1. Introduction

Pulsed mid-infrared (mid-IR) laser sources are of great demand in scientific and technological applications, such as photoelectrical countermeasures, environmental sensing, material processing, non-linear frequency conversion, and molecular spectroscopy. Saturable absorbers (SAs), as the most important part of optical-modulation technique, have attracted significant attention in recent years, especially the emerging of two-dimensional (2D) nano-materials, such as graphene [1–3 ] and layered transition-metal dichalcogenides (TMDs) [4,5 ]. Remarkable progress in graphene indicates their outstanding physical and structural properties, including ultrafast recovery time, controllable modulation depth, and easy fabrication. Nevertheless, the absence of a bandgap in graphene introduces the intrinsic limitation for photonic applications. Although the bandgap can be engineered through size quantization or electric fields [6,7 ], the unexpected defects are also accompanied due to these immature technology, rendering them insufficient for some operations. On the other hand, current researches on TMDs like MoS₂, MoSe₂, and WSe₂ mainly focused on the near-infrared region [4, 8, 9 ], which imply that the TMDs are unsuitable for mid-IR applications, although their bandgap can be controlled by introducing suitable defects.

Currently, black phosphorus (BP) atomic layers, as an alternative 2D nano-material, have been successfully fabricated and widely studied in various areas of electronics and photonics [10–13 ], owing to its tunable direct bandgap [14–16 ], high carries mobility (~1000 cm²/V.s) [17], large on/off ratios (>10⁵), and high in-plane anisotropy [18]. It has been demonstrated that its direct bandgap can be tuned from 0.35 eV (bulk) to 2.0 eV (monolayer) [15,19 ], which fill the present gap between the zero bandgap of graphene and wide bandgap of TMDs. This is particularly interesting for photonics, since it can be capable for broadband optical response with a wavelength range from ~600 nm to ~4 μm, depending on the number of layers of phosphorus. Up to now, the broadband nonlinear optical response of few-layered BP had been demonstrated at wavebands of 400 nm, 800 nm, 1562 nm, and 1930 nm [20,21 ], experimentally proved the potential applications in a wide spectral region. Using BP as SA, Q-switched/mode-locked laser pulses had been frequently reported at ~1.5 μm and ~1.9 μm based on Er/Tm-doped fiber lasers, respectively [22–26 ]. Very recently, our group demonstrated the 6.1 ps pulse generation from a BP mode-locked bulk laser for the first time at 1064 nm [27]. Besides, Ma et al. realized the stable Q-switched pulses with a pulse width of 620 ns in a Yb:CaYAlO₄ laser with the BP-based saturable-absorber mirror (SAM) at 1046 nm [28]. However, there is no report of any Q-switched/mode-locked bulk laser with BP-based SA in mid-IR region.

In this letter, a high-quality SAM was successfully fabricated with the multi-layered BP, prepared by liquid-phase exfoliation (LPE) technique. With the prepared BP-SAM, we experimentally demonstrated the mid-IR pulse generation at 2.4 μm from a Q-switched Cr:ZnSe laser for the first time to our best knowledge. The Q-switched Cr:ZnSe laser delivered a maximum average output power of 36 mW with the pulse duration of 189 ns at a repetition rate of 176 kHz, corresponding to a single pulse energy of 205 nJ. The modulation depth and saturation power intensity of our prepared BP absorber were measured to be 10.7% and 0.96 MW/cm², respectively. The results sufficiently validated that multi-layered BP could be used as an optical modulator for mid-IR pulsed laser sources.

2. Preparation and characterization of BP SAM

The LPE method was employed to produce high quality BP-based SAM in our experiment. The BP powder was prepared from the BP bulk crystal by grinding. Firstly, the powder was dispersed into the isopropyl alcohol (IPA) and ultra-sonicated for 3 hours. In order to remove the large-size phosphorous sheets, the dispersions were processed with the centrifugation at a speed of 1500 rpm for 20 minutes. Then the top half of the dispersions without sedimentation were collected for further use. The processed BP solution was dropped upon the output coupler (OC), which was coated with partial reflection at 2.2-2.6 μm (T = 4%) and used as the SAM substrate. Finally, the treated OC was dried under vacuum at room temperature for 24 hours and then soaked in alcohol followed by 5-min sonication to remove the IPA.

To confirm that the BP nano-platelets (NPs) have been successfully transferred on the OC, we performed a morphology analysis with an atomic force microscopy (AFM). Figure 1 shows the AFM image and the typical height profiles of BP NPs, which implied that the thickness of the transferred phosphorene sheets should be about 6-8 nm. Given that the thickness of 0.6 nm for single layer BP [29], the as-prepared phosphorene sheets are ~10 layers thick, which means that the band gap of our BP sheets is about 0.45 eV [15] and can respond to the photons with the wavelength shorter than 2.76 μm.

 

Fig. 1 AFM image and typical height profile of the transferred phosphorene sheets.

Download Full Size | PDF

To verify that the transferred material is BP, a Raman spectrum analysis was conducted using a 532 nm laser source, as plotted in Fig. 2(a) . The spectrum shows three characteristic Raman peaks located at the wavenumbers of 359.6 cm⁻¹, 437.3 cm⁻¹, and 463.6 cm⁻¹, corresponding to one out-of-plane vibration mode A_g ¹, two in-plane vibration modes B_2g and A_g ² of phosphorus atoms, respectively [30]. According to the previous researches, the A_g ¹ and A_g ² modes will shift toward each other with the increased thickness of BP sheets [20]. Compared with the typical Raman spectrum of bulk BP, the measured value of 104 cm⁻¹ between the A_g ¹ and A_g ² modes implied that the BP powder have been exfoliated down to be several layers in our study, which agrees well with the measured AFM image.

 

Fig. 2 (a) Raman spectrum of the prepared multi-layered BP film at room temperature; (b)Transmittance versus incident optical power intensity on phosphorene SA.

Download Full Size | PDF

We also measured the saturable absorption behavior of BP-based SAM at 2.4 μm, which was conducted by a home-made mid-infrared picosecond optical parametric oscillator (OPO) based on MgO:PPLN. It was synchronously pumped by a mode-locked laser system with a repetition rate of ~120 MHz and pulse duration of ~10 ps. Apart from the 4% of OC, the nonlinear transmission curve of the exploited BP SA is plotted in Fig. 2(b), which shows a clear increase in the transmittance with the increasing pump power intensity. The modulation depth (ΔT), non-saturable loss, and the saturation power intensity (I_s) of our BP SA were estimated to be about 10.7%, 1.6%, and 0.96 MW/cm², respectively. The relatively large modulation depth and low saturation power intensity confirms that the BP is a preferred SA for Q-switched bulk laser.

3. Laser experimental results and discussion

Transition-metal doped II-VI crystal like Cr:ZnSe are ideal mediums in mid-IR region since they can directly generate high-power mid-IR radiations. Therefore, the well-known Cr:ZnSe crystal was chosen as the gain medium to investigate the optical modulation ability of prepared BP-based SAM at 2.4 μm. The 4-mm-long Cr:ZnSe medium had a Cr²⁺ concentration of 6 × 10¹⁸ cm⁻³ and the dimensions of 5 × 2 mm², which was anti-reflection (AR) coated for the pump wavelength on both surfaces. For efficient thermal management, it was wrapped with indium foil and tightly mounted in a copper block cooled to 18°C by circulating water. Figure 3 sketches the BP Q-switched Cr:ZnSe laser experimental setup in a V-typed resonator. A CW Tm-fiber laser pump source (AP-Tm-1950-SM-10) was employed in our work, operating at a wavelength of 1950 nm. Using a focusing lens (f = 100 mm), the pump beam was focused into the Cr:ZnSe crystal with a spot radius of ~110 μm. The input flat and folded concave (R = 200 mm) mirrors were AR-coated from 1.95 μm to 2.05 μm and high reflectivity (HR) coated for the laser wavelength. The distances among mirrors were orderly optimized as 180 mm and 235 mm. With the propagation ABCD matrix theory, the mode radii on the gain medium and the BP SAM were calculated to be about 105 and 139 μm, respectively. Additionally, we expect to see smaller mode sizes on the both Cr:ZnSe and BP SAM as increasing pump power because of the strong thermal lensing effect of laser crystal.

 

Fig. 3 The schematic of Q-switched Cr:ZnSe laser setup based on BP-SAM.

Download Full Size | PDF

When inserting the BP-SAM in the cavity, we achieved the stable PQS operation by optimizing the position where oscillating laser beam located on the BP SAM carefully. Figure 4 depicts the Q-switched laser output characteristics in details, including average output power, pulse repetition rate, pulse width, single pulse energy, and laser pulses signal. The average output power and Q-switched laser emission spectrum were measured by a laser power meter (Fieldmax-II, coherent) and a spectrometer with the resolution bandwidth of 0.4 nm (WaveScan, APE Inc.), respectively. The laser pulses signal was recorded by a Tektronix DPL7104 digital oscilloscope (1 GHz bandwidth) and an infrared detector with a response time of 20 ns (PVI-2TE-5, VIGO System S.A.).

 

Fig. 4 (a) Average output power and single pulse energy versus incident pump power; (b) Evolutions of the pulse repetition rate and pulse width as increasing incident pump power; (c) A typical Q-switched pulse train (inset) and the temporal pulse profile, (d) Emission spectrum of Q-switched Cr:ZnSe laser.

Download Full Size | PDF

As shown in Fig. 4(a), the average output power increased almost linearly with the increasing incident pump power. Under an incident pump power of 1.99 W, a maximum average output power of 36 mW was obtained, corresponding to a slope efficiency of 27.3%. Note that once the incident pump power exceeded 2 W, the Q-switched operation became unstable, which brought about the sub-pulse and instability of pulse. However, the stable Q-switched regime could be realized again just by reducing the incident pump power below 2 W. Therefore, it could be concluded that the pulse instability was mainly caused by the over-saturation of BP absorber. Additionally, no optical damage or degradation was observed during any of the performed experiments, even at the highest available optical intensity in our work. Therefore, it is worthy to emphasize the high optical damage threshold of the exploited BP-based SAM.

Figure 4(b) plots the evolutions of the pulse repetition rate and width with the incident pump power. The stronger the pump intensity was, the easier the BP SA was bleached. Thus we expect to see the narrower pulse duration and larger pulse repetition rate as increasing incident pump power. Consequently, the pulse repetition rate and duration varied from 98 kHz to 176 kHz and 396 ns to 189 ns, respectively, with the range of incident pump power. With the measured average output power and pulse repetition rate, the single pulse energy was calculated and shown in Fig. 4(a). A maximum single pulse energy of 205 nJ was attained in our work. Figure 4(c) shows the temporal pulse profile with the minimum duration of 189 ns and a typical Q-switched pulse train with a repetition rate of 126 kHz, in which the pulse-to-pulse instability was found to be ~10%. The measured Q-switched emission spectrum was shown in Fig. 4(d). The emission central wavelength was located at 2411 nm, with the full width at half maximum (FWHM) of 33 nm.

5. Conclusion

In a summary, we have studied a high-quality BP-based SAM with the multi-layered BP, prepared by LPE method. The modulation depth and saturation power intensity of the prepared BP absorber were measured to be 10.7% and 0.96 MW/cm², respectively. With the BP-SAM, we experimentally demonstrated the mid-IR pulse generation at 2.4 μm from a BP Q-switched Cr:ZnSe laser for the first time to our best knowledge. The Q-switched Cr:ZnSe laser delivered a maximum average output power of 36 mW with the pulse duration of 189 ns at a repetition rate of 176 kHz, corresponding to a single pulse energy of 205 nJ. The results sufficiently validated that multi-layered BP could be used as an optical modulator for mid-IR pulse laser sources, especially for ultrafast photonics applications, which will be our future target.