Bi₂Te₃/Graphene Heterostructure as the Saturable Absorber for ~1.0 μm Passively Q-switched Solid State Pulsed Laser

Abstract

Due to the tunable nonlinear optical properties of the Bi₂Te₃/graphene heterostructure, stable solid state pulsed lasers based on the Bi₂Te₃/graphene saturable absorber have attracted intensive attention. In this work, the Bi₂Te₃/graphene heterostructure with good nonlinear absorption characteristics was synthesized by a self-assembly solvothermal route, and the optical saturable absorption properties of the saturable absorber were investigated. Owing to the large modulation depth of Bi₂Te₃ nanosheets and the high thermal conductivity of graphene, the Bi₂Te₃/graphene heterostructure saturable absorber shown good nonlinear saturable absorber performance and contributed the improved passively Q-switched Yb³⁺: GdAl₃(BO₃)₄ pulsed laser when compared with that of the pure Bi₂Te₃ based Yb³⁺: GdAl₃(BO₃)₄ laser, no matter pulse width or pulse energy. Our work demonstrates that the Bi₂Te₃/graphene heterostructure was a promising saturable absorber in ~1 μm solid-state pulsed lasers.

1. Introduction

For realizing pulsed lasers, an effective method is the passively Q-switching technology because of the compactness and simplicity of the cavity structure. In such a technology, saturable absorber (SA) is the most important component. In the past decades, several kinds of two-dimensional (2D) nanomaterials, including graphene, transition metal dichalcogenides (e.g., MoS₂ and WS₂), topological insulators (TIs, e.g., Bi₂Te₃, Sb₂Te₃, andBi₂Se₃), and black phosphorus, have been used as SAs in both fiber and solid-state Q-switched lasers [1,2,3,4,5,6,7]. Among these 2D nanomaterials SAs, graphene is the most popular one due to its broadband absorption, low nonsaturable loss, high damage threshold, and controllable modulation depth [1,8]. Efficient Q-switching fiber and bulk lasers based on graphene SA from the near to middle infrared region have been realized [9,10,11]. Nevertheless, the pulse energies of these pulsed lasers are prohibited by the relatively low modulation depth of graphene [11]. TIs are another kind of broadband SAs because of the gapless metallic state in their surface states. Compared with graphene SA, Bi₂Te₃ as a typical TI presents much higher modulation depths [12,13], but it is a relatively slower SA with a long relaxation time [14,15,16]. Furthermore, due to the heavily populated intrinsic defects and low thermal conductivity of Bi₂Te₃, most of the peak powers and pulse energies based on Bi₂Te₃ SAs were also very low [17,18].

Recently, to overcome the shortcomings of graphene SA and Bi₂Te₃ SA when they were used individually, a strategy to controllably grow the Bi₂Te₃/graphene heterostructure with controllable optical properties was provided. The modulation depth of the Bi₂Te₃/graphene heterostructure is much higher than that of monolayer graphene, which benefits from the stronger absorption of Bi₂Te₃ nanocrystals. At the same time, the exciton relaxation time is tunable with the change of Bi₂Te₃ coverage. Using the Bi₂Te₃/graphene heterostructure SA, Mu et al. realized both Q-switching and mode-locking fiber pulsed lasers at ~1.5 μm [19]. Wang et al. further demonstrated the broadband saturable absorber property and achieved passively mode-locked pulses in both Er³⁺-doped and Yb³⁺-doped fiber lasers [20].

Compared with fiber lasers, the solid state lasers were considered to be more liable to obtain high-energy short pulse due to less nonlinear pulse-splitting and larger mode areas [21]. For the solid state pulsed lasers, You et al. investigated ∼2 μm and ∼3 μm passively Q-switched lasers based on the Bi₂Te₃/graphene heterostructure SA by using the Tm³⁺: YAP and Er³⁺: YSGG as the laser gain medium, respectively [22]. Due to the widespread applications in communication, laser detection, and industrial processing, ∼1 μm pulsed lasers with high peak power and short pulse duration are still widely concerned. Generally, the Yb³⁺-doped crystals are promising laser gain material for realizing ∼1 μm pulsed lasers because of the simple energy levels, weak electron–phonon coupling, and long upper-state lifetime of Yb³⁺. To our best knowledge, Bi₂Te₃/graphene SA has not been used to Q-switched Yb³⁺-doped solid state lasers so far. So, in this work, the Yb³⁺: GdAl₃(BO₃)₄ (Yb³⁺: GAB) crystal was adopted as the laser gain medium to investigate the high-performance ∼1 μm solid state pulsed laser based on the Bi₂Te₃/graphene heterostructure SA.

2. Materials and Methods

There are various strategies to fabricated two-dimensional materials, such as chemical vapor deposition (CVD) [19], vacuum filtration [23], solvothermal method [22], and so on [24,25,26]. Among them, hydrothermal is a promising method because of its simple preparation process and easiness to achieve large size preparation. So, the Bi₂Te₃/graphene heterostructures in this work were synthesized by a self-assembly solvothermal route [22]. In detail, the used GO was firstly synthesized from graphite powder by the modified Hummers method [27]. Then, all the chemicals used for synthesizing Bi₂Te₃ nanosheets, Na₂TeO₃ (0.1 g), BiCl₃ (0.095 g), NaOH (0.12 g), and PVP (0.15 g), were added in graphene oxide (5 mg) which was dissolved in 15 mL ethylene glycol. After ultrasonic stirring for 30 min, the mixed solution was transferred to a 20 mL hydrothermal reactor and then was heated at 180 °C for 36 h. At last, the Bi₂Te₃/graphene heterostructure powder was obtained after cooling, washing, and drying. In this process, the Bi₂Te₃ nucleus could be tightly adsorbed on the GO by oxygen-containing functional groups and grown into Bi₂Te₃ nanosheets. Additionally, the GO could be reduced by ethylene glycol to graphene [22].

The SA was prepared by spin-coating the Bi₂Te₃/graphene heterostructure on quartz substrate. The optical saturable absorption properties of the SA were investigated by open-aperture z-scan technology [22]; here, a homemade ~1.0 μm Q-switched laser source was used.

To obtain the ~1 μm Q-switched laser, a fiber-coupled 976 nm diode laser with a core diameter of 200 μm and a numerical aperture of 0.22 was adopted as the pump source. The laser cavity is a plano-concave cavity with a length of 18 mm. A plane mirror with high-reflection coating at 1020–1080 nm and antireflection coating at 976 nm was used as the input mirror (IM), while the output coupler (OC) was a concave mirror, which has the partial transmission of 3% from 1020 to 1080 nm. The laser gain medium was uncoated 10 at.% Yb³⁺: GAB crystal with a size of 3 × 3 × 2 mm³. The Bi₂Te₃/graphene SA was located near the output mirror inside the resonant cavity. The corresponding experimental setup of the passively Q-switched laser is shown in Figure 1.

3. Results and Discussion

Figure 2a shows the XRD patterns of the prepared Bi₂Te₃/graphene heterostructure and the standard diffraction peaks of Bi₂Te₃. It is clear that the diffractions peaks of Bi₂Te₃/graphene heterostructure are in good agreement with those of the hexagonal Bi₂Te₃ phase (JCPDF card No. 15-0863), which indicates the successful synthesis of Bi₂Te₃ nanosheets. However, the typical diffraction peak of graphene cannot be found, which suggests that the existence of Bi₂Te₃ plates inhibited the restacking of graphene sheets [28]. To confirm the presence of graphene, the C1s XPS spectra of Bi₂Te₃/graphene heterostructure and bare Bi₂Te₃ fabricated by the solvothermal method were performed and given in Figure 2b. The peaks of the spectra of Bi₂Te₃/graphene heterostructure correspond to different forms of carbons: non-oxygenated carbon (C–C and C=C, 285 eV), carbonyl carbon (C=O, 288.2 eV). Compared to the strong XPS peaks of C=O and C–O of other works about graphene oxide, the omitted XPS peak of C–O and pretty weak peak of C=O in this work clearly demonstrated the reduction of GO into graphene [28].

To analyze the morphology and structure of the obtained Bi₂Te₃/graphene sample, the scanning electron microscope (SEM) and the Raman spectra were characterized. Figure 3a displays the SEM image of bare Bi₂Te₃ nanosheets. Obviously, the sample exhibits a hexagonal plate shape. In Figure 3b, the hexagonal Bi₂Te₃ nanosheets absorbed on the graphene surface presents quite sharp edges and a flat surface, which indicates they are well-crystallized. Figure 3c gives the typical Raman spectrum of the Bi₂Te₃ nanosheets, which was excited by a 532 nm laser. Four characteristic peaks at 62, 98, 116, and 140 cm⁻¹ are consistent with the vibrational modes of A_1g¹, E_g², A_1u², and A_1g² for Bi₂Te₃, respectively. Among them, the A_1u² peak is used to be the infrared-active mode in bulk crystalline Bi₂Te₃, which becomes Raman-active mode when the bulk breaks into thin films for the symmetry breaking [29,30]. In the Raman spectrum of the Bi₂Te₃/graphene heterostructure shown in Figure 3d, three vibrational modes of A_1g¹, E_g², and A_1u² for Bi₂Te₃ are observed in the low frequency range of 40–180 cm⁻¹. It is worth noting that the vibration strength of A_1u² increases as the thickness of Bi₂Te₃ decreases. Consequently, the Bi₂Te₃ nanosheets anchored on the surface of graphene are much thinner than the bare Bi₂Te₃ nanosheets synthesized by the solvothermal route directly, which benefits from the one-sided growth of Bi₂Te₃ on graphene. In the high frequency range of 800–3000 cm⁻¹, there are two Raman peaks of graphene. In detail, the characteristic peak at 1598 cm⁻¹ is identified as G band, and another peak at 1330 cm⁻¹ can be recognized as D band, which owes to the first-order zone boundary phonons in defective graphene [31].

The z-scan transmittance depends on the position of Bi₂Te₃/graphene SA shown in Figure 4a. The sharp peak on the curve indicates the good nonlinear absorption characteristics of SA. Furthermore, the relationship between optical transmittances and increased excited intensity can be determined by translating from the z-scan curve, which is given in Figure 4b. In addition, the modulation depth of 26.5%, the saturation intensity of 1.34 MW/cm², and the non-saturable losses of 3.1% can be calculated through curve fitting with the formula:

$$T\left(I\right) = 1 - \Delta R \times \exp \left(-I/I_{sat}\right) - T_{ns}$$

(1)

where ΔR, I, I_sat, and T_ns are the modulation depth, the exciting intensity, the saturation intensity, and the nonsaturable loss, respectively.

Table 1. Nonlinear absorption parameters of graphene SA, Bi₂Te₃ SA, and graphene/Bi₂Te₃ SA.

Type of SAs	Modulation Depth	Saturation Intensity	Nonsaturable Loss	Ref.
Monolayer graphene	4.5%	0.8 MW/cm2	-	[32]
6.5 nm Bi2Te3	6.5%	35 W/cm2	-
Bi2Te3/grapheme(CVD)	23%	0.6 MW/cm2	-
Bi2Te3	14.29%	16.62 kW/cm2	4.23%	[33]
Bi2Te3/graphene	26.5%	1.34 MW/cm2	3.1%	This work

lists the nonlinear absorption parameters of monolayer graphene SA, Bi₂Te₃ SA, and Bi₂Te₃/graphene SA. It is obvious that the saturation intensity of the bare Bi₂Te₃ nanosheet was pretty low, which can mainly be attributed to the weakened absorption caused by the oxidation of the Bi₂Te₃ nanosheets. When grown on graphene, the surface of the Bi₂Te₃ nanosheet can be protected from oxidation effectively, which is also conducive to improve the modulation depth. Due to the high thermal conductivity of graphene, the nonsaturable loss of Bi₂Te₃/graphene SA was smaller. The relatively lager modulation depth and saturation intensity as well as the lower nonsaturable loss of Bi₂Te₃/graphene SA will be advantageous to the realization of high-performance Q-switched lasers. It is worth noting that the modulation depth and saturation intensity of the Bi₂Te₃/graphene SA in this work are comparable to those of the Bi₂Te₃/graphene SA produced by CVD [32].

A stable passive Q-switching laser operation was realized when the absorbed pump power was increased to be 2.86 W. The output power nearly increased linearly with the absorbed power elevated, and the maximum average output power of 305 mW with a slope efficiency of 11.7% was obtained until the absorbed power was 5.45 W, as shown in Figure 5a. The pulse width and repetition rate depending on the absorbed pump power was presented in Figure 5b. The repetition rate and pulse width varied from 23.6 kHz to 181.1 kHz and 2.65 μs to 280 ns, respectively, in the stable Q-switching regime. Based on the measured repetition rate and average output power, the energy of pulse was estimated and shown in Figure 5a. Obviously, the pulse energy increased monotonically with the pump power up to 4.8 W, with a maximum value of 1.83 μJ. Then, it decreased slightly with the further increased pump power due to the thermal effect in the cavity. Furthermore, the pulse energy was 1.68 μJ under the maximum absorbed power, and the relevant Q-switched laser pulse trains and single pulse are displayed in Figure 6.

Table 2. Comparison of ∼1 μm passively Q-Switched laser performance using 2D SAs.

Type of SAs	Gain Crystal	Output Power (mw)	Pulse Width (ns)	Repetition Rate (kHz)	Pulse Energy (μJ)	Ref.
Bi2Te3	Nd: Lu2O3	79	720	94.7	0.83	[18]
BP	Yb: CYA	37	620	113.6	0.33	[31]
MoS2	Nd: YAP	260	227	232.5	1.11	[34]
Bi2Se3	Yb: KGW	439.4	1600	166.7	2.64	[35]
Bi2Te3	Yb: CYB	161	416	147.7	1.09	[33]
Bi2Te3	Yb: GAB	57	415	111.4	0.51	[36]
Bi2Te3/graphene	Yb: GAB	305	280	181.1	1.83	This work

shows some excellent ∼1 μm passively Q-switched solid state laser performances with 2D SAs. It can be seen that our results are superior to the results yielded by using BP, MoS₂, and Bi₂Te₃. Especially compared with the Q-switched laser performances based on Bi₂Te₃ individually, the shorter pulses and higher power were obtained simultaneously when utilizing the Bi₂Te₃/graphene heterostructure as the SA. Despite the slightly higher output power of Bi₂Se₃ Q-switched Yb: KGW laser, the narrow pulse width in this work will widen the practical application. It is important to note that, besides the SAs, the gain crystals and cavity design strongly influence the Q-switched laser performance. So, it is hopeful that better performance can be realized by coating the crystal or further optimizing the cavity structure.

4. Conclusions

In this work, the Bi₂Te₃/graphene heterostructure was successfully fabricated by the solvothermal reaction, and the relevant nonlinear saturable absorption characteristic was also explored. The saturation intensity, modulation depth, and nonsaturable loss were estimated to be 1.34 MW/cm², 26.5%, and 3.1%, respectively. By using the Yb³⁺: GAB crystal as the laser gain medium, the passively Q-switched pulsed laser of ∼1 μm was demonstrated. The maximum output power of 305 mW was obtained, corresponding to a pulse width of 280 ns and a repetition rate of 181.1 kHz. The experimental results demonstrate that the Bi₂Te₃/graphene heterostructure is an appropriate SA for Q-switched lasers in near-infrared range.