1 Introduction

Stimulated Raman scattering (SRS) in crystal materials has been shown to be an efficient method to extending the laser wavelengths from the ultraviolet to the near infrared. Self-Raman lasers, in which the laser crystal acts as the Raman-active medium simultaneously, can provide multiple advantages of small scale, low loss and high efficiency. Nowadays, the materials most commonly used in Raman lasers include YVO4, GdVO4, LiNbO3, Ba(NO3)2, BaWO4, KGd(WO4)2 and SrWO4. Of the Raman-active media, the neodymium-doped yttrium orthovanadate crystal (Nd:YVO4) was considered as a wonderful Raman medium due to its high Raman gain (4.5 cm/GW) [1] and large emission cross section.

In 2001, Kaminskii et al. [1] predicted that Nd:YVO4 and Nd:GdVO4 would be promising self-Raman laser media. Before long, it was first experimentally proved by Chen [2, 3]. In recent years, solid-state Raman lasers based on Nd:YVO4 crystal has been widely developed [4–7]. In those studies, major attention was focused on the generation of the first-Stokes laser or the visible laser combining with frequency-doubling technology. In 2012, Chen et al. [8] investigated second-Stokes YVO4/Nd:YVO4/YVO4 self-Raman laser operating at the wavelength of 1,313 nm for the first time. However, as far as we know, to date there has been no relevant report on second-Stokes self-Raman laser at 1.76 μm with Nd:YVO4. The other methods for generating 1.7 μm laser include solid-state laser with Er3+-doped crystal or intracavity optical parametric oscillators based on KTiOAsO4 or periodically poled lithium niobate media [9–11]. The disadvantages of the later approach are either the complex experimental setup or the necessary special crystal temperature control. The laser sources operating at the eye-safe region of the spectrum (1.5–1.8 μm) are of great interest for various potential applications such as laser ranging, remote sensing and active imaging. The self-Raman laser with the advantages of compact structure and high efficiency will attract more and more attention.

The thermal effects, which are caused by the absorption of pump light and the SRS cascading frequency conversion process in the Nd:YVO4 crystal, are the most important factors affecting the overall performance of the self-Raman laser. To reduce the influence of the thermal effects, one effective method is using the composite Nd:YVO4 crystal as the laser gain medium [12]. Therefore, a 30-mm-long double-end diffusion-bonded composite Nd:YVO4 crystal was employed in our work that not only mitigated the thermal effects but also increased the Raman interaction length for the SRS frequency conversion.

In this paper, an all-solid-state self-Raman second-Stokes laser at 1,764 nm is demonstrated with a diode-pumped actively Q-switched composite YVO4/Nd:YVO4/YVO4 crystal laser. With the incident power of 34 W at a repetition rate of 20 kHz, the maximal average output power of second-Stokes radiation at 1,764 nm was up to 0.99 W, corresponding to an optical conversion efficiency of 2.9 %. The shortest pulse width was 1.92 ns, and the maximal peak power was 21.5 kW.

2 Experimental setup

The diode-pumped second-Stokes self-Raman laser experiments were carried out in a plano-concave resonator, as shown in Fig. 1. The pump source was a commercially available high-power fiber-coupled diode-laser-array at 808 nm. The core diameter and numerical aperture (N.A.) of the fiber were 0.4 mm and 0.22, respectively. The pump beam from the fiber end at the wavelength of 808 nm was focused into the laser crystal with the spot size of 0.4 mm in diameter by an optical imaging system with the imaging ratio of 1:1. The composite YVO4/Nd:YVO4/YVO4 crystal was a 0.3 at % Nd3+-doped 10-mm-long a-cut Nd:YVO4 crystal bounded with a 2-mm-long pure YVO4 at the pumped end and a 18-mm-long pure YVO4 at another end. It was anti-reflection (AR) coated at 808, 1,064 and 1,342 nm on both of its faces. And the transmittances of the laser crystal at 1,524 and 1,764 nm were measured to be 91 and 70 %, respectively. The absorption efficiency of the incident pump power was measured to be about 95 %. To remove the heat generated in the laser crystal, it was wrapped with indium foil and mounted in a water-cooled copper block heat sink. And the water temperature was maintained around 18 °C during the experiments. A compact resonator with a total length of 108 mm was designed for second-Stokes light generation. The front mirror M1 was a concave mirror with a radius of curvature of 250 mm, high-transmittance (HT) coated at 808 nm and high-reflection (HR) coated at 1,342, 1,524 and 1,764 nm. A flat output coupler M2 was HR coated at 1,342 and 1,524 nm, and HT coated at 1,764 nm. A 46-mm-long acousto-optic Q-switch (AOS, Gooch & Housego Co.) was AR coated at 1,342 and 1,524 nm. And its transmittance at 1,764 nm was measured to be 90 %. To suppress parasitic oscillations at the 1.06 μm region, all of the elements in the cavity were high-transmittance (HT) coated at 1.06 μm.

Fig. 1
figure 1

Schematic diagram of the composite Nd:YVO4 crystal second-Stokes self-Raman laser

Full size image

3 Results and discussion

With the above-mentioned experimental configuration, the second-Stokes average output power at 1,764 nm with respect to the incident pump power has been investigated at different pulse repetition frequencies (PRFs) of 20, 25 and 30 kHz, as shown in Fig. 2. Since the second-Stokes output power was very sensitive to the duty cycle of the Q-switch, the duty cycle was optimized first of all. At the PRF of 20 kHz with the duty cycle of 0.79 % for Q-switch operation, the maximum second-Stokes average output power was up to 0.99 W with an incident power of 34 W and a corresponding optical conversion efficiency of 2.9 %. The maximum pulse energy was calculated to be 49.5 μJ. The threshold pump powers of first- and second-Stokes radiation were measured to be 8 and 12.3 W, respectively. At the pump power from 9 to 16 W, the major portion of the output radiation was the first-Stokes output. When the pump power was over 17 W, the second-Stokes line increased rapidly to be the major portion of the output radiation due to the efficient conversion from first-Stokes line. At the PRFs of 25 and 30 kHz, the maximum second-Stokes average output powers were up to 0.863 and 0.607 W, respectively.

Fig. 2
figure 2

Average output power at 1,764 nm versus the incident pump power for PRFs of 20, 25 and 30 kHz

Full size image

We noted that the PRFs could affect the characteristics of the second-Stokes average output power directly. At the smaller PRF, there had been enough time for the population inversion reaching its peak value during two adjacent pulses. Hence, the extremely high peak power of the fundamental pulse could be generated, which enhanced the SRS cascaded conversion efficiency and resulted in the greater second-Stokes average output power eventually. For this reason, we obtained the maximum second-Stokes average output powers at the PRF of 20 kHz. The output power tended to saturate at the maximum level due to the resonator instability caused by strong thermal lens effect resulting from the SRS conversion process [13, 14]. There were several reasons for the low optical conversion efficiency. Firstly, the stimulated emission cross section of the Nd:YVO4 crystal at 1,342 nm was smaller compared with that at 1,064 nm. Secondly, the laser crystal and the Q-switcher were not AR coated at 1,764 nm, which resulted in more loss in the cavity. Finally, the optimization of the output coupling would be beneficial to the higher conversion efficiency. The fundamental radiation was measured to be linearly polarized along the π-direction, so did the first- and second-Stokes radiation. The polarized output has the advantage that it avoids undesirable thermally induced birefringence [15] and improves the Raman conversion efficiency [16, 17]. The laser average output power instability was measured to be 1.95 % for 36 min with an incident power of 34 W and the PRF of 20 kHz.

The spectrum information of the second-Stokes laser was monitored by an optical spectrum analyzer (Yokogawa AQ6375). As shown in Fig. 3, the optical spectrum was recorded at the incident pump power of 34 W and a PRF of 20 kHz. The central wavelengths of the fundamental, first- and second-Stokes radiation were measured to be 1,342.43, 1,524.67 and 1,764.24 nm, with the corresponding line widths of 0.16, 0.18 and 0.27 nm, respectively. It can be seen that the frequency shift of each adjacent wavelengths interval was about 890 cm−1, which was the optical vibration modes of the tetrahedral VO4 3− ionic groups (890 cm−1). The optical spectrum indicates that very small residual of fundamental emission and first-Stokes Raman emission was present in the output.

Fig. 3
figure 3

The optical spectrum of the actively Q-switched composite YVO4/Nd:YVO4/YVO4 self-Raman laser

Full size image

A grating monochromator was applied to separate the fundamental, first- and second-Stokes output radiation. The temporal pulse behavior was recorded by a mixed signal oscilloscope (Tektronix MSO 4032) with a fast photodiode detector (EOT ET-3500). The pulse widths of the fundamental, first- and second-Stokes radiation, as shown in Fig. 4a, were measured as a function of the incident pump power at the PRF of 20 kHz. The shortest pulse width of second-Stokes was measured to be 1.92 ns with a pump power of 24 W corresponding to the fundamental and the first-Stokes pulse width of 9.8 and 5.83 ns, respectively, as shown in Fig. 4b–d. The pulse shortening was due to the cascading nonlinear frequency conversion process of SRS [18, 19]. It can also be seen that the pulse width of higher-order Stokes radiation is shortened much more.

Fig. 4
figure 4

a Pulse duration for fundamental (1,342 nm), first- (1,524 nm) and second-Stokes radiation (1,764 nm) with respect to the incident pump power at PRF of 20 kHz; b–d typical temporal pulse profiles for three wavelengths at pump power of 24 W and the PRF of 20 kHz

Full size image

The second-Stokes output beam profiles were measured by a CCD camera (Spiricon PY-III), as shown in Fig. 5a. The beam quality factor (M 2) of the second-Stokes was measured to be 1.43 × 1.45 in the horizontal and vertical directions at the incident pump power of 30 W and the PRF of 20 kHz. It indicated that the second-Stokes radiation was mainly contributed by TEM00 mode. As shown in Fig. 5b, the artificial X and Y beam widths plotted against their Z-axis sample locations. The beam widths were plotted in micrometers (μm) and the Z values in millimeters (mm). The solid lines were the resulting curve fits of the beam propagation equation according to the plotted data. Figure 6 shows the peak power of second-Stokes radiation as a function of the incident pump power at different PRFs. The highest peak power was up to 21.5 kW at the pump power of 34 W and PRF of 20 kHz, with the corresponding single pulse energy of 49.5 μJ. At the same incident pump power, we can see that the single pulse energy and the peak power with a PRF of 20 kHz are both greater than that with PRFs of 25 and 30 kHz.

Fig. 5
figure 5

a Measured beam profile of second-Stokes emission at 1,764 nm; b measured beam widths and the resulting curve fits of the beam propagation equation

Full size image
Fig. 6
figure 6

a Single pulse energy and b peak power at 1,764 nm versus the incident pump power for different PRFs

Full size image

4 Conclusions

In summary, the second-Stokes self-Raman radiation at 1,764 nm was achieved in a diode-end-pumped actively Q-switched composite YVO4/Nd:YVO4/YVO4 crystal laser. At the pump power of 34 W and PRF of 20 kHz, the maximum average output power was obtained to be 0.99 W, corresponding to the optical conversion efficiency of 2.9 %. The shortest pulse width was 1.92 ns, and the highest peak power and pulse energy were 21.5 kW and 49.5 μJ, respectively. Note that the laser crystal and the Q-switcher are not AR coated at 1,764 nm. Thus, it is expected that a higher output power of second-Stokes radiation can be achieved with 1,764 nm HT-coated laser crystal and the Q-switcher.