The experiment was first conducted without any of the SA devices to observe the operation of the fiber laser. As expected, the TDFL without any SA only operated in the continuous wave (CW) regime as no pulses were observed in the oscilloscope as the pump power was increased to a maximum of 1 W. The experiment was then continued by integrating the Ta2AlC-deposited tapered fiber into the laser cavity. With fine-tuning of the PC, the fundamental mode-locking of the TDFL was observed at a threshold pump power of 245 mW. The fundamental mode-locking could be obtained until the pump power reached 480 mW. As the pump power was further increased until the maximum, the mode-locked operation was sustained but the laser operated at higher harmonics. As our interest was mainly on the fundamental operation of the mode-locked laser, the laser characteristics were only recorded when the pulsed laser operated at the fundamental frequency.
The characteristics of the mode-locked laser using the Ta2AlC-deposited tapered fiber are shown in Fig. 10. The optical spectrum recorded in Fig. 10 (a) shows a broad laser spectrum at a center wavelength of 1937 nm, having a 3-dB bandwidth of 2.8 nm. It was apparent that Kelly sidebands were observed in the soliton spectrum when the mode-locking operation was initiated, which was expected as the mode-locked TDFL was operating in the anomalous dispersion regime. From the values obtained from the optical spectrum, the transform-limited pulse width could be estimated by the equation;
where c is the speed of light, TBP is the time-bandwidth product, λc is the center wavelength, and ∆λ is the 3-dB bandwidth. Taking the TBP to be 0.315 for a sech2 pulse profile, the center wavelength to be 1937 nm, and the 3-dB bandwidth to be 2.8 nm, the transform-limited pulse width was calculated to be 1.407 ps. Figure 10 (b) shows that the oscilloscope trace of the mode-locked pulses had a repetition rate of 10.73 MHz, which correlates with the cavity round trip time estimated from the cavity length of approximately 19.3 m. A comparable pulse train was achieved by Zhou et al.59, where the slight fluctuation of the pulse peak intensities was affected by the relatively low sampling rate of the photodetector bandwidth and the low sampling points per pulse. Nonetheless, it was found that the pulse train's amplitude jitter was within reasonable limits. The radio frequency (RF) spectrum of the mode-locked pulse is shown in Fig. 10 (c), whereby a sharp peak was observed at about 10.73 MHz with a measured signaltonoise ratio (SNR) of ~ 55 dB. Figure 10 (d) shows the autocorrelation trace of the modelocked pulse, where the pulse width was measured to be 1.678 ps when fitted with a sech2 fitting. It was only about 19% longer than the calculated transform-limited pulse width. The corresponding TBP was 0.375, indicating that the pulse width is only slightly chirped.
The experiment was then further conducted by replacing the Ta2AlC-deposited tapered fiber with the Ta2AlC-based side-polished fiber that has been described in Sect. 3.3. The fundamental mode-locking operation was achieved at a threshold pump power of 351 mW. The laser spectrum is given in Fig. 11 (a) also shows a typical soliton spectrum with distinct Kelly sidebands. However, it was observed that the sidebands were uneven, with the longer being higher than the shorter-wavelength sidebands. It is highly likely due to the optical fiber birefringence filtering effect in the cavity, as was theoretically and experimentally proven by Man et al.72. The center wavelength and the 3-dB bandwidth of the TDFL were recorded to be 1931 nm and 3.1 nm, respectively. As for the frequency of the mode-locked pulses, the oscilloscope trace shown in Fig. 11 (b) shows a repetition rate of 9.52 MHz. The frequency was lower compared to the mode-locked laser with the Ta2AlC-deposited tapered fiber, which was due to the slightly longer length of the SPF. The RF spectrum of the mode-locked laser in Fig. 11 (c) shows a sharp peak at around 9.52 MHz, having an SNR of ~ 50.5 dB. From the autocorrelation trace of the mode-locked pulse shown in Fig. 11 (d), the pulse width was measured to be 1.743 ps, fitted with a sech2 profile. The corresponding TBP was then calculated to be 0.434, also indicating that the pulse was chirped.
The TDFL cavity was further tested by inserting the fabricated D-shaped fiber with the Ta2AlC solution. The mode-locked pulses were generated at a pump power of 380 mW. The output characteristics of the mode-locked laser are shown in Fig. 12. From the optical spectrum plotted in Fig. 12 (a), the mode-locked laser had a center wavelength of 1929 nm with a 3-dB bandwidth of 2.2 nm. The presence of minor dips in the optical spectrum could be attributed to water absorption lines in 2 µm 73. The mode-locked pulse had a repetition rate of 10.16 MHz with a 9.84 ns interval between peaks, measured from the oscilloscope trace in Fig. 12 (b). It tallies well with the fundamental frequency of the mode-locked laser, which was estimated by f = c/nL where c is the speed of light, n is the refractive index of an optical fiber, and L is the length of the cavity. By taking n to be approximately 1.44 at 2000 nm and L to be 20.43 meters, the fundamental frequency was about 10.2 MHz. Figure 12 (c) shows the RF spectrum with a peak that corresponds to the repetition rate of the mode-locked laser, in which it has an SNR value of ~ 47 dB. The autocorrelation trace of the pulse recorded using the autocorrelator is shown in Fig. 12 (d). When fitted with a sech2 fitting curve, the pulse width was measured to be 1.817 ps.
The stability of the mode-locking operation with all three SA devices was evaluated by conducting a long-term stability test over two hours. The laser output for each of the three SA devices was monitored at every 10-minute interval, in which their optical spectrum was recorded and plotted in Fig. 13. As seen from Fig. 13 (a), the mode-locked TDFL with the Ta2AlC-deposited tapered fiber exhibit a steady output intensity, and the contour plot shows no changes in the central wavelength of the output spectrum. For Ta2AlC-deposited SPF, the output spectrum displayed in Fig. 13 (b) also shows a very stable output as the central wavelength and the Kelly sideband exhibit no changes throughout the stability test. Meanwhile, the D-shaped fiber demonstrates a steady output as depicted in Fig. 13 (c).
The pump power against the output power of the mode-locked lasers is plotted in Fig. 14. As seen from the graph, the average output power of the laser with each type of SA device increased almost linearly after the mode-locking operation was initiated. The mode-locking threshold was 245 mW for the TDFL with the Ta2AlC-deposited tapered fiber, 351 mW for the Ta2AlC-deposited SPF, and 380 mW the Ta2AlC-deposited D-shaped fiber. At this pump power, the fundamental mode-locking (FML) was observed. The FML operation was sustained until a pump power of 480 mW for the Ta2AlC-deposited tapered fiber and 479 mW and 550 mW for the Ta2AlC-deposited SPF and D-shaped fiber, respectively. At the maximum pump power in which the FML was sustained, the average output power recorded for the mode-locked TDFL with the tapered fiber, SPF, side polished, and D-shaped fiber was 1.91 mW, 0.8 mW 1.37 mW, respectively. When the pump power was further increased beyond these pump power, harmonic mode-locking was observed and could be maintained up until the maximum pump power of 1 W. At the maximum pump power of 1 W, the maximum average output power obtained were 3.48 mW, 2.27 mW and 2.71 mW for the TDFL with the tapered fiber, SPF and D-shaped fiber, respectively. The TDFL with the Ta2AlC-deposited tapered fiber had the highest output power as the structure of the tapered fiber was maintained, only its dimension was reduced74.
In contrast, the mode-locked TDFL with the Ta2AlC-deposited SPF and D-shaped fiber had a lower output power since the fiber structure had been modified during the grinding or polishing process. It caused a higher amount of light to escape easily by scattering light due to imperfection of the surfaces. Nevertheless, all the three SA devices could operate even when the pump power was increased until a maximum pump power of 1 W, without being optically damaged. Compared with materials embedded in polymer hosts, SAs in the film form typically had a lower damage threshold and can be easily burnt when the power was high 75. It limits the application of polymer-based SAs in ultrafast fiber laser systems, eliminating the possibility of power-scaling of fiber lasers.
The performance of the mode-locked TDFLs with each of the three SA devices is summarized in Table 3.
Table 3
Summary of the optical characteristics of the TDFL with each mode-locking device in the fundamental operation.
Type of device
|
Threshold pump power (mW)
|
Repetition rate (MHz)
|
Signal to noise ratio (SNR)
|
Pulse width (ps)
|
Max. Output power (mW)
|
Max. Peak Power (W)
|
Tapered
|
245
|
10.73
|
55
|
1.678
|
1.91
|
106
|
SPF
|
351
|
9.52
|
50.5
|
1.743
|
0.8
|
43
|
D-shaped
|
380
|
10.16
|
47
|
1.817
|
1.37
|
82
|
From Table 3, the lowest pump power needed to initiate the mode-locking operation was that of the Ta2AlC-deposited tapered fiber at 245 mW. This was followed by the Ta2AlC-deposited SPF and the D-shaped fiber at the pump power of 351 mW and 380 mW, respectively. A low pump power threshold to induce the mode-locking operation was favorable as it could reduce the energy consumption. The mode-locked TDFL using the Ta2AlC-deposited tapered fiber generated the highest maximum average output power, which was as high as 3.48 mW. Compared to the output power of the TDFL using the Ta2AlC-deposited SPF and D-shaped fiber, the generated output power was only 2.71 mW for the former and 2.27 mW for the latter. It was lower by about 22% from the output power generated by the TDFL with the Ta2AlC-deposited tapered fiber. The corresponding peak power could be calculated by dividing the average output power with the repetition rate and then dividing the value with the pulse width. It gives peak power values of 106 W, 43 W, and 82 W for the TDFL with the Ta2AlC-deposited tapered fiber, SPF, and D-shaped fiber. The lowest and highest repetition rate recorded was 9.52 MHz with the Ti2AlC-deposited SPF and 10.73 MHz with the Ta2AlC-deposited tapered fiber. The difference in the repetition rate was only due to the length of the fiber used for the evanescent field-based fibers. In this regard, a higher repetition rate of the mode-locked TDFL could be obtained by having a shorter cavity length. The SNR values recorded from the RF spectrum for all three cases were more than 47 dB, which indicates that all three evanescent field-based SAs could generate stable mode-locked pulses. The shortest pulse width was recorded to be 1.678 ps, obtained using the Ta2AlC-deposited tapered fiber. The pulse widths obtained using the Ta2AlC-deposited SPF and D-shaped fiber were slightly longer, with values of 1.734 ps and 1.817 ps being measured. Overall, it is seen that all the evanescent field-based SA devices could generate mode-locked pulses in the 2 µm region, with the Ta2AlC-deposited tapered fiber providing the best performance in terms of the low mode-locking threshold, the highest output, and peak power, as well as having the shortest pulse width.