using a high-power 885-nm diode laser as pump source, we have realized high-power and high-energy Nd:YAG lasers at the rarely investigated 1.8-μm emission band in continuous-wave and passively Q-switched regimes.

2. Experimental setup

The experimental setup is shown schematically in Fig. 1. The pump source is a fiber-coupled diode laser emitting at 885 nm at maximum output power of about 50 W. The coupling fiber has a core diameter of 400 μm and numerical aperture of 0.22. The pump beam was injected into the Nd:YAG laser crystal by two doublet lenses both with focal lengths of 75 mm. The Nd:YAG crystal acting as gain medium has a dimensions of 3 × 3 × 10 (mm³) and a nominal dopant concentration of 1at.%. To protect the laser crystal from thermal fracture, we wrapped the crystal with indium foil and then mounted it inside a copper block connecting to water-cooled chiller with temperature set at 8°C. In the previous investigation, we have found that an output power variation of about 0.36%/mm·K for Nd:YAG crystal [18]. The present temperature value is a tradeoff between high output power generation and avoiding moisture.

 

Fig. 1 Laser experimental setups of 885-nm diode-pumped Nd:YAG bulk lasers in (a) continuous-wave and (b) passively Q-switched regimes.

Download Full Size | PDF

Two kinds of laser resonators were used during the laser experiments for different purposes. For continuous-wave laser operation, a simple two-mirror plane-concave cavity was configured. The used flat mirror (M1) has a transmission of more than 94% at pumping wavelength and high reflection of more than 99.9% at laser wavelength. Two mirrors (M2) both with curvature radii of 50 mm were used for output couplers (OCs). Transmissions of the two mirrors are 5.82% (OC1) and 4.06% (OC2) at laser wavelength. For passively Q-switched laser operation, we arranged a three-mirror V-shaped cavity aiming at avoiding thermally induced degradation of the laser performance considering 32% of residual pump power. The mirror M3 has a curvature radius of 50 mm with high transmission at pumping wavelength and high reflection of more than 99.9% at laser wavelength. Two M2, again, were used as OCs here. The used saturable absorber was a Cr:ZnSe crystal with initial transmission of about 96% at laser wavelength.

3. Results and discussion

The 885-nm absorption corresponds to transition from ⁴I_9/2 to ⁴F_3/2, i.e. pumping directly to the laser upper level. The absorption cross section of Nd:YAG crystal at 885 nm is about 1 × 10⁻²⁰ cm². Single-pass absorption ratio of the pump power was measured to be about 68%. Taking the crystal length into account, the effective absorption coefficient of the Nd:YAG crystal is simply estimated to be about 1.14 cm⁻¹ at this specific wavelength. With these data, we can further deduce that the real dopant concentration of the used Nd:YAG crystal is about 0.83at.%. In addition, emission cross section for the 1834-nm line of Nd:YAG crystal is about 3 × 10⁻²¹ cm² [7], while the absorption cross section of Cr:ZnSe crystal is about 85 × 10⁻²⁰ cm² [24]. The latter has far larger emission cross section than the former, which is favorable for passively Q-switched operation.

Figure 2 shows the laser output power varying with the increase of absorbed power in continuous-wave mode. The laser cavity length was optimized to be about 46 mm for the best laser performance with maximum output power. Under this situation, a maximum output power of 2.72 W was achieved using OC1 and the threshold was about 4.19 W of absorbed power, which leads to a slope efficiency of about 12.1%. The output power is more than two times of our previous result achieved with an 808-nm diode pump source. The thermal mitigation originating from relatively low quantum defect which allows pumping the Nd:YAG at high level is of benefit to the power scaling. Moreover, slope efficiency achieved in this work is also higher than previous one, which has further verified the advantage of in-band pumping. Using OC2, the maximum output power reduced to 1.89 W, while the threshold was about 3.08 W. The corresponding slope efficiency was linearly fitted to be about 8.2%. Comparing with our previous results [18], the present efficiency improvement is not remarkable, which could be explained by worse overlap efficiency between pump beam and cavity mode with a thickness doubled crystal. Moreover, the present crystal has a higher dopant concentration to ensure sufficient absorption of pump power, which could lead to quenching effect that affects the laser performance.

 

Fig. 2 The laser output power varying with absorbed power of 1834-nm continuous-wave Nd:YAG laser; inset: typical laser spectrum and laser beam spots at maximum output powers.

Download Full Size | PDF

With these data, we estimate intracavity loss of the four-level Nd:YAG laser based on Findlay-Clay analysis. The corresponding formula can be written as $P_{th1,abs}/P_{th2,abs}=\left(L-\ln R_1\right)/\left(L-\ln R_2\right)$, where P_thi,abs (i = 1,2) is the threshold power, L is the intracavity loss and R_i is the reflection of the OC. The intracavity loss can be deduced to be about 0.99%, which indicates a good optical crystal of the laser crystal. It should be pointed out that thermally induced power saturations have still been observed in Fig. 2 when the absorbed power increased to over 26 W. We consider further power scaling by adopting a composite Nd:YAG crystal with two undoped YAG crystals at the ends. The inset shown in Fig. 2 gives a typical spectrum of the continuous-wave laser with peak at about 1834.46 nm. The insets in Fig. 2 show the beam spots of the output lasers at maximum output powers. Further, the beam qualities characterized by M² factor were measured to be about 3.3 and 3.2 for the 2.72 W and 1.89 W lasers, which are acceptable for such high absorbed powers.

In the case of passively Q-switched Nd:YAG lasers, the laser cavity was configured to be about 85 mm in total. The M3 mirror was orientated to form a small folded angle of about 15°. According to the standard ABCD matrix analysis, with this angle as small as 15°, the astigmatism induced by the folded curvature mirror is negligible. Moreover, under this situation, we found the smallest cavity mode size (in radius) inside the Nd:YAG to vary slightly from about 130 to 145 μm with the increase of the thermal lensing effect. Furthermore, the smallest cavity mode size in the Cr:ZnSe saturable absorber is estimated to be about 98 μm. The relatively small beam size inside the saturable absorber is favorable for blenching. Figure 3 shows the passively Q-switched results. Using OC1, we obtained a maximum average output power up to 1.25 W and the slope efficiency is about 5.7%. Using OC2, the corresponding maximum output power and slope efficiency are 0.89 W and 3.7%. The inset shown in Fig. 3 gives the lasing wavelength of the passively Q-switched lasers with peak at 1834.25 nm. Under both cases, the power stabilities were measured to be better than about 1.8% (rms) in 20 minutes.

 

Fig. 3 The laser output power varying with absorbed power of 1834-nm passively Q-switched Nd:YAG laser; inset: typical laser spectrum.

Download Full Size | PDF

Figure 4 shows the typical single pulse profiles at maximum output power with insets of the corresponding pulse trains. The achieved shortest pulse time duration is about 54.6 ns at a pulse repetition rate of 10.09 kHz with OC1. The pulse time duration increased to about 59.8 ns with a pulse repetition rate of 8.66 kHz using OC2. The intensity instabilities of the pulse train at maximum output powers were less than about 9.8% for both cases and the pulse-to-pulse timing jitters were less than about 6.7%, i.e. about 4 ns. Correspondingly, repetition rate of the pulse trains was found to have instability of about 4.3%.

 

Fig. 4 Single pulse profile of the Q-switched 1834 nm laser using (a) OC1 and (b) OC2 with inset showing the corresponding pulse trains.

Download Full Size | PDF

Figure 5(a) and 5(b) show the pulse width and pulse repetition rate varying with the increase of the absorbed powers. At threshold, the pulse widths achieved with OC1 and OC2 are 95.5 ns and 110.6 ns, respectively. Correspondingly, the pulse repetition rates are 1.06 kHz and 0.87 kHz. With the increasing of the absorbed power, the pulse widths for the two cases decreased monotonously until the absorbed power reached 20 W. Above this absorbed power, pulse widths showed saturations with the shortest pulse widths of about 54.2 ns and 58.8 ns. At the same time, the pure repetition rate increased almost linearly. With these data, we calculated the pulse energy and pulse peak power, as shown in Figs. 5(c) and 5(d), increasing from 35.8 μJ to 125.9 μJ (maximum achieved in this work) for OC1 and from 14.9 μJ to 108.3 μJ for OC2. The pulse peak powers correspondingly increase from 0.37 kW to 2.27 kW for OC1 and from 0.13 kW to 1.81 kW for OC2. The pulse energies and peak powers all showed saturations mainly arising from the pulse width saturations.

 

Fig. 5 The variations of (a) pulse width, (b) pulse repetition rate, (c) pulse energy and (d) pulse peak power with the increase of absorbed power for OC1 and OC2 cases.

Download Full Size | PDF

Comparing the present results with some Tm³⁺-doped lasers reported in literature, we have found that the Nd:YAG bulk laser at 1834 nm is in fact quite competitive in output power and pulse energy. For instance, Zhang et al. reported a Tm,Y:CaF₂ laser at 1969 nm with average output power of 400 mW and pulse energy of 21.7µJ [25]. 3.2-W average output power and 70.5-μJ pulse energy were also obtained in Tm,Ho:GdVO₄ laser at 2.05 μm [26]. We, previously, operated a Tm:CaGdAlO₄ laser with 0.37-W average output power and 37.4-μJ pulse energy by specifically coating the cavity mirrors for supporting lasing at wavelength blue shift to 1854 nm [27]. Better performance with mJ-scale pulse energy of the passively Q-switched Nd:YAG bulk laser at 1834 nm could be expected by adopting a low-transmission Cr:ZnSe saturable absorber, like reported in [28] (average output power only 98 mW, pulse width 14 ns and repetition rate 120 Hz), for further narrowing the pulse width and reducing the repetition rate.

4. Conclusions

To summarize, we have operated continuous-wave and passively Q-switched Nd:YAG bulk lasers at 1834 nm under an in-band diode laser pumping at 885 nm. We have obtained a maximum output power up to 2.72 W at 1834 nm and the corresponding slope efficiency is about 12.1% with respect to the absorbed power. Passively Q-switched operation is also demonstrated with maximum average output power of 1.25 W. The achieved shortest pulse width is about 54 ns with maximum pulse energy and peak power of 125.9 μJ and 2.27 kW. Finally, it should be pointed out that Tm³⁺ and even Ho³⁺ lasers emerge in mind generally when dealing with this specific laser emission band. However, they cannot be efficiently shifted to wavelength less than 1.85 μm. Hence, the 1834-nm Nd³⁺ laser bridges this wavelength gap and it could have potential applications in shallow water absorption of biological tissues.