1. INTRODUCTION

Two-dimensional (2D) nanomaterials have attracted much attention recently due to their remarkable electronic and optical properties. A well-known 2D nanomaterial, graphene, has been widely investigated as an excellent saturable absorber (SA) for the wavelength range between 0.8 and 3 μm due to its unique zero bandgap, high thermal stability, and the excellent saturable absorption properties [1–5]. With the wide application of graphene-based SAs, other graphene-like 2D materials also have attracted extensive attention. Recently, molybdenum disulfide (${\mathrm{MoS}}_2$), a new type of 2D material, transition metal dichalcogenides [6], has received attention due to its thickness-dependent electronic and optical properties. ${\mathrm{MoS}}_2$ is composed of a hexagonal structure of molybdenum atoms sandwiched between two layers of chalcogen atoms [6]. Reference [7] demonstrated that the ${\mathrm{MoS}}_2$ has a nonlinear optical response stronger than that of graphene. ${\mathrm{MoS}}_2$ further has an indirect bandgap of 1.29 eV in bulk form and a direct bandgap of $\sim 1.85\mathrm{eV}$ in a single layer [8,9]. Because of the direct bandgap of $\sim 1.85\mathrm{eV}$ in single layer, ${\mathrm{MoS}}_2$ has a potential application at visible wavelengths [10]. In addition, the few-layer ${\mathrm{MoS}}_2$ could be regarded as a promising broadband SA for pulsed lasers and has been demonstrated in [11]. Using broadband few-layer ${\mathrm{MoS}}_2$ as SAs, passively $Q$-switched and ultrafast lasers have been realized. The 1054 nm passively mode-locked Yb-doped fiber laser and the 1.5 μm passively mode-locked erbium-doped fiber laser were realized by Zhang et al. [12] and Xia et al. [13], respectively. A passive $Q$-switching and $Q$-switched mode-locking Tm:CLNGG laser at a wavelength of 2 μm has been realized [14], from which the minimum pulse duration, maximum output power, and maximum pulse energy of 4.84 μs, 62 mW, and 0.72 μJ were obtained, respectively. In addition, Lou et al. reported a fewer-layer ${\mathrm{MoS}}_2$ dual-wavelength $Q$-switched Yb:LGGG laser at 1025.2 and 1028.1 nm [15]. A maximum average output power of 0.6 W was obtained, corresponding to single pulse energy up to 1.8 μJ.

$\mathrm{Tm}\text{:}{\mathrm{GdVO}}_4$ offers many favorable properties [16,17]. The absorption band of that is considerably stronger and broader (770–820 nm) than that in YAG and YAP. Furthermore, the thermal conductivity of ${\mathrm{GdVO}}_4$ is $11.7{\mathrm{Wm}}^{-1}{\mathrm{K}}^{-1}$ at 300 K, larger than that in YAP and YLF, which is more efficient cooling the crystal. A maximum output power of 2.6 W at 1910 nm in a diode-pumped continuous wave (CW) $\mathrm{Tm}\text{:}{\mathrm{GdVO}}_4$ laser had been reported by Cerny et al. [18]. An acousto-optical (AO) $Q$-switch output has also been proved in $\mathrm{Tm}\text{:}{\mathrm{GdVO}}_4$ [19]. However, a diode-pumped passively $Q$-switched $\mathrm{Tm}\text{:}{\mathrm{GdVO}}_4$ laser has not been reported.

In this paper, the ${\mathrm{MoS}}_2$ SA is fabricated by thermally decomposing ammonium thiomolybdate dip coating the mica and successfully realizing a stable passively $Q$-switched $\mathrm{Tm}\text{:}{\mathrm{GdVO}}_4$ laser at 1902 nm. A maximum output power of 100 mW was achieved at 1902 nm, corresponding to the maximum single pulse energy of 2.08 μJ.

2. FABRICATION OF ${\mathrm{MoS}}_2$ SA

Figure 1 schematically illustrates a convenient and cost-effective approach by thermally decomposing the ammonium thiomolybdate [20]. ${({\mathrm{NH}}_4)}_2{\mathrm{MoS}}_4$ (Alfa Aesar, purity of 99.99%; 0.05 g) dissolved into 5 mL organic solvent of dimethylformamide (DMF) to form a uniform and stable solution through sonication for 30 min. We deposited a thin ${({\mathrm{NH}}_4)}_2{\mathrm{MoS}}_4$ film onto the fresh surface of mica of just dissociation with a dropper and used a spin coater to ensure uniformity of solution on the mica. The sample should be placed in the hot zone of the furnace for drying naturally horizontally. After this, the annealing process was carried out in the quartz tube. First, the quartz tube was pumped to low pressure accompanied by a flow of ${\mathrm{H}}_2$ and Ar (20/80 sccm). Subsequently, a temperature of 500°C was kept for an hour in the growth process. Finally, the furnace was quickly cooled to room temperature by opening the furnace.

 

Fig. 1. Schematic illustration of the formation of ${\mathrm{MoS}}_2$.

Download Full Size | PDF

The XRD results of exfoliated ${\mathrm{MoS}}_2$ NPs in Fig. 2(a) are in good fit the hexagonal structure of ${\mathrm{MoS}}_2$ (PDF Card 04-0880). Several diffraction peaks are easily indexed, assigned to (110), (100), (102), and (106) reflection.

 

Fig. 2. (a) XRD pattern. (b) SEM images of as-grown ${\mathrm{MoS}}_2$ on mica substrates. (c) Trilayered ${\mathrm{MoS}}_2$ Raman spectra.

Download Full Size | PDF

Figure 2(b) exhibits the SEM image of the ${\mathrm{MoS}}_2$ membrane after the decomposition of ammonium thiomolybdate on mica substrate. A large area continuous ${\mathrm{MoS}}_2$ layer is obtained. As shown in the SEM image, the surface of the sample is uniform and compliant.

Figure 2(c) shows the Raman spectroscopy on trilayered ${\mathrm{MoS}}_2$. The characteristic peaks of the two samples at 382.9 and $405.4{\mathrm{cm}}^{-1}$ are assigned to the ${\mathrm{E}}_{2\mathrm{g}}^1$ and ${\mathrm{A}}_{1\mathrm{g}}$ modes of ${\mathrm{MoS}}_2$, respectively, while the two characteristic peaks ${\mathrm{E}}_{2\mathrm{g}}^1$ and ${\mathrm{A}}_{1\mathrm{g}}$ of the bulk ${\mathrm{MoS}}_2$ occur at 380 and $410{\mathrm{cm}}^{-1}$ [21]. Compared with the thickness-related distance value of bulk ${\mathrm{MoS}}_2$ between the two peaks is $\sim 30{\mathrm{cm}}^{-1}$, the thickness-related distance value of the trilayered ${\mathrm{MoS}}_2$ between the two peaks in $\sim 22.5{\mathrm{cm}}^{-1}$. The frequency difference of the two modes can be used to determine the layer thickness of ${\mathrm{MoS}}_2$. In our result, the thickness-related distance of $\sim 22.5{\mathrm{cm}}^{-1}$ corresponds to three or four layers of ${\mathrm{MoS}}_2$ [22,23].

3. EXPERIMENT SETUP

A schematic of the experimental arrangement is shown in Fig. 3. A fiber-coupled diode laser was used as the pump resource, whose emission wavelength was around 803 nm at 25.4°C. The pump core diameter and the numerical aperture were 400 μm and 0.22, respectively. By using a couple of convex lens (1:0.5), the pump beam was focused on the laser crystal with the radius of 100 μm. The laser crystal was $a$-cut $\mathrm{Tm}\text{:}{\mathrm{GdVO}}_4$ crystal with the ${\mathrm{Tm}}^{3+}$ doping concentration of 0.5 at. % and the dimension of $3\mathrm{mm}\times 3\mathrm{mm}\times 3\mathrm{mm}$, which both surfaces were antireflection coating at the pump and laser wavelengths. It was wrapped with an indium foil and then mounted on a water-cooled copper crystal holder with the cooling water temperature set at 15°C to preserve the laser crystal from thermal fracture. The absorption efficiency of the diode pump by the crystal was 77.86%. The input mirror M1 was a plane-concave mirror with radius of 100 mm, which was high-transmission coated at 780–810 nm and high-reflection coated at 1900–2000 nm. A flat mirror M2 with the transmission of 2% was used as the output coupler (OC). A compact concave-plane resonator was designed to keep the mode matching in crystal between the pump beam and the fundamental resonant mode. The laser pulse trains was recorded by a 1 GHz digital oscilloscope (Tektronix DPO 4104) and a fast photodiode detector (ET-5000) with a rising time of 250 ps. The average output power was measured by a laser power meter (30A-SH-V1, made in Israel).

 

Fig. 3. Schematic configuration of the $Q$-switched $\mathrm{Tm}\text{:}{\mathrm{GdVO}}_4$ laser.

Download Full Size | PDF

4. RESULTS AND DISCUSSION

The average output powers of the CW and passively $Q$-switched $\mathrm{Tm}\text{:}{\mathrm{GdVO}}_4$ laser versus absorbed pump power are shown in Fig. 4. In the CW operation, the laser threshold pumping power was 2.12 W without inserting ${\mathrm{MoS}}_2$ SA. When the absorbed pump powers were increased to 3.08 W, the maximum output power of 304 mW was obtained with corresponding slope efficiency of 19%. In the passively $Q$-switched operation, the ${\mathrm{MoS}}_2$ SA was positioned close to the M2. The radius of the laser beam at the ${\mathrm{MoS}}_2$ SA was calculated to be about 125 μm by ABCD matrix. By carefully aligning the laser cavity, the passively $Q$-switched pulse train was searched. The laser threshold pumping power was 2.41 W. The maximum output power of 100 mW was obtained with a corresponding slope efficiency of 7.3% with the absorbed pump powers of 3.08 W. To avoid damage to the laser crystal and ${\mathrm{MoS}}_2$, the pump power was not increased more than 4.0 W.

 

Fig. 4. Average output power of CW and passively $Q$-switched versus the absorbed pump power.

Download Full Size | PDF

The CW and $Q$-switched pulse spectrum were measured by an optical spectrum analyzer (AvaSpec-NIR256–2.2-RM) with a resolution of 10 nm. Figure 5 shows the CW and $Q$-switched spectra under the absorbed pump power of 3.08 W. We can see the central wavelengths were 1938 and 1902 nm. The central wavelength of $Q$-switching was shorter than that of the CW, which was attributed to the stimulated emission cross section in the $Q$-switched operation becoming a key factor because the energy stored in the crystal far exceeds the CW operation threshold [24].

 

Fig. 5. Output spectra from $\mathrm{Tm}\text{:}{\mathrm{GdVO}}_4$ lasers in CW operation and passively $Q$-switched operation.

Download Full Size | PDF

The pulse width and the repetition rate of the passively $Q$-switched operation as the function of the absorbed pump power are shown in Fig. 6. It can be seen that the repetition increases as the pulse width decreases rapidly. In Fig. 6, we can see that, when the absorbed pump power is increased from 1.88 to 3.08 W, the pulse duration decreased from 2 to 0.8 μs with the pulse repetition rate increasing from 25.58 to 48.09 kHz. The maximum repetition rate of 48.09 kHz and the minimum pulse width of 0.8 μs were achieved under the absorbed pump power of 3.08 W, corresponding to the maximum single pulse energy of 2.08 μJ.

 

Fig. 6. Pulse width and repetition rate as a function of absorbed pump power.

Download Full Size | PDF

Figure 7 gives a recorded typical oscilloscope pulse train with the time of 40 and $2\mathrm{\mu s}/\mathrm{div}$. The experimental result was obtained at the absorbed pump power of 3.08 W. The pulse-to-pulse amplitude fluctuation of the $Q$-switched pulse train was less than 5%. In order to protect the laser crystal and ${\mathrm{MoS}}_2$ from damage, we no longer increased the pump power. In future experiments, we will continue to optimize cavity type to obtain excellent $Q$-switching stability and $Q$-switch mode-lock.

 

Fig. 7. Passively $Q$-switched pulse at 40 and $2\mathrm{\mu s}/\mathrm{div}$.

Download Full Size | PDF

5. CONCLUSIONS

In conclusion, using a ${\mathrm{MoS}}_2$ as a SA, which was fabricated by thermally decomposing the ammonium thiomolybdate, a passively $Q$-switched $\mathrm{Tm}\text{:}{\mathrm{GdVO}}_4$ laser was realized. A maximum output power of 100 mW and pulse duration of 0.8 μs were achieved at 1902 nm, corresponding to the slope efficiency of 7.3%, and the single pulse energy was 2.08 μJ. As far as we know, this is the first report on diode-pumped passively $Q$-switched $\mathrm{Tm}\text{:}{\mathrm{GdVO}}_4$ lasers with the ${\mathrm{MoS}}_2$. With further optimization of the ${\mathrm{MoS}}_2\text{-}\mathrm{SA}$, the higher output power and CW mode-locking laser performance can be anticipated.