1. Introduction

It is well known that the multi-phonon nonradiative transitions between some related multiplets of the doped active ions, and the quantum defect caused by the energy difference between pump and laser photons, can generate pump-induced heat in the gain medium of diode-pumped solid-state laser [1]. Then, thermal effect of the gain medium will limit the maximum average output power, reduce the efficiency of laser operation and output beam quality, and destroy the stability of resonator.

Compared with those of Nd³⁺ or Yb³⁺ singly doped materials, there is more heat generation in the Er³⁺ and Yb³⁺ co-doped materials operating at 1.5-1.6 μm, which can be used in some applications such as laser-range-finding, lidar, and medicine [2]. Firstly, for the diode-pumped Er-Yb laser, the commonly used pump wavelength λ_p is 976 nm and laser wavelength λ_l is around 1.55 μm. Therefore, the quantum defect (1−λ_p/λ_l) of the Er-Yb laser is about 37%, which is larger than the 24% and 8.5% of the common diode-pumped Nd³⁺ or Yb³⁺ lasers, respectively [1]. Secondly, high effective phonon energy of host crystal enhances the multi-phonon relaxation from ⁴I_11/2 to ⁴I_13/2 multiplet of Er³⁺, and then restrains the back energy-transfer from Er³⁺ to Yb³⁺ and upconversion loss originated from the ⁴I_11/2 multiplet. Therefore, the selection of the material with high effective phonon energy, such as phosphate glass and borate crystal [2–5], as Er³⁺ and Yb³⁺ co-doped host is advantageous for realizing the 1.5-1.6 μm laser with high slope efficiency. However, high effective phonon energy also increases the multi-phonon relaxation from ⁴I_13/2 to ⁴I_15/2 multiplet of Er³⁺ and then reduces the fluorescence quantum efficiency of the upper laser multiplet ⁴I_13/2. For example, fluorescence quantum efficiency of the ⁴I_13/2 multiplet of the Er³⁺:Yb³⁺:RAl₃(BO₃)₄ (Er:Yb:RAB, R = Y, Gd and Lu) crystal, which is the best 976 nm-diode-pumped 1.5-1.6 μm laser crystal reported presently [3], is only about 7% [3, 6]. Then, a large amount of absorbed pump power is converted into heat during the pump process. Moreover, 1.5-1.6 μm laser corresponding to the ⁴I_13/2→⁴I_15/2 transition of Er³⁺ operates in a quasi-three-level scheme. Ions populated in the lower laser level of ⁴I_15/2 multiplet through Boltzmann distribution act as reabsorption loss, which decreases the operation efficiency and increases the laser threshold. Energy level splitting of ⁴I_15/2 multiplet of Er³⁺ is smaller than that of ²F_7/2 multiplet of Yb³⁺ in the same crystal, such as 308 cm⁻¹ for Er:YAB and 581 cm⁻¹ for Yb:YAB [7, 8]. Therefore, with the increment of crystal temperature caused by the pump-induced heat, the influence of reabsorption loss on laser performances is more serious in the Er-Yb laser. As a result, the slope efficiency and maximum output power realized presently in the Er-Yb 1.5-1.6 μm lasers are limited by the strong thermal effect of the Er³⁺ and Yb³⁺ co-doped materials [9, 10].

Compared with Er:Yb:YAB crystal, Er³⁺:Yb³⁺:LuAl₃(BO₃)₄ (Er:Yb:LuAB) crystal has higher optical quality and better laser performances because the radius and mass of Lu³⁺ are the closest to those of Yb³⁺, respectively [4, 6]. In this paper, in order to reduce the thermal effect of the Er:Yb:LuAB crystal, a sandwich-type sapphire/Er:Yb:LuAB/sapphire micro-laser, in which the sapphire crystals act as heat diffuser and the cavity mirrors were deposited onto the outside surfaces of the sapphire crystals, was fabricated. Pumped by a continuous-wave (cw) 976 nm laser diode (LD), laser performances including maximum output power, slope efficiency and beam quality can be enhanced in the sandwich-type micro-laser compared with those of a monolithic Er:Yb:LuAB micro-laser operating in a same experimental condition.

2. Experimental arrangement

The experimental setup is depicted in Fig. 1. A 0.7-mm-thick Er(1.1 at.%):Yb(24.1 at.%):LuAB microchip with cross section of 5 mm × 5 mm was used as the gain medium. The uncoated microchip was placed between two polished 1.5-mm-thick sapphire crystals with cross section of 5 mm × 5 mm. In order to avoid the birefringent effect, both Er:Yb:LuAB and sapphire crystals were c-cut. Input mirror (IM) and output mirror (OM) were directly deposited onto the outside surfaces of the sapphire crystals. The sandwich-type micro-laser with thickness of about 3.7 mm was mounted in a copper holder, which was cooled by water at about 20 °C, and the close contact between the crystals was completed using screws. All the surfaces of the micro-laser were contacted with copper and there is a hole with radius of about 1 mm in the center of the holder to permit the passing of the laser beams. In this sandwich-type micro-laser, heat generated in the pump region of the gain medium can effectively diffuse out through the sapphire crystals because they have a high thermal conductivity of 42 Wm⁻¹K⁻¹ [11]. The sapphire and Er:Yb:LuAB crystals have similar refractive indexes (about 1.75) at 1.5-1.6 μm. Therefore, the interface reflectivity between them is close to zero without the addition of index matching oil. Furthermore, the sapphire crystal is easily available and cheap at present because its mass production has been realized. A 976 nm fiber-coupled LD (100 μm diameter core) was used as the pump source. After passing a telescopic lens system (TLS) consisted of two convex lenses with the same focal length of 45 mm, pump beam was focused to a spot with radius of about 50 µm in the microchip. IM has 90% transmission at 976 nm and 99.8% reflectivity at 1.5-1.6 μm. For avoiding the laser oscillating simultaneously at longer wavelengths of about 1560 and 1600 nm [6], OM has 1.5% transmission around 1540 nm and transmission larger than 2.5% at wavelength longer than 1560 nm.

 

Fig. 1 Experimental setup of the cw 976 nm diode-end-pumped sapphire/Er:Yb:LuAB/sapphire micro-laser.

Download Full Size | PDF

3. Results and discussion

Figure 2 shows cw output power of the sapphire/Er:Yb:LuAB/sapphire micro-laser versus incident pump power. Maximum output power of 1.17 W was obtained at incident pump power of 5.42 W and the threshold was about 1.1 W. The fluctuation of output power was less than ± 2.5% in an hour. No output power saturation phenomenon was observed in the micro-laser for the maximum incident power available in our lab at present. Therefore, higher output power will be realized in the micro-laser when a higher pump power is used in the future. When the pump power was higher than 4 W, slope efficiency η with respect to incident pump power was 33%. For the pump power lower than 4 W, slope efficiency decreased to 21%. One reason responsible for the decrement of slope efficiency may be the reabsorption loss of the quasi-three-level laser caused by the ions population in lower laser energy, which is decreased with the increment of the fundamental laser intensity in the cavity [12]. Furthermore, Due to the poor temperature control of the LD assembled in our lab, its emission wavelength is changed from 970 to 976 nm with the increment of LD current. Therefore, the better matching between the LD emission wavelength and the absorption spectrum of Er:Yb:LuAB crystal at higher pump power can also lead to the higher slope efficiency. Spectra of the micro-laser were recorded with a monochromator (Triax550, Jobin-Yvon) associated with a TE-cooled Ge detector (DSS-G025T, Jobin-Yvon), and they are similar at various pump power. The spectrum at pump power of 5.42 W is also shown in Fig. 2 and laser wavelength was centered at about 1543 nm. The wide operating wavelength range is caused by the broad gain spectrum of the Er:Yb:LuAB crystal at 1.5-1.6 μm [6]. The spacing between the adjacent laser lines is about 0.5 nm and in agreement with the theoretical spacing ∆λ (about 0.46 nm) of laser lines caused by the etalon effect associated to the 1.5-mm-thick sapphire crystal, which is calculated by $\Delta \lambda \text{=}\frac{{\lambda }^{\text{2}}}{\text{2}nL}$ [13]. Here λ is the laser wavelength, n is refractive index of the crystal at λ, and L is the thickness of the sapphire crystal. Then, by the use of etalon effect associated to the sapphire crystal with different thickness, it is possible to control laser wavelength of the micro-laser. In order to investigate the effectiveness of the sapphire crystal as heat diffuser, output power of a monolithic Er:Yb:LuAB micro-laser, in which the cavity mirrors with the same transmissions were directly deposited onto the surfaces of a 0.7-mm-thick Er:Yb:LuAB microchip, was recorded in the same experimental condition except for the using of the sapphire crystals. With the increment of pump power, output power saturation phenomenon was observed in the monolithic micro-laser and maximum output power of 460 mW was obtained at pump power of 4.5 W. The fluctuation of output power was less than ± 4.3% in an hour. When the higher pump power was used, the fracture of the microchip was observed in this work. For the monolithic micro-laser, the slope efficiency and threshold with respect to incident pump power were about 17% and 1.3 W, respectively. Obviously, adoption of the sapphire crystals can effectively reduce the thermal effect of the Er:Yb:LuAB microchip and then enhance its output laser performances.

 

Fig. 2 1543 nm cw output power of the sandwich-type sapphire/Er:Yb:LuAB/sapphire and monolithic Er:Yb:LuAB micro-lasers versus incident pump power for OM transmission of 1.5%. Spectrum of the sandwich-type micro-laser at pump power of 5.42 W is also shown.

Download Full Size | PDF

In order to analyze beam quality of the micro-laser, a lens with a 70-mm focal length was used to focus the output beam and then the spatial profile of the focused beam was recorded with a Pyrocam III camera (Ophir Optronics Ltd.). The beam diameter at various distances from the focusing lens was calculated by the 4-sigma method. By fitting these data to the Gaussian beam propagation expression, the beam quality factors $M_X^2$ and $M_Y^2$ of the sandwich-type micro-laser were estimated to be about 1.32 and 1.45 at pump power of 1.84 W, respectively, as shown in Fig. 3(a). When pump power increased to 5.42 W, $M_X^2$ and $M_Y^2$ were changed to be about 2.78 and 2.86, respectively, as shown in Fig. 3(b), which may be originated from the thermal effect. Furthermore, it can be seen from the inset of Fig. 3(b) that the ratios of $M_X^2$ to $M_Y^2$ were almost near to 1 at various pump powers. With the increment of pump power, the thermal effect of gain medium becomes more serious and then the thermal focal length of gain medium is shortened, which causes the decrement of the fundamental laser waist radius w₀ of the micro-laser [13]. When pump power increased from 1.84 to 5.42 W, w₀ of the sandwich-type micro-laser decreased from about 70 to 60 μm. It can be seen that at higher pump power, the waist radius of fundamental laser in the gain medium was closer to that (about 50 μm) of pump laser. Then, the reabsorption loss can be reduced and more energy can be extracted, which results in the higher slope efficiency as shown in Fig. 2. For the monolithic Er:Yb:LuAB micro-laser, the beam quality factors $M_X^2$ and $M_Y^2$ at pump power of 4.5 W were measured to be about 4.0 and 3.8, respectively. Therefore, the use of the sapphire crystal can also enhance the output beam quality of the micro-laser.

 

Fig. 3 Beam quality factor M² of the sandwich-type micro-laser at 1543 nm: (a) beam diameter as a function of the distance from the focusing lens at pump power of 1.84 W. (b) variation of $M_X^2$ with pump power. Ratio of $M_X^2$ to $M_Y^2$ versus pump power is also shown in the inset.

Download Full Size | PDF

When a sapphire crystal with OM transmission of 2% around 1520 nm was used, in which transmission was higher than 3% at wavelength longer than 1540 nm, 1521 nm laser was realized in the sandwich-type micro-laser, as shown in Fig. 4. Maximum output power of 1.2 W was obtained at pump power of 5.42 W and the threshold was about 1.15 W. When the pump power was higher than 4 W, slope efficiency η with respect to incident pump power was 33%. Taking into account the 90% transmission of IM and about 95% absorption efficiency of the microchip at 976 nm, slope efficiency with respect to absorbed pump power was about 40%. The maximum output power and slope efficiency obtained in the sandwich-type micro-laser are larger than 800 mW and 16% of cw diode-pumped monolithic Er:Yb:YAB micro-laser, respectively [9]. For the same incident pump power, the beam quality factors of the 1521 nm laser were similar to those of the 1543 nm laser of the sandwich-type micro-laser.

 

Fig. 4 1521 nm cw output power of the sandwich-type sapphire/Er:Yb:LuAB/sapphire micro-laser versus incident pump power for OM transmission of 2.0%. Spectrum of the sandwich-type micro-laser at pump power of 5.42 W is also shown.

Download Full Size | PDF

To investigate pulse performances of the passively Q-switched sandwich-type micro-laser, a 0.5-mm-thick uncoated Co²⁺:Mg_0.4Al_2.4O₄ crystal with an initial transmission of about 99% around 1520 nm was placed between the Er:Yb:LuAB microchip and the sapphire crystal with OM transmission of 2%. Like Co²⁺:MgAl₂O₄ crystal, Co²⁺:Mg_0.4Al_2.4O₄ crystal is a novel saturable absorber for 1.5-1.6 μm Q-switching lasers [14]. The experimental setup is shown in Fig. 5(a). When the pump power was 5.42 W, maximum average output power of 420 mW was obtained. Pulse profiles of the passively Q-switched micro-laser were measured by a 2 GHz InGaAs photodiode connected to an oscilloscope with bandwidths of 1 GHz (DSO6102A, Agilent). Pulse train and oscilloscope trace at pump power of 5.42 W are shown in Figs. 5(b) and 5(c), respectively. Pulse repetition frequency was about 69 kHz and pulse duration was about 12.6 ns. Pulse-to-pulse amplitude fluctuation and interpulse time jittering were less than ± 2% and ± 4.5%, respectively. Pulse energy was about 6.1 μJ. The beam quality factors were about 2.95 and 2.71, respectively, as shown in Fig. 5(d).

 

Fig. 5 (a) Experimental setup of the passively Q-switched sandwich-type micro-laser. Pulse train and oscilloscope trace of the micro-laser at pump power of 5.42 W were shown in (b) and (c), respectively. (d) shows the beam diameter of the micro-laser as a function of the distance from the focusing lens at pump power of 5.42 W.

Download Full Size | PDF

4. Conclusion

By tightly pressing two sapphire crystals and a Er:Yb:LuAB microchip together, and directly depositing the cavity mirrors onto the outside surfaces of the sapphire crystals, a sandwich-type sapphire/Er:Yb:LuAB/sapphire micro-laser was fabricated. The sapphire crystals act as heat diffuser and then the thermal effect of gain medium can be reduced in this sandwich-type micro-laser compared with the monolithic micro-laser. For the sandwich-type micro-laser, 1543 nm laser with maximum output power of 1.17 W and slope efficiency of 33% was obtained. However, for the monolithic micro-laser, the maximum output power and slope efficiency were only 460 mW and 17%, respectively. Furthermore, 1521 nm cw laser with maximum output power of 1.2 W and slope efficiency of 33%, and passively Q-switched pulse laser with 6.1 μJ energy, 12.6 ns duration, and 69 kHz repetition frequency were also realized in the sandwich-type micro-laser.