1. Introduction

Passively Q-switched fiber lasers are the focus of research efforts due to their potential in several applications including medicine, security, sensing and optical communication [1–3]. To date, various materials have been demonstrated as efficient saturable absorbers (SA) for both mode-locking and Q-switching of lasers [4]. SA acts as a Q-switcher in a laser cavity to generate short optical pulse trains. Conventional bulk SAs such as ion-doped crystals [5] and color filter glasses [6] have many limitations in the generation of pulsed fiber lasers, which includes low optical damage threshold, slow response time and narrow spectral range [4]. Although semiconductor saturable absorber mirrors (SESAMs) have been developed as a mature technology for fiber laser [7–9], they still have the disadvantages of narrow operating bandwidth, complex fabrication and expensive packaging. In recent years, there were also extensive investigations on the usage of low-dimensional nanomaterials due to their remarkable optical and optoelectronic properties [7–12].

Carbon nanotubes (CNTs) and graphene have been successfully demonstrated as SAs in fiber lasers, as they are easy to be fabricated and cost-effective. However, CNTs often demand bandgap engineering or charity control which limits its operation at certain wavelengths [13]. On the other hand, graphene has the drawback of zero bandgap which limited its capability in several applications. Different techniques have been proposed to open the bandgap in graphene, but that only made graphene less attractive for particular applications [14,15]. Transition metal dichalcogenides (TMDs) are another example of two-dimensional layered materials where their optoelectronic properties are strongly thickness-dependent, offering the possibility to engineer their optical properties for desirable performances [16]. Black phosphorus (BP) [17] and gold nanomaterials [18–20] have also attracted much attention due to their attractive performance in pulsed fiber lasers. However, these materials (i.e. TMD, BP and gold nanomaterials) suffer from a low optical damage threshold [21,22], additionally, complex fabrication process and insufficient purity and non-uniformity are common factors that degrade their performance as SAs.

Nevertheless, there have been increasing concerns regarding the effect of long-term exposure of these nanomaterials on human health [23]. On the other hand, finding a new material to be used as a high-performance SA is of great interest. In that regard, organic materials might have the potential to be used as a bio-compatible, environmentally-friendly and high-performance SAs. They have the advantages of large and ultrafast nonlinear response and broad spectral tunability [24–26]. Moreover, they offer mechanical flexibility, light-weight and low production cost [27], additionally, organic materials might increase the applications in biomedical laser processing and sensing [28]. These materials, in comparison to inorganic materials, require simple fabrication and offer versatile molecular design [25,29]. Due to their customized properties for optoelectronics, organic materials have been utilized in various applications including organic solar cells (OSCs) [30–33], organic light-emitting diodes (OLEDs) [34–37], organic bistable memory devices [38,39] and organic thin film transistor [40,41]. Despite numerous works reported on the use of organic materials in the linear regimes and in a relatively low power domain, the applications of organic materials in nonlinear optics, especially in the fiber lasers pulse generation, have yet to be fully explored. Tris(8-hydroxyquinolino)aluminum (Alq₃) is considered the most rigorously studied small molecule organic semiconductor[42], thus it became the choice for organic SA in our research. Alq₃ which has superior thermal stability can be easily synthesized and purified [43]. Alq₃ also has high photoconductivity and long operational lifetimes [44]. Furthermore, Alq₃ has the advantages of highly stable thin film formation and good heat resistance [45].

This work tries to draw the attention to the potentials that organic material might have in fiber laser technology. We used Alq₃ in a simple all-compact fiber cavity and produced a very interesting result. The laser cavity produced very high pulse energy and very high peak power of 0.81 µJ and 237.62 mW, respectively, in two different setups at 1 µm region. Thus, we believe that Alq₃ might be a strong potential candidate for developing pulsed fiber laser with a very interesting performance. Compared with the materials mentioned above, Alq₃ has relatively high optical damage threshold (higher than 350 mW) and high output pulses energy. Furthermore, it has an optical band gap of 2.5 eV with a simple, low-cost and straight forward fabrication process, which may be useful for other photonics applications. However, further research is required to develop the cavity structure and adopt an enhanced fabrication technique. It should be mentioned that we successfully used this SA to produce mode-lock in 1 µm, in a different cavity setup. Furthermore, we used this SA to fabricate Q-switch, tunable wavelength Q-switch (from 1520 nm to 1563 nm) and mode-locked lasers in the 1.5 µm region.

2. SA preparation and characterization

Figure 1(a) shows the molecular structure for Alq₃. Alq₃ powder which was purchased from Sigma Aldrich (product no. 697737-1G) [46] was embedded into polyvinyl alcohol (PVA) polymer to form a thin film SA. The powder has a purity of 99.995% and molecular weight of 459.43 g/mol. PVA is a suitable host to produce high-performance SA due to its favourable physical properties such as biocompatibility and good chemical resistance [47,48]. Additionally, PVA has an excellent film forming, mechanical and adhesive properties. The Alq₃:PVA thin film was prepared by drop casting technique at room temperature. Briefly, 10 mg of Alq₃ powder was fully dissolved (except for few particles) in 1 ml distilled water by stirring at 50^o C for one hour. Then, 2 drops of acetone were added to the solution to ensure equal distribution of Alq₃ inside the solution. On the other hand, the PVA was prepared by adding 1 g of PVA to 100 ml distilled water then the solution was ultrasonicated for one hour at room temperature. After that, the Alq₃ solution prepared previously was mixed with 5 ml of PVA solution and stirred for about three hours at room temperature. The mixture was then poured into a petri dish and dried at room temperature for three days to form a thin solid Alq₃:PVA film. The thickness of the thin film was around 50 µm and it was cut into a tiny piece and attached to a fiber ferrule.

 

Fig. 1. (a) Molecular structure of Alq₃. (b) Optical absorption spectrum of Alq₃: PVA thin film (c)Tauc plot of Alq₃: PVA thin film where dashed lines represent the extrapolation of the linear part of the curve to energy axis to calculate the band gaps. (d) Scanning electron microscopy (SEM) image of the SA, (e) Nonlinear transmission and (f) linear optical absorption spectrum of Alq₃: PVA SA

Download Full Size | PDF

The optical absorbance spectrum of the Alq₃:PVA thin film was measured using a spectrophotometer (Perkin Elmer, Lambda 750), see Fig. 1(b). The figure shows two peaks at around 300 and 385 nm which are assigned to the $\pi \to {\pi ^{\ast }}$ transition of PVA and Alq₃ materials, respectively due to unsaturated bonds [49–51]. The band gaps of the Alq3 and the modified PVA thin film were determined, from the Tauc plot shown in Fig. 1(c), to be about 2.5 and 2.4 eV, respectively [52,53]. The bandgap value of Alq3 is in good agreement with the previous literature [54]. The optical band gap of Alq3 can be detected due to some undissolved particles of Alq3 on the Alq3: PVA thin film which has been verified through the scanning electron microscopy (SEM) (The size of these particles was between 2–6.5 µm). Further research is required to address the particles size distribution and their effect on the optical properties. Figure 1(d) illustrates the SEM image of the fabricated thin film. Figure 1(e) illustrates the nonlinear transmission of Alq3: PVA SA. The saturation intensity and modulation depth of the SA were 3 MW/cm2 and 8.1%, respectively. The figure shows that the transmission increases with the power intensity, which is a typical characteristic of saturable absorption. The linear optical absorption of the SA is shown in Fig. 1(f). The SA shows a flat absorption of about 2% in the range from 900–1600 nm. It should be noted that even though the bandgap of Alq₃ is about 2.5 and 2.4 eV (which means that absorption peaks of Alq₃ is located near 500 nm), the saturable absorption was observed near 1000 nm which is sub-bandgap absorption. The sub-bandgap absorption is attributed to some defects in the material. The evidence of these defects and their physical origin in Alq₃ is addressed in Ref. [55]. These defects create states in the bandgap. These states absorb sub-bandgap photons and affect the absorption spectra, causing artifacts and features that interpreted as lower bandgap than the actual non-defected value. These defects lead to decreasing the optical energy gap and shifting the absorption edge towards the higher wavelength of the incident [56]. A similar case has been reported in Ref. [48]. The authors used WS₂ as SA in which it has a direct bandgap of ∼2.0 eV (∼630 nm) and the indirect bandgap is ∼1.4 eV (∼886 nm) [57], but they still produced a mode-lock operation in 1.5 µm region. The authors attributed the sub-bandgap absorption to some defects as well as absorption of edge modes and two-photon absorption (TPA). MoS₂ also has a direct bandgap of ∼1.8 eV (688 nm) and the indirect bandgap is 0.86 - 1.29 eV (1443 -962 nm) [56], but the sub-bandgap absorption was observed in which the saturable absorption property was observed in a wide band beyond this limitation. Further investigation is still required to understand sub-bandgap absorption in Alq₃.

Based on the modulation depth result, it is expected that the recovery and response time of Alq₃ SA were fast enough to allow for both Q-switching and mode-locking pulses generation. In more details, the recovery time should be between the pulse duration (of the Q-switching operation) and the upper-state lifetime of the gain medium. As mentioned above, this SA was used to produce a mode-locking operation. It was achieved by balancing between- dispersion and nonlinearity (i.e. 20 m SMF was added to the cavity). The pulse width produced was in picoseconds, while the repetition rate was ∼ 7 MHz. The short pulse width is attributed to the very fast recovery time of the SA, resulting in a fast loss modulation.

3. Laser configuration

The cavity setup is presented in Fig. 2. A 1.40 m of ytterbium-doped fiber (ytterbium ions concentration: 1500 ppm, core diameter: 4 µm, numerical aperture: 0.20) was used as a gain medium. The cavity length was approximately 25 m. The cavity length was then reduced to 5 m in a second experiment. The ring cavity was coupled with a 980-nm laser diode (LD) through a 980/1064 nm wavelength-division multiplexer (WDM). A 50:50 optical coupler (OC) was spliced to another end of the YDF to allow half of the output to be extracted for performance monitoring. The remaining output was directed to the Alq₃: PVA thin film which was sandwiched between two fiber ferrules. An optical isolator was addressed to eliminate back-reflection. The characteristics of the laser were analyzed using a digital oscilloscope with 2 GHz photodetector, an optical spectrum analyzer and a 7.8 GHz radio-frequency (RF) spectrum analyzer.

 

Fig. 2. Q-switched YDFL cavity setup

Download Full Size | PDF

4. Results and discussion

4.1 High pulse energy Q-switched YDFL

At first, a continuous wave (CW) emission was produced at input LD power of 60 mW. While a self-starting and stable Q-switched YDF laser operation was recorded at threshold LD power of 84 mW. Figure 3 describes the performance of the Q-switched YDFL operation. The Q-switched fiber laser showed a high-energy, relatively high output-power and stable performance as the LD power was raised to 147 mW. Figure 3 (a) shows the output spectrum of the Q-switching operation centered at 1066.7 nm with a 3 dB bandwidth of 0.2 nm.

 

Fig. 3. Characteristic of Q-switched YDFL in 25 m cavity. (a) optical spectrum, (b) pulse train at the maximum input LD power, (c) pulse duration and pulse repetition rate and as a function of input LD power, (d) pulse energy and average output power as a versus input LD power, (e) peak power against LD power and (f) Radio frequency spectrum, inset figure shows SNR at span of 40 kHz.

Download Full Size | PDF

Figure 3 (b) shows the temporal characteristics of the Q-switched pulse trains at a maximum input pump power, the inset figure shows the pulse train in the time scale of 3.5 ms. The inset figure shows a uniform intensity distribution without amplitude modulation nor fluctuation (as almost all the pulses have an amplitude of 168 mV $\mp $ 0.05 mV when they were recorded at a sampling period of 0.7 µs), which indicates the fiber laser has a stable performance. The pulse repetition rate raised from 12.95 kHz to 29.62 kHz and the pulse duration shortened from 24 µs to 9 µs as the input LD power was tuned from threshold to maximum value, i.e. from 84 mW to 147 mW, see Fig. 3 (c). Increasing the input pump power (up to 249 mW) caused the pulse train to be unstable and then disappeared. That might be attributed to the SA being over-saturated, as similar behaviour has been reported previously [58–60]. The pump power was further increased to maximum available input of 350 mW, when the LD power was gradually reduced, the Q-switched operation was observed again. Hence, we concluded that the thermal damage of the SA is higher than 350 mW. The average output power of the cavity was relatively high. That is attributed to the high performance and low insertion loss of the SA and low inter-cavity loss [61]. Additionally, a large portion of signal power was extracted from the cavity due to the high taping ratio of the OC. The average output power increased almost linearly from 4 mW to 24 mW with an increase in the input pump power, see Fig. 3 (d). The slope efficiency of our YDF laser is 30%. The setup produced high pulse energy ranging from 0.3 to 0.81 µJ as the pump power was increased from 84 mW to 147 mW. The high pulse energy is attributed to the high average output power and low repetition rate. The maximum pulse energy is 0.81 µJ, which is the is the very high among passive Q-switched fiber lasers in 1 µm region as reported so far [62–65]. As shown in Fig. 3 (e), the peak power also increased almost linearly from 12.9 to 90 mW within the given pump power range. To check the stability of Q-switched laser operation, the radio frequency (RF) was measured with a span of 300 kHz, see Fig. 3 (f). The inset figure shows the fundamental RF peaks at 29.62 kHz with an optical signal to noise ratio (SNR) of about 50 dB, which indicates excellent Q-switching stability.

4.2 Short pulse width Q-switched YDFL

In this experiment, the fiber laser cavity length was reduced to 5 m. Correspondingly, the pulse width was reduced, and the repetition rate was increased, affecting both the pulse energy and the peak power. The CW emission was observed at the input LD power of 50 mW, while the threshold and maximum LD power for generating the Q-switching were 87 mW and 147 mW, respectively. The optical spectrum showed almost similar shape in comparison with the previous cavity setup with slightly broader 3-dB bandwidth of about 0.4 nm, as can be seen in Fig. 4 (a). The centre wavelength slightly shifted to a shorter wavelength to compensate for the round-trip time difference due to the change in physical cavity length. Figure 4 (b) shows the temporal characteristics at the maximum input LD power. The pulse train is more stable and uniform than the one in the previous setup which indicates the performance and stability of the fiber laser is enhanced. The inset presents the pulse train at a time scale of 2.5 ms. The repetition rate was raised from 25.2 kHz to 66.3 kHz and the pulse duration was shortened from 9.9 µs to 1.9 µs as the input LD power was raised from threshold to the maximum value, see Fig. 4(c). The pulse repetition rate was increased as compared to the previous setup, as the decrease in the cavity length resulted in a shorter time for a pulse to travel through the cavity. Besides, the saturation of the SA is more rapid as compared to the 25 m cavity, resulting in a shorter pulse width and larger repetition rate. The output power was enhanced in comparison to the previous setup due to lower optical loss with shorted cavity length. This setup produced an output power of 8.62 mW which is about twice the power produced in the previous experiment at the same input LD power. At the same time, higher slope efficiency was obtained, which is 33%. As can be seen from Fig. 4(d), the output power reached the maximum value of about 30 mW at maximum input LD power. The maximum pulse energy achieved in this setup is 451.5 nJ which is about half of the value in the previous setup. That can be attributed to the pulse energy being inversely proportional to the pulse repetition rate. The increment in the pulse repetition rate resulted in a drop in the pulse energy.

 

Fig. 4. Characteristic of Q-switched YDFL in 5 m cavity: (a) optical spectrum, (b) pulse train at the maximum LD power, (c) pulse repetition rate and pulse width duration versus input LD power, (d) output power and pulse energy as a function of input LD power, (e) Peak power versus input LD power (f) and RF spectrum, inset figure shows SNR at span of 80 kHz.

Download Full Size | PDF

The peak power was enhanced tremendously in this setup reaching linearly to a maximum value of 237.62 mW at maximum input LD power of 147 mW, see Fig. 4(e). That can be attributed to the peak power being correlated inversely with the pulse width and proportionally with the pulse energy. In this setup, the pulse width was decreased by about eight times while the pulse energy was reduced only to the half as compared with the previous setup under similar considerations. Figure 4(f) shows an improvement in Q-switching stability. The SNR was recorded to be 52.78 dB, see the inset figure. Additionally, the RF spectrum shows more harmonics up to 3 MHz which further indicates better stability of fiber laser.

The reduction in RF intensities (in Figs. 3 (f) and Figs. 4 (f)) confirms the Q-switching operation, which has a relatively bigger pulses width than the mode locking. To check long the term stability of the fiber laser, the output spectrum was sampled every $\sim $ 12 minutes for the total duration of 138 minutes, see Fig. 5. The samples were almost identical, as the peaks were located at 1063 nm with an optical power intensity of $- $22 dBm $\mp \;0.4$ dBm. This further confirms the stability of the pulsed laser and shows that the SA can withstand long term operation. An optimized Q-switching operation can be obtained by further reducing the cavity length and/or enhancing the cavity design [61,66]. To the best of our knowledge, this work has achieved (by using SA based compact all-fiber ring cavity and through two different setups) a very high pulse energy and a very high peak power at 1 µm region. Additionally, this is the first demonstration of passive Q-switched fiber laser using an organic material as SA in 1 µm region.

 

Fig. 5. Shows the Optical spectrum output for the total duration of 135 minute

Download Full Size | PDF

5. Conclusion

This work has demonstrated two different experimental setups of YDFL Q-switched fiber laser based on Alq₃ as SA. A stable and high-performance Q-switching operation was successfully achieved in both experiments. In 25 m laser cavity, the Q-switched fiber laser produced maximum pulse energy and peak power of 0.81 µJ and 90 mW, respectively. By reducing the cavity length from 25 m to 5 m in the second experiment, the pulse width was reduced from 9 µs to 1.9 µs and peak power was enhanced tremendously to the maximum value of 237.62 mW, but the pulse energy was reduced to 0.45 µJ. The pulse energy of 0.81 µJ and the peak power of 237.62 mW represent a very good results in any SA-based Q-switched YDF lasers using a ring cavity. The organic SA showed a promising performance as passive Q-switcher at the 1 µm region.