1 Introduction

Soliton formation is an interesting topic in photonics field that has been aggressively investigated in recent years due to the wide applications in optical communications, sensors and instrumentations. To date, most of the reported works were focused on soliton mode-locked lasers that operate in the bright pulse regime [1–3]. Besides bright pulse regime, there are other types of mode-locking pulses that operate in dark pulse regime, called dark solitons. Dark solitons are solutions of the nonlinear Schrödinger equation (NLSE) [4] and also further identified as solutions of the complex Ginzburg–Landau equation (CGLE) [5, 6]. The formation of dark pulses can be classified into three mechanisms, which are NLSE dark pulse [7–9], cubic–quintic nonlinear Schrödinger equation (CQNLSE) dark pulse [10] and domain-wall (DW) dark pulse [10, 11]. NLSE dark pulses relies on the change of the linear refractive index coefficient and occur under a normal dispersion cavity. When the incident field in the cavity experiences strong non-Kerr nonlinearities, it is possible to generate CQNLSE-type dark pulse in anomalous dispersion. Unlike the NLSE-type dark pulse that is dominated by Kerr nonlinearities, CQNLSE-type dark pulse is dominated by fundamental and higher-order nonlinearities, such as third-order (cubic) and fifth-order (quintic) nonlinearities. On the other hand, the mathematical model of optical DW has well established [12] and the domain wall between traveling waves is also well agreed with GLE [13]. The fundamental concept of DW dark pulse is based on two or more lasing at different wavelengths oscillating and causing topological defects in temporal domain. Compared to NLSE-type dark pulse, DW dark pulses can operate in both normal dispersion and anomalous dispersion cavity. Moreover, DW-type dark pulse implementation can be more compact compared to CQNLSE-type dark pulse, as it does not require any highly nonlinear medium. In our earlier works, we have demonstrated NLSE dark pulse [9], switchable DW dark pulse and CQNLSE dark pulse [10] based on nonlinear polarization rotation (NPR) technique. Furthermore, dark pulse generation has also been demonstrated using different types of saturable absorber, such as carbon nanotubes (CNT) [14] and graphene oxide [15].

In the post-graphene discovery, transition metal dichalcogenides (TMD) materials have attracted great attention as next generation 2D materials. Among the TMD materials, molybdenum disulfide (MoS2) has shown great potential in photonics applications. MoS2 exhibits 1.8 eV direct bandgap, whereas graphene possesses zero bandgap. Therefore, the MoS2 optical-range bandgap provides higher absorption coefficient in comparison with graphene. Besides, MoS2 also presents remarkable performance over graphene in terms of modulation depth and photoresponsivity [16]. These properties have led to the exploration of MoS2 to function as saturable absorber. In this paper, we have experimentally demonstrated the dual-wavelength DW dark pulse based on MoS2. To the best of our knowledge, this is the first demonstration of dual-wavelength DW dark pulse generation based on MoS2 film saturable absorber.

2 MoS2 film saturable absorber fabrication

The bulk and few-layer MoS2 possess discernible properties. Thus, it is imperative to exfoliate the few-layer material from the bulk material. Several approaches including micromechanical cleavage, chemical vapor deposition (CVD) and solution based methods are available to exfoliate the bulk materials. Among these methods, liquid phase exfoliation (LPE) method is considered to be the most efficient and cost effective approach since it does not require any post-chemical treatments of the prepared solvents and permits large scale production.

In brief, the bulk MoS2 is added with dimethylformamide (DMF) solvent with an initial concentration of 5 mg/ml. The solution is ultrasonicated for 24 h to break the weak van der Walls force between the atomic layers using a bath ultrasonicator. The suspension is then centrifuged at 3000 rpm for 60 min, and the supernatant is extracted from the suspension to obtain the few-layer MoS2. The supernatant solution is diluted to 10 vol %, and the linear optical absorption characterization is carried out using spectrophotometer (JASCO V-670 UV–Visible–NIR). In Fig. 1, the four observed peaks at ~664 nm (a), ~605 nm (b), ~439 nm (c) and ~395 nm (d). a and b transitions originate from the interband excitonic transitions which indicates 2H polytype, whereas c and d result from the transitions between higher density of states regions of the MoS2 band structure. This result confirms the presence of few-layer MoS2 nanosheets.

Fig. 1
figure 1

Linear optical absorption characterization of few-layer MoS2 suspension diluted to 10 vol %

Full size image

To confirm the crystalline nature of few-layer MoS2 nanosheets, we further characterized the X-Ray Diffraction (XRD) using Bruker D8 Advance equipment at an excitation wavelength of 1.5406 Å. From Fig. 2, it is clearly evident that all the labeled peaks of the bulk MoS2 can be readily indexed to rhombohedral MoS2 (JCPDS. No:06-0097). The few-layer MoS2 showed a high [0 0 2] orientation and disappearance of some characteristic peaks, which attests that the bulk MoS2 had been successfully exfoliated to few-layer MoS2.

Fig. 2
figure 2

The XRD pattern of the bulk and few-layer MoS2

Full size image

Next, the Raman spectrum is characterized by using Renishaw inVia confocal Raman microscope with an excitation wavelength of 488 nm and 1.5 mW power. From Fig. 3, the Raman Spectrum confirms the exfoliation of few-layer MoS2 since there is a peak shift for few-layer MoS2 compared to bulk MoS2. The Raman peaks at 383 and 407 cm−1 correspond to the in-plane (\(E_{2g}^{1}\)) and out-of plane (A 1g ) vibrational modes of the bulk MoS2. The frequency difference between these modes will give the details of the number of layers. From this analysis, it can be confirmed that the average number of layers should be around ~3–4 layers, since each MoS2 layer possess thickness of 0.65 nm.

Fig. 3
figure 3

Raman spectroscopy characterization of bulk and few-layer MoS2

Full size image

The next step is the fabrication of MoS2 film. Firstly, the few-layer MoS2 solution is placed in a bath sonicator for 10 min. After that, 10 ml of the few-layer MoS2 solution is added with 200 mg of polyvinyl alcohol (PVA) dissolved in 20 ml of deionised (DI) water (concentration of 10 mg/ml). The solution mixture, which is now 30 ml, is stirred using magnetic stirrer and heated continuously at a fixed temperature of 80 °C till the solution is reduced to approximately 10 ml. This process takes approximately 7 h to finish. This is followed by drying the remaining solution on a glass substrate in an oven at 80 °C for another 4 h.

3 Experiment arrangement

The experimental setup of the proposed mode-locked YDFL with MoS2 as SA is shown in Fig. 4. The ring cavity consisted of a 974 nm center wavelength laser diode, with maximum output power of 600 mW and launch power of 142.1 mW (Oclaro Model LC96A74P-20R) connected to a 980/1060 nm wavelength division multiplexing (WDM) coupler. One port of the WDM was fusion spliced to a 70 cm YDF gain medium (DF1100 Fibercore) that had peak absorption of 1300 dB/m at 977 nm, and another port was attached to the 90 % port of an optical coupler (OC). The laser output emerging from the YDF was fusion spliced to the isolator to ensure a unidirectional propagation of the laser inside the cavity. The output of the isolator was subsequently connected to the MoS2 thin film (sandwiched between the fiber ferrules), which was connected to a 90/10 fused bi-conical optical coupler (OC1). The fraction of the propagating signal tapped out by the 10 % port of the coupler was connected to 50/50 optical coupler (OC2) in order to monitor the performance of the mode-locked YDFL simultaneously via output 1 and output 2. Both outputs were used to observe the mode-locked spectrum and pulse train by using optical spectrum analyzer (OSA) (YOKOGAWA AQ6373) with a resolution of 0.02 nm and a high-speed photodetector attached to digital oscilloscope (YOKOGAWA DLM2054), respectively. Apart from that, an optical power meter (ILX LIGHTWAVE OMM-6810B) and RF spectrum Analyzer (ANRITSU MS2683A) with frequency range between 9 kHz and 7.8 GHz were also used to measure average output power of the system and the stability of the pulse repetition rate, respectively.

Fig. 4
figure 4

The experimental setup of the proposed mode-locked YDFL with MoS2 SA

Full size image

4 Results and discussion

The generation of dark pulse in the proposed cavity is based on dual-wavelength DW. Under the optimum polarization orientation, dark pulse was achieved as pump power is increased above the pumping power threshold of 175.4 mW. The DW dark pulse train is sustained until the maximum pump power of 242.5 mW. Figure 5 shows the typical DW dark pulse emission of the proposed laser at pump power of 242.5 mW, which operates at fundamental repetition rate of 17.4 MHz. Figure 6 depicts the output spectrum of the dual-wavelength DW dark pulse. The dual peaks are located at 1037 and 1039 nm. When the lasing oscillates simultaneously with dual peak wavelengths in the cavity, it leads to the emergence of topological defects in temporal domain and thus forms a narrow intensity dip in the strong CW laser emission background as shown in Fig. 5.

Fig. 5
figure 5

Emission of dark pulse train at pump power of 242.5 mW

Full size image
Fig. 6
figure 6

Output spectrum at pump power of 242.5 mW

Full size image

Figure 7 shows the output power and pulse energy of the proposed dark pulse cavity against pump power. Output power shows directly proportional relationship against pump power. Output power increases from 0.4 to 0.8 mW by varying the pump power from 175.4 to 242.5 mW. Throughout the DW dark pulse operation in the cavity, the laser efficiency is around 0.6 %. The dependence of the pulse energy can be seen to increase almost linearly with the pump power. Pulse energy increases from 23 to 46 pJ as pump power increases from 175.4 to 242.5 mW. The stability of the DW dark pulse is then investigated by RF spectrum analyzer. As shown in Fig. 8, the laser operates at fundamental frequency of 17.4 MHz, which tallies well with the pulse repetition rate as shown in Fig. 5.

Fig. 7
figure 7

Output power and pulse energy of the proposed DW dark pulse laser

Full size image
Fig. 8
figure 8

DW dark pulse RF spectrum of the proposed laser at pump power of 242.5

Full size image

The SNR of the laser is around 50 dB, which indicates that the laser is operating in a stable condition. Even though there are two operating wavelengths in the cavity, only a single frequency component is observed, which confirms a typical DW operation whereby the mutual coupling of two wavelengths generates a single frequency component.

5 Conclusion

We have demonstrated dual-wavelength DW dark pulse operation based on MoS2 film saturable absorber. The DW dark pulse exhibited two peaks, which are located at 1037 and 1039 nm. The pulse repetition rate was maintained at 17.4 MHz throughout the DW dark pulse operation. The highest pulse energy of 46 pJ was obtained at pump power of 242.5 mW. Besides that, DW dark pulse operation was demonstrated at high stability with SNR of 50 dB.