1. Introduction

Stimulated Raman scattering (SRS) process in crystals becomes the increasingly important way to generate new laser spectral lines. Actively Q-switched Raman lasers have been widely investigated [1–6]. However, active Q-switching techniques employ acousto-optic or electro-optic modulators, which require external driving devices, and passive Q-switching techniques have the advantages of simplicity, compactness and low cost. Cr⁴⁺:YAG crystal has been proved to be a suitable passive Q-switcher for 0.9~1.2µm wavelength.

Some rare earth doped laser crystals are found to be Raman-active, for example Nd:KGd(WO₄)₂, Yb:KLu(WO₄)₂, Nd:PbWO₄, Nd:YVO₄, Nd:GdVO₄, Yb:YVO₄ and Nd:SrMoO₄ [7–17]. These crystals can be utilized to realize the self-Raman lasers, in which the laser crystal also acts as Raman medium. Self-Raman lasers can generate new laser lines with very compact, stable and simple structure.

With the self-Raman configuration and passively Q-switched element, compact, stable and pulsed Raman lasers can be realized. This kind of lasers can generate new laser lines with high peak intensity and narrow pulse width, which have wide range of applications. In 2004, Y. F. Chen realized the passively Q-switched c-cut Nd:YVO₄/Cr⁴⁺:YAG self-Raman lasers, the maximum output power of first Stokes at 1178.6nm is 125mW, and the conversion efficiency is 6.3% [11]. In the same year, Y. F. Chen reported a diode-pumped passively Q-switched c-cut Nd:GdVO₄/Cr⁴⁺:YAG self-Raman laser, and the self-Raman laser produced the average output power of 140mW at 1175.6 nm [12]. In 2005, Yb:YVO₄/Cr⁴⁺:YAG self-Raman laser was realized to generate the 1119.5nm laser with pulse energy of 3.6μJ at a repetition rate of 25 kHz [13]. The passively Q-switched Yb:KLu(WO₄)₂ self-Raman laser with 0.4 W output power at 1137.0 nm was reported in 2006 [14]. In 2009, Basiev et al. realized the passively Q-switched SrMoO₄:Nd³⁺/LiF:F^-₂ self-Raman laser, and average output power of 0.25W was obtained at 1163 nm [15]. In the same year, passively Q-switched output from a a-cut Nd:YVO₄/YVO₄/Cr:YAG self-Raman laser was reported, and the maximum Raman output of 570mW was obtained at 1176 nm [16]. In 2012, Y. X. Shi et al. obtained the maximum average output power of 347mW at 1176nm with the passively Q-switched a-cut Nd:YVO₄/Cr⁴⁺:YAG self-Raman laser [17].

Nd:YVO₄ crystal is an excellent laser medium for LD pumping with many advantages, such as large absorption coefficient, large stimulated emission cross section, and wide absorption bandwidth. At the same time, Nd:YVO₄ crystal is good Raman crystal, and can be obtained easily. In 2001, Kaminskii et al. predicted that YVO₄ crystals were the promising Raman media, and the Raman gain coefficient in the NIR spectral region is estimated to be about 4.5cm/GW [7]. In 2004, Chen et al. realized the efficient and compact solid-state self-Raman lasers with Nd:YVO₄ crystals [10, 11].

Nd:YVO₄ crystal crystallizes in a D_4h tetragonal space group of the zircon type, and its unique optical axis is along the four-fold symmetric axis c. The π-polarized (electric field∥c) transition has much larger stimulated emission cross section (σ_// = 25 × 10⁻¹⁹cm²) than the σ-polarized (electric field⊥c) one (σ_⊥ = 6.5 × 10⁻¹⁹cm²) for the emission at 1.06 μm. However, high laser emission cross section limits energy storage capacities in the passively Q-switched regime. When a Nd:YVO₄ crystal is cut along the c axis, i.e., c-cut, the effective stimulated-emission cross section is σ_⊥ instead of σ_//. A c-cut Nd:YVO₄ crystal is more appropriate for passive Q-switching operation [11, 18, 19]. For intracavity Raman lasers, the normalized Raman gain is larger for the laser medium with smaller stimulated-emission cross section [20].

In this paper, with c-cut Nd:YVO₄ crystal of 20mm in length as the laser and Raman media, we realize the efficient, compact LD end-pumped passively Q-switched self-Raman laser. The temporal, polarization, spectrum and spatial features of the self-Raman laser are studied experimentally in detail. Its optimized conditions are analyzed, and the most efficient Raman conversion is obtained under the conditions of Cr⁴⁺:YAG with initial transmission of T₀ = 89% and output couplers with reflectivity of R_S = 90% at 1178nm. The maximum average output power at 1178nm obtained is 800mW with respect to the incident pump power of 7.9W, and the corresponding conversion efficiency is 10.1%. To the best of our knowledge, this is the highest average output power ever obtained from passively Q-switched self-Raman lasers. With c-cut Nd:YVO₄ crystal as laser medium, laser output should be unpolarized. However, the outputs of fundamental and Stokes laser are found to be linearly polarized along the diagonals of the rectangular cross section of c-cut Nd:YVO₄ crystal in the experiments, and the polarization mode competition is observed in the outputs of fundamental and first Stokes laser.

2. Experimental arrangement

Figure 1 shows the experimental setup of the diode-pumped passively Q-switched c-cut Nd:YVO₄/Cr⁴⁺:YAG self-Raman laser. The pump source is a 30W fiber-coupled 808nm laser diode with a core diameter of 400μm and a numerical aperture of 0.22. A focusing lens system (1:1 magnification) with a coupling efficiency of 95% is used to reimage the pump beam into the laser crystal. The resonator has a linear configuration. The rear mirror M₁ is a concave mirror. It is coated for high reflection (HR) at 1066nm (R>99.8%) and 1178nm (R>99.5%), and high transmission (HT) at 808nm (T>90%). The plane output coupler M₂ is coated for high reflection at 1066nm (R>99.8%), and two output couplers (OC) with different reflectivities at 1178nm (R_S = 80% and 90%) are used in our experiment. The laser material is a 0.3%-Nd-doped c-cut Nd:YVO₄ crystal with a dimension of 3 × 3 × 20mm³, and both faces are AR coated at 1066nm (R<0.2%), 1178nm (R<0.2%), and 808nm (R<3%). The c-cut Nd: YVO₄ crystal acts as the Raman active material at the same time. The saturable absorbers Cr⁴⁺: YAG are AR coated at 1066nm (R<0.2%). Nd:YVO₄ and Cr⁴⁺:YAG crystals are wrapped with indium foil and held in water-cooled copper blocks. The water temperature is controlled to be 23°C throughout the experiment. The saturable absorber is put close to the output coupler to obtain small cavity mode size in it, which is beneficial to the passively Q-switched criterion [19]. The overall laser cavity length is approximately 30mm. The distance between M₁ and Nd:YVO₄ crystal is about 3mm, and Cr⁴⁺:YAG crystal is apart from Nd:YVO₄ crystal and M₂ by 4mm and 2mm, respectively. A dichroic mirror separates the output of 1178nm from those of 808nm and 1066nm. The average output power is measured by a power meter (Ophir Laserstar). The pulse temporal behavior is recorded by an Agilent digital oscilloscope (DSO7104A, 1GHz Bandwidth) with a fast PIN photodiode.

 

Fig. 1 Experimental setup of the diode-pumped passively Q-switched c-cut Nd:YVO₄ /Cr⁴⁺:YAG self-Raman laser

Download Full Size | PDF

3. Experimental results and discussion

3.1 Output power at 1178nm

The curvature radius of rear mirror is related to the resonator mode size of fundamental laser, and affects the Raman laser performance mainly in two aspects. On one hand, with other parameters invariable, smaller curvature radius of rear mirror leads to larger A/A_a, and fundamental pulses can be generated of narrower pulse width and higher peak power [18, 19], which are preferred for Raman conversion. The other effect is mode matching condition between the fundamental laser and pump beams. The experiments are carried out for mirrors M₁ with curvature radii of 150mm, 300mm, and 500mm, respectively. The results show highest Stokes conversion efficiency relative to the pump power is realized with M₁ of 300mm due to good mode matching condition and relative large A/A_a, therefore M₁ of 300mm is adopted in the following experiments.

Figure 2 shows the average output power (a) and pulse energy (b) at 1178nm with respect to the incident pump power for Cr⁴⁺:YAG of initial transmission T₀ = 89% and 86%, and M₂ of R_S = 80% and 90%, respectively. The most efficient Raman conversion is obtained under the conditions of T₀ = 89% and R_S = 90%, and the first Stokes average output power is much higher than the other three lines as shown in Fig. 2(a). The maximum average output power obtained is 800mW with respect to the incident pump power of 7.9 W, and the corresponding conversion efficiency is 10.1%. This is the highest average output power ever obtained from LD pumped passively Q-switched Nd:YVO₄ self-Raman lasers. The average output power increases with incident pump power, and finally drop down beyond the incident pump power of 8.8 W with T₀ = 89% and R_S = 90%, which is due to unstable resonator caused by thermal lensing in the laser medium. The pulse energy decreases slightly with incident pump power as shown in Fig. 2(b). The average pulse energy is about 18μJ with T₀ = 89% and R_S = 90%, and maximum pulse energy is over 30μJ with T₀ = 86% and R_S = 90%. The Stokes pulse energy decreases with increasing the pump power, and there are two main reasons. One is the thermal lens effect in the laser crystal, which reduces the laser mode volume, resulting in a decrease of the pulse energy. The other is the pump-induced bleaching of the saturable absorber. With increasing the incident power, the laser medium could not completely absorb the pump power, and the remaining pump power, to some extent, could bleach the saturable absorber, resulting in a significant decrease of the pulse energy [21].

 

Fig. 2 Average output power (a) and pulse energy (b) at 1178nm with respect to the incident pump power for Cr⁴⁺:YAG of initial transmission 89% and 86%, and M₂ of R_S = 80% and 90%, respectively

Download Full Size | PDF

3.2 Temporal characteristics

Pulse repetition frequency (a) and pulse width (b) at 1178nm with respect to the incident pump power is shown in Fig. 3 for Cr⁴⁺:YAG of T₀ = 89% and 86%, and M₂ of R_S = 80% and 90%, respectively. As the incident pump power increases, the PRF of the Raman laser grows, and increases from 8.9 kHz to 46.4 kHz and from 4.5 kHz to 29.5 kHz for Cr⁴⁺:YAG of T₀ = 89% and 86%, respectively. The PRF for T₀ = 89% is higher than that for T₀ = 86%, and the output coupling rate has a little effect on PRF.

 

Fig. 3 Pulse repetition frequency (a) and pulse width (b) at 1178nm with respect to the incident pump power for Cr⁴⁺:YAG of T₀ = 89% and 86%, and M₂ of R_S = 80% and 90%, respectively

Download Full Size | PDF

As in Fig. 3(b), for T₀ = 89% and R_S = 80%, the pulse duration of first Stokes laser fluctuates widely, and has unstable output. With higher initial transmission of Cr⁴⁺:YAG, the pulse energy and peak intensity of fundamental laser is lower, at the same time, the output coupling rate is too high to maintain the efficient Raman conversion inside the resonator, resulting in the unstable output of first Stokes. For T₀ = 89% and R_S = 90%, the pulse duration shows fluctuation when pump power is below 4.5W, and becomes stable for higher pump power. Figure 4 shows profiles of first Stokes in (a) generated under the pump power of 5.0W for T₀ = 89% and R_S = 90%, and the train of first Stokes in (b). In the experiment, very stable pulse train of first Stokes is obtained with T₀ = 89% and R_S = 90% as shown in Fig. 4(b). The fluctuation of pulse peak power is less than 7%. There is the intensity modulation of first Stokes pulses as shown in Fig. 4(a). In the experiment, the intensity modulation of fundamental laser pulses is also observed, and is not as obvious as that of first Stokes pulses. The intensity modulation of fundamental laser pulses is related to the excited state absorption (ESA) of Cr:YAG crystal. Because the output coupler is HR coated at the fundamental laser in the experiment, its intracavity intensity is so high that most of the Cr⁴⁺ ions are excited to the first excited state, and further excitation will lead to an ESA saturation. The relaxation time of the ESA is in sub nanosecond region, and sub nanosecond intensity modulation of Q-switched pulses of fundamental laser is present. In 2009, the evolutions of first Stokes from spontaneous Raman scattering were simulated numerically by our group for the intracavity Raman lasers [22], and the numerical results showed the intensity modulation and self-mode-locking effect of the first Stokes. In the spontaneous Raman scattering emission, the lucky spike whose intensity surpasses the SRS threshold experiences larger Raman gain due to the exponential growth of the first Stokes intensity, and grow rapidly into pulse profile inside the resonator. The numerical results show that intensity modulation of fundamental laser pulses will strengthen the intensity modulation of first Stokes. The pulse duration is stable on the whole for T₀ = 86%, however, there is an unexpected pulse duration reduction for R_S = 80% when the pump power ranges form 6.5W to 8W. Profile of first Stokes pulse generated under the pump power of 7.3W for T₀ = 86% and R_S = 80% is shown in Fig. 4(c), and it is found that intensity modulation of the first Stokes pulse lead to this unexpected reduction of pulse width (800ps).

 

Fig. 4 Profiles of first Stokes pulse in (a), the train of first Stokes in (b) generated under the pump power of 5.0W for T₀ = 89% and R_S = 90%; unexpected reduction of first Stokes pulse duration generated under the pump power of 7.3W for T₀ = 86% and R_S = 80% in (c).

Download Full Size | PDF

The normalized rate equations of passively Q-switched intracavity Raman lasers were deduced, and numerical analysis were carried out in [20]. According to the definitions in [20], the normalized Raman gain M for the experiment shown in Fig. 1 is calculated to be 24, and saturable absorber synthetic parameter α is 23. The normalized initial population inversion density N and normalized first Stokes loss K for different T₀ and R_S are estimated as shown in Table 1, and the optimum values of normalized Raman gain M_opt are obtained to achieve the maximum integral of the normalized first Stokes photon density for different conditions as given in Table 1. Figure 5 shows temporal profiles of first Stokes pulses generated with the pump power of 5W for T₀ = 89% and R_S = 90% in (a); T₀ = 89% and R_S = 80% in (b); T₀ = 86% and R_S = 90% in (c); T₀ = 86% and R_S = 80% in (d). As shown in Figs. 5(a)-5(c), the satellite pulses of first Stokes laser are present. With the larger R_S and smaller T₀, the multi-pulse phenomenon is more serious. There are two reasons for this phenomenon. One is the long upper level lifetime (3 μs~4 μs) of the slow recovery saturable absorber Cr:YAG crystal, which provides the preconditions for the emergence of satellite pulses. The other is that too large Raman gain will lead to satellite pulses of the fundamental and first Stokes, according to the numerical modeling of passively Q-switched intracavity Raman lasers [20]. When the SRS threshold is reached, the intracavity Stokes photon densities go up sharply while the fundamental photon densities decrease rapidly. However, the population inversion is not depleted completely, and another fundamental pulse will occur later. When the SRS threshold is reached again, the satellite Stokes pulse emerges. For our experiments, normalized Raman gain M = 24 is larger than M_opt given in Table 1, and the satellite pulses of first Stokes were observed in the experiment as shown in Fig. 5. In Fig. 5(a), the main pulse is much larger than the satellite pulses, and conversion efficiency of first Stokes is highest for T₀ = 89% and R_S = 90% as shown in Fig. 2, which means the Raman laser works under the good conditions. In Fig. 5(b), a very small satellite pulse presents, which indicates the Raman laser works just slightly above M_opt, and conversion efficiency is low. In Fig. 5(c), satellite pulse phenomenon is serious, and the number of satellite pulses is four, and their energies are high, which indicates the Raman laser works far above M_opt, with which the conversion efficiency of Raman laser is low. In Fig. 5(d), the number of the satellite pulses is 2, and pulse width of the main pulse is much smaller than that in Fig. 5(c), which means the output coupling rate is too high to maintain the efficient Raman conversion inside the resonator. In conclusion, when the normalized Raman gain M is moderately larger than M_opt as with T₀ = 89% R_S = 90%, the efficient and stable Raman laser is realized as shown in Fig. 2 and Fig. 4. The working conditions given in Figs. 5(b), 5(c) and 5(d) are far from optimized ones.

Table 1. The normalized parameters for passively Q-switched c-cut Nd:YVO₄/Cr⁴⁺:YAG self-Raman lasers

View Table | View all tables in this article

 

Fig. 5 Temporal profiles of first Stokes pulses generated with the pump power of 5W for T₀ = 89% and R_S = 90% in (a); T₀ = 89% and R_S = 80% in (b); T₀ = 86% and R_S = 90% in (c); T₀ = 86% and R_S = 80% in (d).

Download Full Size | PDF

3.3 Polarization characteristics

The unique optical axis of Nd:YVO₄ crystal is along the four-fold symmetric axis c. Perpendicular to this axis are the two indistinguishable a and b axes. With c-cut Nd:YVO₄ crystal as laser medium, the stimulated emission cross section is same for electric field ∥a and ∥b, and the output laser should be unpolarized. However, we get the linearly polarized fundamental and first Stokes outputs, and also observe the polarization mode competition in our experiment.

In the experiment, the c-cut Nd:YVO₄ crystal has a rectangular cross section of 3X3mm², which are cut along a and b optical axes, respectively. The water cooled copper blocks including Nd:YVO₄ crystal and indium foil is depicted in Fig. 6. In the experiment, the outputs of fundamental and Stokes laser are found to be linearly polarized along the diagonals of the rectangular cross section, named horizontal and vertical directions in Fig. 6.

 

Fig. 6 Water cooled copper blocks including Nd:YVO₄ crystal and indium foil

Download Full Size | PDF

Table 2 gives horizontally and vertically polarized average output powers at 1066nm of passively Q-switched c-cut Nd:YVO₄/Cr⁴⁺:YAG laser for different pumping power with Cr⁴⁺:YAG of T₀ = 86%, output coupling rate of R = 90%@1066nm, and the laser cavity length of 30mm. For most of the times, the polarization of the fundamental is along horizontal direction, however, polarization mode competition are observed under certain pumping power, e.g. 7.7 W as given in Table 2.

Table 2. Polarized average output powers at 1066nm of passively Q-switched c-cut Nd:YVO₄/Cr⁴⁺:YAG laser

View Table | View all tables in this article

A Glan prism is placed in front of the fast PIN photodiode. When the transmission axis of the polarizer is rotated, two different sets of pulse trains are recorded with the oscilloscope in the fundamental laser output. With the transmission axis nearly parallel to the vertical direction, two pulse trains of 1066nm are recorded simultaneously as shown in Fig. 7(a) for pump power of 6.9W with Cr⁴⁺:YAG of T₀ = 86%, output coupling rate of R = 90%@1066nm, and the laser cavity length of 30mm. The vertically polarized pulse train has high intensity, and horizontally polarized one has low intensity in Fig. 7(a). The polarized mode competition between two polarization directions is obvious. The polarized mode competition is also present in the first Stokes output. Pulse trains at 1178nm output of passively Q-switched self-Raman laser are shown in Figs. 7(b) and 7(c). For pump power of 5.9W with Cr⁴⁺:YAG of T₀ = 86% and output coupling rate of R_S = 90%@1178nm, horizontally polarized pulse train is in (b), and vertically polarized one in (c). The horizontal polarization is advantageous to the vertical one.

 

Fig. 7 Pulse trains of 1066nm output of passively Q-switched laser in (a); pulse trains at 1178nm of passively Q-switched self-Raman laser, horizontally polarized pulse train in (b) and vertically polarized one in (c).

Download Full Size | PDF

In our experiment, the LD pumping beam is tested to be unpolarized. With the polarization interference method, c-cut Nd:YVO₄ crystal used in the experiment is verified to be perfect uniaxial crystal. We attribute the unusual polarization phenomenon to thermally induced birefringence for the moment. The laser medium is subject to convective cooling by the surrounding water, causing non-uniform temperature distribution, furthering resulting in thermal stress and stress birefringence. The temperature, thermal stress, and thermal strain distribution in laser medium are solved both analytically and numerically in [23]. It is found that the distributions of temperature gradient, thermal stress and thermal strain are not axisymmetric on the rectangular cross section of laser medium. The temperature gradient is smallest along the two diagonals of the rectangle, and the thermally induced birefringence leads to the laser output polarized along these two directions. In the experiment, water cooled copper blocks comprise of two parts as shown in Fig. 6, and clasp laser crystal between them, whose contact area is covered with thermally conductive silicone. The thermally conductive conditions of vertical and horizontal directions are different, and the horizontal polarization is found to be advantageous to the vertical one in the experiments. The unusual polarization phenomenon caused by thermally induced birefringence will be studied in more detail. We believe that higher efficient passively Q-switched c-cut Nd:YVO₄/Cr⁴⁺:YAG self-Raman laser can be realized by eliminating this unusual polarization phenomenon.

3.4 Spectrum characteristics

The spectrum information of the intracavity Raman laser was registered with an optical frequency analyzer (ANDO, AQ-6317B, 600nm-1750nm) with the resolution of 0.5nm. Figure 8 depicts spectra of passively Q-switched c-cut Nd:YVO₄/Cr⁴⁺:YAG self-Raman laser under the pump power of 4.3W for T₀ = 89% and R_S = 90%. As shown in Fig. 8, four spectrum lines were observed in the Raman laser output, which are registered in 1066.7nm, 1097.3nm, 1129.5nm and 1178.8nm, respectively. The line of 1178.8nm is the most intensive, and the corresponding SRS shift is calculated to be 889.7cm⁻¹ relative to the fundamental laser registered in 1066.7nm, which is in accordance to the reported Raman shift 890cm⁻¹ of the totally symmetric A_1g optical vibration modes of tetrahedral VO₄^2--ionic groups [7]. The lines of 1097.3nm and 1129.5nm are very weak, and they are first and second Stokes lines caused by the interaction of fundamental laser with the optical vibration mode of 259.8cm⁻¹ [7]. As shown in Fig. 8(b), besides the fundamental laser line registered in 1066.7nm (originates from the R₂ component of the ⁴F_3/2 level, and terminates at the Y₃ component of the ⁴I_11/2 level.), another weaker line in 1062.8nm is observed, which originates from the R₁ component of the ⁴F_3/2 level, and terminates at the Y₂ component of the ⁴I_11/2 level.

 

Fig. 8 (a) Spectrum of passively Q-switched c-cut Nd:YVO₄/Cr⁴⁺:YAG self-Raman laser output under the pump power of 4.3W for T₀ = 89% and R_S = 90%, (b) the spectrum of fundamental laser corresponding to (a)

Download Full Size | PDF

3.5 Spatial characteristics

The spatial features of laser output are studied with laser beam analyzer BeamOn-IR-1550 (Duma Optronics Ltd., Israel). Figure 9 shows the spatial profile of fundamental laser in (a) of passively Q-switched c-cut Nd:YVO₄/Cr⁴⁺:YAG laser under the pump power of 4.8W for T₀ = 89% and output coupling rate of R = 90%@1066nm, and spatial profile of first Stokes laser in (b) of passively Q-switched c-cut Nd:YVO₄/Cr⁴⁺:YAG self-Raman laser output under the pump power of 6.2W for T₀ = 89% and R_S = 90%. M₂ factor of fundamental laser beam is measured to be 1.0 under the pump power of 4.8W for T₀ = 89% and output coupling rate of R = 90%@1066nm, and perfect fundamental laser beam of TEM₀₀ mode is generated as shown in Fig. 9(a). For passively Q-switched c-cut Nd:YVO₄/Cr⁴⁺:YAG self-Raman laser under the pump power of 6.2W for T₀ = 89% and R_S = 90%, M₂ factor of the first Stokes laser beam is measured to be 2.8 as shown in Fig. 9(b). For passively Q-switched c-cut Nd:YVO₄/Cr⁴⁺:YAG self-Raman laser, because the output coupler M₂ is coated for high reflection at 1066nm (R>99.8%), fundamental laser of high order spatial mode and TEM₀₀ mode will oscillate simultaneously under the pump power of 6.2W resulting in the M₂ factor of the first Stokes laser to be 2.8.

 

Fig. 9 Spatial profile of fundamental laser in (a) of passively Q-switched c-cut Nd:YVO₄/Cr⁴⁺:YAG laser under the pump power of 4.8W for T₀ = 89% and output coupling rate of R = 90%@1066nm, spatial profile of first Stokes laser in (b) of passively Q-switched c-cut Nd:YVO₄/Cr⁴⁺:YAG self-Raman laser output under the pump power of 6.2W for T₀ = 89% and R_S = 90%.

Download Full Size | PDF

4. Conclusion

For compact LD end-pumped passively Q-switched c-cut Nd:YVO₄/Cr⁴⁺:YAG self-Raman laser, when the normalized Raman gain M is moderately larger than M_opt as with Cr⁴⁺:YAG of initial transmission T₀ = 89% and output couplers of reflectivity R_S = 90% at 1178nm, the efficient and stable Raman laser is realized. The maximum average output power at 1178nm obtained is 800mW with respect to the incident pump power of 7.9W, the pulse repetition frequency is 44kHz, pulse width is 2.6ns, pulse energy is 18μJ, and peak power intensity is 7.0kW. The corresponding conversion efficiency for LD pump power to first Stokes laser is 10.1%. With c-cut Nd:YVO₄ crystal as self-Raman medium, the outputs of fundamental and Stokes laser are found to be linearly polarized along the diagonals of the rectangular cross section in the experiments, and the polarization mode competition is observed in the output of fundamental and first Stokes laser. The mechanism of the unusual polarization phenomenon is explained from the thermally induced birefringence aspect. It is predicted that higher efficient passively Q-switched c-cut Nd:YVO₄/Cr⁴⁺:YAG self-Raman laser can be realized by eliminating the unusual polarization phenomenon. M₂ factor of the first Stokes laser beam is measured to be about 2.