1. Introduction

Q switching is an important and major technique for generating short laser pulses in μs and ns time durations. Q-switched diode-pumped solid state lasers (DPSSLs) with high peak power and short pulse width have attracted a great deal of attention for applications in laser communication, remote sensing, scientific research and medicine. In general, the generation of passively Q-switched laser pulses depends strongly on the availability of suitable saturable absorbers (SAs). From the viewpoint of materials, SAs used so far can be grouped into three categories, namely transition-metal-ion-doped materials, semiconductors and nanosheets materials. Transition-metal-ion-doped SAs mainly include Cr⁴⁺-doped host crystals like Cr:YAG [1], Cr:GSGG [2], Cr:YSO [3], Cr:SFS or Cr:CGS [4], but also Cr²⁺-, V³⁺- and Co²⁺-doped crystals like Cr:ZnSe, Cr:ZnS [5], Cr:CdMnTe [6], V:YAG [7], Co:MALO and Co:LMA [8]. SESAMs (semiconductor saturable absorber mirrors) [9] are the representative of semiconductor SAs. Both types of conventional SAs have specific advantages, but the same drawback: their operating wavelengths are limited to particular spectral bands. Moreover, the transition-metal-ion-doped crystals as well as semiconductors require specific and expensive fabrication techniques. Compared with transition-metal-ion-doped materials and SESAMs, the third category of SAs, namely the newly discovered nanosheets materials like CNTs (carbon nanotubes [10, 11]), graphene [12–14] and TIs (topological insulators like Bi₂Se₃ and Bi₂Te₃ [15–17]) now appear more and more attractive because of ultra-broadband saturable absorption, and relatively easy and low cost fabrication processes. TI-based SAs, more particularly, have attracted a lot of attention in the recent several years. For example, in 2012, Zhao et al. reported an efficient ultra-short pulse generation operating at 1558 nm by using Bi₂Te₃ [16] and one year later we demonstrated a Q-switched Yb³⁺-doped fiber laser with Bi₂Se₃ [15]. However, this does not mean that they have already reached the maturity of the former materials. CNTs, for instance, were reported to have cluster-induced losses and suitable diameter control of the CNT is often demanded for broadband saturable absorption [13]. Relatively low modulation depth and damage threshold could prohibit the generation of large pulse energies using graphene, and in the case of TIs, some important optical properties are still unknown. Therefore, these broadband materials seem promising but further works are necessary for a real development. Very recently, a new kind of 2D nanosheets material, MoS₂ (Molybdenum sulfide), has been found to be promising for optoelectronic applications [18–22]. Wang et al. [20] firstly observed the saturable absorption phenomenon of MoS₂ at 800 nm. In 2014, Zhang et al. [21] investigated the 1054 nm passively mode-locked laser operation of an Yb³⁺-doped fiber laser, and Woodward et al. [22] demonstrated the Q-switched laser operation of an Yb³⁺-doped fiber laser at 1068 nm. Liu et al. [23] also recently reported a femtosecond pulsed Er³⁺-doped fiber laser operating at 1.55 μm and Huang et al [24] the Q-switched laser operation of a similar type of fiber laser by using MoS₂, the latter giving a more deeper explanation about the origin of the broadband saturable absorption in this type of material. At present, the study of these MoS₂-based SAs are just at the beginning and further investigations for entirely exploiting their saturable absorption properties are still very necessary, especially in the field of SSLs based on bulk crystals in order to achieve higher pulse peak powers and pulse energies than with fibers.

Nd:YAP (YAlO₃) is a robust laser crystal with thermo-mechanical and laser properties comparable to Nd:YAG. The effective emission cross section of the strongest emission line in Nd:YAP occurs at 1079.5 nm for light polarized along the c crystallographic axis and is comparable to that found for Nd:YAG at 1064 nm [25]. Moreover, the Nd:YAP 1079.5 nm laser line is even broader than the 1064 nm laser line of Nd:YAG, which is more favorable for short pulse generation.

In this paper, we demonstrate the first diode-pumped and Q-switched Nd:YAP laser at 1079 nm using a few-layer MoS₂ material deposited on a BK7 glass substrate as saturable absorber. By inserting the MoS₂ saturable absorber into a plano-concave Nd:YAP laser cavity, we could obtain a stable Q-switched laser operation with laser pulses of 1.11 μJ and 227 ns at a repetition rate of about 232 kHz at an absorbed diode pump power of 4.45 W.

2. Fabrication and characterization of MoS₂ SA

The synthesis of the few-layer MoS₂ was performed by using a liquid-phase exfoliation method. The bulk MoS₂ was firstly put into a dimethyl formamide (DMF) solution and after a 20-hour sonication the few-layer MoS₂ suspension was centrifuged for 30 minutes at 1000 rpm to remove the residual bulk MoS₂.

The bulk MoS₂ and the as-prepared few-layer MoS₂ were both characterized by X-ray diffraction (XRD). The results are shown in Fig. 1(a). All the labeled peaks of the bulk MoS₂ can be easily indexed and assigned to rhombohedral MoS₂ (JCPDS NO. 06-0097). The bulk MoS₂ was successfully exfoliated because the XRD pattern of the few-layer MoS₂ showed a high [002] orientation and some characteristic peaks disappeared. The Raman spectra shown in Fig. 1(b) were also recorded to estimate the thickness of the few-layer MoS₂ sample. The two characteristic peaks $E_{2g}^1$ and $A_{1g}$ of the bulk MoS₂ occur at 380 and 410 cm⁻¹. Compared with the bulk MoS₂, the few-layer MoS₂ sample shows a red shift of the $E_{2g}^1$ peak and a blue shift of the $A_{1g}$ peak, respectively. This implies a successful exfoliation of the bulk MoS₂ with a thickness of 1-4 layers [26]. Finally, atomic force microscopy (AFM) image was registered for characterizing the thickness of the as-prepared few-layer MoS₂ (see Fig. 1(c)). The height profile diagram (see Fig. 1(d)) indicates the MoS₂ nanosheets are about 3 layers since the thickness of single layer is 0.65 nm [27].

 

Fig. 1 (a) XRD and (b) Raman spectra of the bulk MoS₂ and few-layer MoS₂ samples; (c) AFM image and (d) height profile of an as-prepared few-layer MoS₂ sample.

Download Full Size | PDF

Following the synthesis of the few-layer MoS₂ suspension, it was proceeded to the following steps to simply transfer the MoS₂ nanosheets onto uncoated BK7 glass plates (0.5 mm in thickness). First, proper amounts of homogeneous MoS₂ solution were directly dripped onto the BK7 glass substrate. Second, the glass substrate was rotated at low speed to uniformly disperse the MoS₂ solution. Third, we dried it inside an oven with a constant temperature of 70°C for two hours. The transmission of the MoS₂ SA was measured using a Perkin Elmer Lambda 750 Spectrophotometer. As shown in Fig. 2, the few-layer MoS₂ sample gives rise to a transmission of 88.6% at 1079 nm. A blank BK7 glass was also measured with a flat transmission of around 92.7%. The final transmission of the MoS₂ SA itself could be deduced to be around 95.5%, which means a linear loss of about 4.5%. From Fig. 2, it is also demonstrated that the MoS₂ sample exhibits a broad and flat transmission in the whole near- to mid-infrared spectral range. Using the same kind of experimental set-up as in [24], we estimated that the saturation optical intensity of the sample used in our laser experiments should be larger than about 13 MW/cm².

 

Fig. 2 Transmission spectra of the blank BK7 and MoS₂-transferred BK7 glass substrates.

Download Full Size | PDF

3. Laser operation conditions

The laser experimental set-up is reported in Fig. 3. The pump source is an 18 W fiber-coupled continuous wave diode laser with a numerical aperture of 0.22 and a core diameter of 400 μm. The emitting wavelength varies from about 798 to about 803 nm with the increasing injected current. With a collimating doublet of 50-mm focal length and a focusing doublet of 40-mm focal length, the end face of the coupling fiber was imaged into the laser crystal with a minimum spot radius of about 160 μm. The laser cavity was a typical plano-concave cavity of about 12 mm.

 

Fig. 3 Schematic experimental set-up of the Q-switched Nd:YAP laser

Download Full Size | PDF

The laser gain medium was a nicely polished b-cut Nd:YAP crystal with a Nd³⁺ doping concentration of 0.8at.% and dimensions of 3x3x5 mm³ (5 mm in thickness). The front facet S1 of the Nd:YAP crystal was coated and used as input mirror by depositing dichroic coating with a high transmission (more than 98%) at the pump wavelength and a high reflection (more than 99.5%) at the laser wavelength. The rear facet S2 was coated with an anti-reflection coating at laser wavelength to reduce the Fresnel losses. An output coupler (OC) with a radius of curvature of 1000 mm and a broadband transmission of about 6.5% around 1079 nm was employed. The laser spot size on the OC is estimated to be about 105 μm. In order to alleviate the thermal lensing effect, the Nd:YAP crystal was wrapped inside an indium foil and mounted in a water-cooled copper holder maintained at a temperature of 18°C. During Q-switching operation, the as-fabricated MoS₂ SA was inserted into the cavity.

4. Laser results and discussion

For the b-cut Nd:YAP crystal, the highest emission peak is found at about 1079 nm with an emission cross section of 2.4 × 10⁻¹⁹ cm² [25]. A free-running continuous-wave laser experiment was first fulfilled and a maximum output power of about 5.1 W was obtained for an absorbed pump power of about 14 W without appearance of any roll-off. With a threshold absorbed pump power of about 0.74 W, it means a laser slope efficiency of about 38.4%. In these conditions, the laser emission was monitored and measured with a peak wavelength of 1079.57 nm by using an optical spectrum analyzer HP 70951B.

The BK7 glass plate with MoS₂ thin film was then inserted inside the laser cavity. By continuously increasing the pump power from the continuous-wave threshold and finely tilting the glass substrate, a stable Q-switched pulsed operation was achieved when the absorbed pump power exceeded about 2.16 W, thus for an incident pump power of about 2.87 W. Such a stable Q-switched laser operation could be maintained up to an incident pump power of 5.75 W, thus up to an absorbed pump power of 4.45 W. Beyond that limit, the output laser pulses became unstable probably because of increasing thermal effects and laser damage due to the poor thermo-mechanical properties of the MoS₂ layer [28]. The stable laser output versus absorbed pump power curve is reported in Fig. 4(a). A maximum output power of 260 mW is achieved with a slope efficiency of about 10.6%. We then used a digital oscilloscope (Tektronix TDS 1012, 100MHz) and a fast photodetector (Thorlabs, DET10A/M) to characterize the corresponding laser pulses. The pulse trains and single pulse profile shown in Fig. 4(b) are those obtained at the maximum pump power of 5.75 W. The pulse duration was measured to be about 227 ns and the corresponding pulse repetition rate was recorded to be 232.5 kHz. Figure 4(c) shows the evolution of the pulse repetition rate and of the pulse width. The pulse repetition rate almost linearly increased from 32 to 232.5 kHz, while the pulse width shortened from 580 to 227 ns. According to the recorded pulse repetition rate and pulse width, we can approximately calculate the pulse peak power and pulse energy (see Fig. 4(d)). The maximum single pulse energy was estimated to be 1.11 μJ with a corresponding peak power of 4.92 W.

 

Fig. 4 (a) Absorbed pump power versus output power for the Q-switching regime; (b) Single pulse profile of the Q-switched Nd:YAP laser. The inset shows the pulse train; (c) Evolutions of the pulse repetition rate and the pulse width with the absorbed pump power; (d) Evolution of the pulse peak power and the pulse energy with the absorbed pump power.

Download Full Size | PDF

Our results can be compared with that reported in Ref [18]. for another passively Q-switched SSL operating at 1064 nm and by using MoS₂ SA, i.e. a maximum output power of 227 mW, a pulse width of 970 ns, a pulse energy of 0.31 µJ and a repetition rate of 732 kHz. Therefore, we obtain here in this work about the same output power but significantly shorter and more energetic laser pulses at a smaller pulse repetition rate. Furthermore, if we compare our MoS₂-based Q-switched laser results with the various reported laser results based on other nanomaterials like CNTs, graphene and TIs (see in Table 1), our results can be considered as superior or at least comparable to the other ones. The present results can be certainly improved firstly by optimizing the laser cavity, by optimizing the modulation depth of the MoS₂ SA and by removing the generated heat in MoS₂ SA.

Table 1. Comparison between the laser results obtained by passively Q-switching various 1µm Nd-based SSLs with MoS₂ and other nanosheets SAs.

View Table

5. Conclusion

We have experimentally presented a passively Q-switched Nd:YAP solid state laser using MoS₂ as saturable absorber. An average output power of 260 mW associated with 227 ns laser pulses of 1.11 μJ (4.92 W peak power) at a repetition rate of 232.5 kHz is obtained for an absorbed pump power of 4.45 W. Therefore, these laser results definitely demonstrate that few-layer MoS₂ could be a very promising saturable absorber for the realization of compact and low cost short pulse Q-switched SSLs and also for the realization of ultra-short pulsed mode-locked SSLs.