Narrow-Linewidth Tunable Fiber Laser Based on Laser-Induced Graphene Heated Fiber Bragg Grating with Low Voltage

Abstract

In this paper, we demonstrate a narrow-linewidth tunable fiber laser based on laser-induced graphene (LIG) paper-heated fiber Bragg grating (FBG) with low voltage. A linewidth of less than 600 Hz is achieved by the combination of a piece of unpumped Er-doped fiber and an FBG. Changing the temperature of the FBG will result in the central transmission spectrum shifting, and hence the laser wavelength tuning. LIG-heated (LIG-H) fabrication on polyimide (PI) paper by CO₂ laser is used to offer temperature control of the FBG. By adjusting the voltage of the LIG-H from 0 to 5 V, the temperature of the LIG-H can be changed from room temperature up to 220 °C, while the central wavelength of the output laser can be continuously adjusted from 1549.5 nm to 1552 nm with a full range of 2.5 nm. The proposed technique by electric control of LIG-H can provide a low-cost and compact wavelength tunable laser design.

1. Introduction

C-band tunable narrow linewidth erbium-doped fiber lasers (EDFLs) have been widely used in the area of optical communications [1,2], fiber optic sensing [3], and LIDAR system [4]. Fiber lasers are also used to achieve high precision methane [5] and protein detection [6], which is relevant to human life. To achieve wavelength tuning, several methods have been proposed, such as employing stress-stretched FBG [7], heated FBG [8], optically tunable band filter [9], Fabry–Perot tunable filters [10], and Mach–Zehnder interferometers [10]. Among them, the use of FBG is cheap and simple. Stress-stretched FBG has certain requirements for mechanical devices, and it is difficult to achieve a good balance between cost, accuracy and volume. Therefore, heated FBG was chosen by us to tune the wavelength. Commercial heating methods, such as employing water batch heating or heating tables, have previously been widely used for FBG heating. However, water bath heating is limited by the boiling point temperature of the water, and incurs safety risks. The size of the heating table is normally large and high voltages are required. Therefore, it is important to find a suitable heating device. Graphene has become an indispensable material for modern life due to its excellent physical properties, and is widely used in industrial fields, such as defense and military [11], aerospace [12], transportation and freight [13], etc. Compared with the commercial heating method, LIG-H, with an extremely high thermal conductivity of about ~5000 W/mK [14], is a promising material for temperature control. It has the advantages of ultra-high operating temperature at relatively low voltage, fast and stable cyclic heating temperature and uniform temperature distribution [15]. Moreover, LIG-H fabricated on PI substrates by direct laser writing has the benefit of flexible shape control [16], through which it is feasible to fabricate a LIG-H of suitable size compared with FBG.

In this paper, we introduce a LIG-H into a narrow linewidth fiber laser. Laser linewidth of less than 600 Hz is achieved by using the combination of unpumped Er-doped fiber and FBG. The wavelength tuning is achieved by the electric control of the LIG-H attached to the FBG. By designing a suitable size of LIG-H, the highest temperature of ~220 °C can be achieved at a low voltage of only 5 V. The wavelength tuning from 1549.5 nm to 1552 nm is achieved by changing the voltage from 0 V to 5 V. During the whole process, the laser exhibits good stability while maintaining a narrow linewidth of less than 600 Hz. We believe that this method inspires the integration of LIG-H into wavelength tunable fiber lasers and will show great potential.

2. Fabrication and Characteristics of the LIG-H

In order to produce LIG heating medium, a CO₂ laser (λ = 10.6 μm) is used to irradiate the PI paper, as shown in Figure 1a. The localized high temperature accumulated by the laser makes the chemical bonds of the PI break and rearrange into aromatic compounds, forming the graphene structure [17]. The samples were fabricated using PI films with a thickness of 125 μm, and the pattern of different sizes of LIG-H can be adjusted through the program. To generate the homogeneous LIG-H, scanning speed of 20 mm/s, and power of 3.5 W were used. To explore the performance of LIG-H, copper electrodes were attached to both sides of the sample and the gaps between electrodes and LIG-H were filled with conductive silver paint, as shown in Figure 1b. Foam tape was used to fix the LIG-H on the glass slide as shown in Figure 1c. This keeps the LIG-H flat and allows the back surface of the LIG-H to have less heat loss, hence a higher temperature at the same voltage.

To confirm the LIG structure, scanning electron microscopy (SEM) images were obtained from SU8220 Hitachi. Figure 1d,e, taken from the top-surface of LIG-H, revealed a uniform porous morphology of the laser-induced medium with high similarity to previous studies [18]. This structure is formed due to the generation of gas during the process of LIG, and the transient high pressure formed by the gas generation contributes to the porous structure of the LIG. The LIG structure was subsequently verified by spectroscopic analyses, including X-ray diffraction (XRD) through a Rigaku Ultima IV instrument with Cu Kα radiation λ = 1.54 Å and Raman microscopy carried out with a Thermo Dxr2xi Raman microscope (532 nm excitation laser) [19]. Figure 1f further explores XRD (top panel) and Raman (bottom panel) spectroscopy. In XRD, the (002) and (100) XRD peaks centered at 2θ = 23° and 2θ = 43°, evidence of the presence of the multilayered in-plane structure of graphene. Comparatively, Raman spectra show the outstanding D (~1350 cm⁻¹), G (∼1590 cm⁻¹), and 2D (∼2692 cm⁻¹) characteristic peaks, providing additional evidence of a few layered structures of graphitic crystals. These plots (SEM, XRD, Raman) prove that the material we produced is graphene [20].

In a further evaluation of the electrothermal performance of LIG-H, the IR thermograms of different sizes of LIG-H were measured, and shown in Figure 2a. In Figure 2a, the LIG-H marked with black rectangle exhibit homogeneous temperature. Then, a voltage source and a temperature sensor (Melexis MLX90614) were combined with a microcontroller (Espressif ESP32) to detect and record the numerical value between the voltage and the temperature. The voltage source and temperature sensor have a resolution of 0.025 V and 0.01 °C, respectively. To better understand the experimental observation, the thermal properties of LIG-H have been studied and its Joule heating model can be expressed as [21]:

$$T-T_0=\frac{P}{hS}\left(1-e^{-\frac{t}{\tau }}\right),$$

(1)

$$T=\frac{U^2}{R}hS+T_0$$

(2)

where T is the temperature of the heater, T₀ is the ambient temperature, and S, h, P, U, and τ are area, convective heat-transfer coefficient, electrical power input, voltage input, and time constant of the heater, respectively. Equation (1) shows the temperature increase trend of LIG-H well, while Equation (2) presents the stable temperature of the sample as a function of the applied voltage. To evaluate the temperature characteristics of different sizes of LIG-H, we made LIG-H in 5 sizes from 20 × 10 mm² to 20 × 30 mm². As demonstrated in Figure 2b, the heating temperature ramps up instantaneously and gradually stabilizes at a certain level with a voltage of 5 V. The experimental results conformed well with Equation (1), and also prove that the smaller size of LIG-H has a faster rate of temperature rise. By further studying the stable temperature of the sample versus voltage, Figure 2c shows that the smaller size of LIG-H has a higher stable temperature at the same voltage; the result conforms to Equation (2). According to Equation (2), the smaller size of LIG-H has a smaller R and S; this made the smaller size of LIG-H have the higher temperature. Thus, the LIG-H with a size of 20 × 10 mm² was chosen, which has enough area to place the whole FBG. Moreover, to study the LIG-H with a size of 20 × 10 mm² in more detail, Figure 2d generally shows the temperature increasing and stabilizing by stepwise enhancing the level of current from 0 V to 5 V. The inset plot provides the magnification when the voltage is 3 V. It can be observed that the temperature rises quickly from one voltage to a higher voltage. In addition, a 1200 s heating and cooling test was also carried out by continuously monitoring the heating temperature of a 20 × 10 mm² LIG-H over 20 cycles of on and off power. As shown in Figure 2e, in every cycle, the fixed electrical voltage (5 V) was turned on for 30 s and then off for another 30 s. Acquired by the sensor of MLX90614 to detect the heater, the profile clearly indicates that the temperature heats from ∼25 to ∼220 °C during the first powered stage; and it is back to the ∼71 °C level after the cooling stage. On the basis of the overall observation of 20 cycles, the heating performance of LIG-H shows high stability and reproducibility without any tendency for degradation or variation. Measured by a multimeter, the resistance of the LIG-H with the size of 20 × 10 mm² was 17 Ω, and the maximum power was less than 1.5 W (within 5 V voltage). Based on its excellent performance, we chose the LIG-H with a size of 20 × 10 mm² as the heat source for the FBG application in the fiber laser.

3. Experimental Scheme and Results

The fiber laser setup is shown in Figure 3a. The fiber laser consisted of a wavelength division multiplex (WDM), a 976 nm pump source, two optical isolators (ISOS1, ISO2), an erbium-doped fiber (EDF1), an optical circulator, a polarization controller (PC), an unpumped erbium-doped fiber (EDF2), an FBG (Guangyou Technology) and a 1 × 23 dB coupler (OC). The EDF1 (EDF22/6/125-23, YOFC) was a 4.3 m fiber, and the absorption coefficient was 20 dB/m at 1530 nm. The narrow linewidth running is achieved by using the combination of the EDF2 (EDF7/6/125-23, YOFC) and FBG, which were added into the ring cavity via a circulator. The absorption coefficient of EDF2 at 1530 nm is 7 dB/m, and the length is 6 m. The central wavelength of FBG is 1549.5 nm with 99.5% reflectivity and 0.1 nm reflectance bandwidth. When the signal light passes through the circulator, the outgoing light from port 2 and the reflected light from the FBG meet in the EDF2 to form a standing wave. At this time, the saturable absorption based on the unpumped EDF causes an ultra-narrowband auto-tracking filter due to spatial hole burning [10], which has a very narrow filtering bandwidth and makes the laser output a single longitudinal mode (SLM).

The pump power-dependent output power is measured and shown in Figure 3b. The pumping threshold is 28.4 mW, and it has the maximum SLM output power (10.4 mW) when the pumping power is 104.5 mW. The optical conversion slope efficiency is 13.8%. As shown in Figure 3c, the maximum power fluctuation was less than 0.1 dB during continuous monitoring for up to 1 h, showing good long-term power stability. An optical delay self-heterodyne measurement system, as shown in Figure 3d, is used to check the laser output linewidth. In this system, the acousto-optical frequency shifter instrument (AOM) operates at 1520–1580 nm with a center frequency of 80 MHz; the delay fiber path is up to 40 Km of G652D fiber. The 3 dB bandwidth of the photodetector (PD) is 2 GHz. Electrical Spectrometer Analysis (ESA) offers superb performance with an accuracy of up to 5 Hz. The test laser is split into two beams after OC1, the light passing through the delay fiber and the light passing through the AOM produce a beat signal at OC2, which is detected by the PD and transmitted to the ESA. From Figure 3e,f, we could find that the single longitudinal mode running of linewidth below 520 Hz with a side mode suppression ratio greater than 43 dB is achieved.

In order to understand the electrical tuning performance of the EDFLs more easily, it is necessary to discuss the temperature sensitivity of FBG. For FBG, the central wavelength can be expressed as follows [22]:

$${\lambda }_B=2n_{eff}\mathsf{\Lambda };$$

(3)

$$\Delta {\lambda }_B={\lambda }_B\left(\zeta +\alpha \right)\Delta T=\eta \Delta T ;$$

(4)

$$\Delta {\lambda }_B=\eta \frac{U^2}{R}hS={\eta }_0{U}^2.$$

(5)

where, ${\lambda }_B$, $n_{eff}$ and $\mathsf{\Lambda }$ represent the central wavelength, effective refractive index, and the period of the fiber grating, respectively; $\Delta {\lambda }_B,$ $\zeta$ and $\alpha$ denote the variation of the central wavelength, thermo-optical and linear expansion coefficients, respectively. Because the variation of $n_{eff}$ is small enough, it can be neglected. Therefore, in the thermal expansion model of FBG, the central wavelength shows a linear relationship with $\mathsf{\Lambda }$ [23]. Then, Equation (3) can be further expressed as Equation (4) with the change of temperature [22]. Where $\Delta {\lambda }_B$ is linearly positively correlated with $\Delta T$, and the ratio coefficient is the fiber grating temperature sensitivity coefficient $\eta$, respectively. Based on the Joule heat model of Equation (2), Equation (4) can be described as Equation (5); where ${\eta }_0$ is a second-order scaling factor for the relationship between wavelength and voltage.

Therefore, when the FBG is placed on the LIG-H and fixed by high temperature tape, the central wavelength of the laser can be changed by adjusting the voltage of the LIG-H. The lasing spectra corresponding to different voltages of LIG-H were measured, as presented in Figure 4a. The voltage of the LIG-H was set from 0 V to 5 V, and the central wavelength was tuned from 1549.5 nm to 1552.0 nm with a full range of 2.5 nm, which is shown in Figure 4b. The optical signal-to-noise ratio (OSNR) at different wavelengths of the EDFLs are shown in Figure 4b. The OSNAs in the wavelength tuning range are all greater than 50 dB. The dependence between the lasing wavelength and voltage shows a quadratic function which is consistent with Equation (5). We also measured the power fluctuation of each wavelength for ten minutes, which is shown in Figure 4c. From Figure 4c, we can see that power fluctuation of less than 0.2 dB/0.4 mW is achieved. Moreover, we also measured the linewidth of the EDFLs at different tuning voltages, as shown in Figure 4d–f. Laser running at all the wavelengths has a narrow linewidth of less than 600 Hz and a side mode suppression ratio greater than 40 dB.

4. Conclusions

In this paper, we proposed and demonstrated a narrow-linewidth tunable fiber laser based on laser-induced graphene-heated FBG with voltage of less than 5 V. When the voltage is not higher than 5 V, the maximum power of LIG-H is less than 1.5 W, which provides the possibility of using a regular battery to tune the laser wavelength. Compared with conventional mechanically tunable as well as piezoelectric controlled filters, our proposed wavelength tuning scheme has a simple and compact structure, low cost and high flexibility. The fiber laser achieves a maximum output power of around 10 mW, with linewidth of less than 600 Hz achieved at a wavelength tuning range from 1549.5 nm to 1552.0 nm. The high optical signal-to-noise ratio (more than 50 dB) and power stability (less than 0.2dB) of the proposed fiber laser make it possible for it to be used in high-precision fields. Moreover, the scheme has the potential to reduce the cost of tunable narrow-linewidth fiber lasers and is expected to be used in a variety of fiber lasers.