Passively Mode-Locked Ytterbium-Doped Fiber Laser Based on Fe₃O₄ Nanosheets Saturable Absorber

Abstract

Two-dimensional material nanosheets have always been a research hotspot because of their unique structure and properties. We report mode-locked operation in ytterbium-doped fiber laser (YDFL) for the first time by adopting Fe₃O₄ nanosheets as a saturable absorber (SA). The laser is capable of generating 456 ps pulses, centered around 1039 nm. Our results manifest that Fe₃O₄ nanosheets are expected to become a new type of saturable absorber, which can better promote the development of mode-locked lasers.

1. Introduction

Ultrashort pulse lasers have important applications in medical [1], industrial [2], and scientific research fields [3]. Compared with active mode-locked techniques, passive mode-locked techniques are widely used to realize ultrashort pulse lasers because of their simple implementation and cavity structure. A saturable absorber (SA) is the key to the realization of passive mode-locked techniques, and the emerging new saturable absorber greatly promotes the development of passive mode-locked techniques. The mode-locked operation can be achieved simply and quickly by inserting a saturable absorber into the resonant cavity. Recently, various SAs have been proposed to achieve a mode-locked fiber laser, including the use of semiconductor saturable absorber mirrors (SESAM) [4], carbon nanotubes [5], topological insulators [6], graphene [7] and black phosphorus [8]. Among them, SESAMs are widely used because of their high thermal damage threshold and strong structural stability [9]. In addition, the researchers also found that transition metal dichalcogenides, such as WS₂ [10,11], Re_xNb_(1−x)S₂ [12], W_xNb_(1−x)Se₂ [13], Nb_xRe_(1−x)S₂ [14], PdSe₂ [15]; transition metal oxides, such as titanium dioxide [16], zinc oxide [17], indium tin oxide [18], aluminum oxide [19] and Fe₃O₄, have nonlinear saturable absorption effects. Among the transition metal oxides, Fe₃O₄ nanoparticles (FONPs) are widely used as SAs in mode-locked lasers, due to their high damage threshold [20], large third-order nonlinear effect and fast response time [21]. FONPs are generally regarded as semiconductor materials, and their semiconductor band gap width can be adjusted by adjusting the particle size [22].

In ytterbium-doped fiber lasers (YDFL), FONPs are used to achieve Q-switched and mode-locked operation. In 2018, Al-Hayali et al. demonstrated the operation of a dual-wavelength passively Q-switched YDFL, by using FONP-based SAs in a magnetic fluid [23]. In 2019, in the study of Li et al., FONPs were synthesized by chemical co-precipitation as SAs and applied as a Q-switch in the 1 μm region [24]. In 2020, Chen et al. reported a passively Q-switched mode-locked YDFL at a 1084.4 nm wavelength, by employing FONPs as SAs [21]. In 2021, Li et al. demonstrated passively Q-switched 1-, 1.5-, and 2-μm fiber lasers by utilizing FONPs as broadband SAs [25].

In this work, we prepared Fe₃O₄ nanosheets by a co-precipitation method for the study of the mode-locked operation of YDFLs. The saturation intensity of the prepared Fe₃O₄ nanosheets was 0.538 GW/cm², and the modulation depth was 14.5%. In addition, we realized the mode-locked operation of Fe₃O₄ nanosheets as SAs in a YDFL for the first time. Compared with FONPs, Fe₃O₄ nanosheets can obtain narrower pulse widths in mode-locked operation. By inserting the SA into the cavity, the mode-locked operation occurs when the pump power reaches 150 mW. The center wavelength of the mode-locked laser is around 1039 nm, the repetition frequency is 18.576 MHz, and the pulse duration is 456 ps. The maximum output power of the mode-locked laser is 13.6 mW, and the output slope efficiency is 11.36%. The results of this study demonstrate that Fe₃O₄ nanosheets can be used as one of the materials for realizing ultrashort pulse lasers.

2. Preparation and Characterization of Fe₃O₄ Nanosheets SAs

The Fe₃O₄ nanosheets used in this experiment were grown from commercial water-based magnetic fluids, as shown in Figure 1a. First, a self-made cuvette from a piece of quartz with a thickness of about 0.2 mm is made. Then, approximately 0.05 mL of the water-based magnetic fluid solution is placed into the cuvette. Next, the cuvette is placed in the magnetic field, and the length of the standing time is determined by the environment in which the experiment is performed. On the one hand, the water-based ferrofluid solution will be attracted by the magnetic force of the magnet, and on the other hand, it will be affected by the evaporation of water vapor. Under this dual influence, the ferromagnetic nanoparticles eventually grow into nanosheets. Finally, we take out the nanosheets to obtain the solution of the Fe₃O₄ nanosheets, as reflected in Figure 1b.

To prepare the SA, we first drop the Fe₃O₄ nanosheets solution onto the clean end face of a fiber ferrule. Then, the ferrule and the Fe₃O₄ nanosheets solution are allowed to stand for half an hour at room temperature, so that the deionized water in the solution is completely evaporated. Finally, we obtain the SA based on Fe₃O₄ nanosheets, as illustrated in Figure 2.

The surface morphology of the as-prepared Fe₃O₄ nanosheets is observed and analyzed by optical microscopy and scanning electron microscopy (SEM) (JEOL, JSM-7800, Tokyo, Japan). As depicted in Figure 3a, under the lens of a 25 × 10 optical microscope, different shades of color can be observed, which indicates that the darker the color, the thicker the nanosheet. Under SEM, as shown in Figure 3b, we can observe the obvious layered structure. The area with a darker color under the optical microscope and the area with more layers under SEM correspond to each other.

The transmission spectra of the Fe₃O₄ nanosheets were recorded by a UV–Vis–NIR spectrophotometer (PerkinElmer, Lambda 950, MA, USA), as shown in Figure 4. We can observe that the nanosheets have an absorptivity of 24.19%, at a wavelength of 1039 nm. The factors affecting this linear loss may be the thickness and uniformity of the Fe₃O₄ nanosheets [21].

The nonlinear absorption properties of the Fe₃O₄ nanosheets were recorded by a mode-locked pulsed laser (pulse duration: 17.31 ps, center wavelength: 1050 nm, repetition rate: 31.29 MHz). The recorded data were fitted using the following formula [20,21]:

$$T\left(I\right)=1-\mathsf{\Delta }T\cdot exp\left(-I∕I_{sat}\right)-T_{ns}$$

(1)

where $T\left(I\right)$ is the transmission, $\mathsf{\Delta }T$ is the modulation depth, $I$ is the input intensity, $I_{sat}$ is the saturation intensity, and $T_{ns}$ is the non-saturable loss. According to the fitting results displayed in Figure 5, the saturation intensity, modulation depth, and non-saturation loss are approximately 0.538 GW/cm², 14.5%, and 32.77%, respectively. We note that the energy density hitting the Fe₃O₄ nanosheets reaches the GW level. Compared with Fe₃O₄ nanoparticles, Fe₃O₄ nanosheets can withstand more energy.

3. Experimental Setup

Figure 6 demonstrates the experimental setup of a passively mode-locked YDFL, based on Fe₃O₄ nanosheets as the SA. The experimental structure adopts a fiber optic ring cavity. The pump light is emitted from a 976 nm laser diode (LD), with a maximum output power of 1 W. The pump light is coupled into the ring cavity through a 980/1064 wavelength division multiplexer (WDM). Ytterbium-doped gain fiber of about 0.4 m is used to realize the laser operation of 1 μm. A polarization independent isolator (PI-ISO) ensures the unidirectional transmission of the laser in the ring cavity. The 87:13 optical coupler (OC) outputs 13% of the laser energy in the ring cavity, which is convenient for observing and testing the state of the laser in the cavity. The polarization state and birefringence of the intracavity laser are regulated by a polarization controller (PC). The SA of the Fe₃O₄ nanosheets is inserted to realize mode-locked operation. The single-mode fiber (Nufern LMA-GDF-10/125-M) is applied to connect each device. The total cavity length is about 11 m.

The pulse trains in the experimental results were received by a wavelength range 800–1700 nm InGaAs photodetector (Thorlabs, DET08C/M, NJ, USA) and displayed in a 3 GHz bandwidth digital oscilloscope (LeCroy, WavePro 7300A, NY, USA). The spectral data were measured and analyzed using a wavelength range 600–1750 nm optical spectrum analyzer (Yokogawa, ANDO AQ6317B, Tokyo, Japan). The frequency was measured by a radio-frequency (RF) analyzer (Agilent, N9020A, CA, USA) from 10 Hz to 3.6 GHz.

4. Experimental Results and Discussion

To ensure that the SA is the key factor affecting the mode-locked operation, we tested the experimental setup without the SA in advance. By adjusting the pump source and PC, no mode-locked pulses appear on the oscilloscope. On the contrary, by inserting the SA into the cavity and adjusting the PC and pump power appropriately, the mode-locked operation happens when the pump power reaches 150 mW. The mode-locked operation is maintained when the pump power is increased from 150 mW to 220 mW. To avoid burning out the fiber ferrule and SA, we did not increase the pump power over 220 mW.

We summarize the experimental data and formulate the figures for analysis to discuss the characteristics of the mode-locked operation. Figure 7a demonstrates a typical pulse sequence with a period of 53.83 ns under 100 ns/div, which is consistent with the total laser cavity length of 11 m. Figure 7b shows the oscilloscope trace of a single pulse under 780 ps/div, with a full width at half maximum (FWHM) of 456 ps.

Figure 8 manifests the output optical spectrum when the pump power is 180 mW. The central wavelength is 1039.32 nm, and the 3 dB bandwidth is 0.047 nm.

To obtain the mode-locked pulse repetition frequency, we used the RF spectrum, recorded by a spectrum analyzer as shown in Figure 9. The fundamental repetition frequency is 18.576 MHz, and the signal-to-noise ratio (SNR) is greater than 20 dB with the resolution bandwidth of 30 kHz.

The variation in average output power and single pulse energy with pump power are shown in Figure 10. The pump power is between 100 mW and 150 mW, and the system is in continuous wave operation. The mode-locked pulse appears until the pump power reaches 150 mW. The slope efficiency of the mode-locked laser can reach 11.36%. As the pump power increases from 150 mW to 220 mW, the average output power increases from 5.7 mW to 13.6 mW, and the maximum single pulse energy is 0.7321 nJ.

In recent years, FONPs have been continuously adopted in YDFLs to achieve Q-switched and mode-locked operations. In

Table 1. Performance comparison of YDFL based on magnetic fluid as SA.

SA Materials	Modulation Depth	Pulse Duration	Reference
FONPs	NA	3.4 μs	[23]
FONPs	7.8%	1.68 μs	[24]
FONPs	6.6%	880 ps	[21]
FONPs	16.42%	3.78 μs	[25]
Fe3O4 nanosheets	14.5%	456 ps	This work

, we can observe that Fe₃O₄ nanosheets exhibit some advantages in pulse duration and modulation depth, compared to FONPs.

5. Conclusions

In conclusion, we fabricated Fe₃O₄ nanosheets using water-based ferrofluid and applied them as the SA in a YDFL, and successfully realized mode-locked operation. Our research extends the application of magnetic fluids in YDFLs. In addition, we believe that Fe₃O₄ nanosheets have great potential in achieving mode-locked operation, due to their excellent nonlinear optical properties and extremely high damage threshold. At the same time, we noticed that when the Fe₃O₄ nanosheets were grown, their thickness is random and we cannot control it at present. Thick Fe₃O₄ nanosheets can easily accumulate too much heat and cause damage to the fiber ferrule. Therefore, an appropriate thickness can better realize the mode-locked operation. Hopefully, this can be improved in future research.