Open Access
Add to BibSonomy
Abstract
We present an efficient tunable all-silica-fiber 2nd-order cascaded Raman pulse laser utilizing 2-µm dissipative-soliton-resonance (DSR) rectangular pulses for pumping and highly GeO2-doped silica fiber as Raman gain medium. When pumped at 1966.5 nm, the maximum 1st-order Raman optical conversion efficiency is up to 64.4% at 2153 nm, with 92.4% spectral purity and 0.39-W average power. The maximum 2nd-order Raman optical conversion efficiency is 19.3% at 2370 nm, with 39.2% spectral purity and 0.25-W average power. To our knowledge, these conversion efficiencies and spectral purities represent the highest levels achieved in a mid-infrared all-silica-fiber cascaded pulsed Raman laser. Additionally, by adjusting the central wavelength of the DSR seed pulse, the 2nd-order Raman light can be tuned within a range of 41 nm (2354∼2395 nm). Our system provides a simple and easy-to-implement solution for realizing efficient tunable cascaded pulsed Raman lasers in the 2.4-µm band.
© 2024 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
The mid-infrared (MIR) spectral range holds significant technical and scientific importance as a majority of molecules exhibit fundamental vibrational absorption bands within the MIR region, resulting in unique spectral fingerprints [1]. This distinctive feature positions MIR pulsed fiber lasers as indispensable tools in various fields such as spectroscopy, medical surgery, military radar, and industrial processing [2,3]. Typically, rare-earth-ion-doped (Tm-, Ho-, Er-, and Dy-doped) fiber lasers are employed to generate MIR lasers directly. Nevertheless, the limited emission bandwidth of these gain fibers confines the output wavelength of these fiber lasers to the range below 2.1 µm [4] and within the 2.7∼3.7 µm [5,6] bands. To address this limitation, Raman fiber lasers (RFLs) offer a promising solution. RFLs utilize Raman gain fiber as the nonlinear medium to transfer energy from pump light to long-wave Raman light through the stimulated Raman scattering (SRS) process. In principle, RFLs can emit laser pulses at any desired wavelength across the entire transmission window of the optical fiber. To date, some studies have utilized soft glass fibers as Raman gain media to expand the wavelength range [7–10]. In 2012, Fortin et al. demonstrated a fluoride glass RFL with a Raman wavelength of 2231 nm and an output power of 3.7 W [7]. In 2014, Bernier et al. presented an As2S3-based cascaded RFL with a 2nd-order Raman wavelength of 3.77 µm [8]. In 2022, Jiao et al. employed a novel fluorotellurite fiber as the Raman gain medium and observed the 3rd-order cascaded Raman shifts under 1.55-µm nanosecond pulsed pumping [10]. The 1st-, 2nd-, and 3rd-order Raman wavelengths were 1765, 2049, and 2438 nm, respectively. Although various MIR Raman laser sources have been developed using diverse soft glass fibers, all-silica-fiber structured Raman laser sources still possess significant advantages in terms of operational stability and potential compactness at 2∼3 µm region.
Highly GeO2-doped silica fibers have garnered the attention of researchers owing to their low transmission loss, high damage resistance threshold, and ease of all-fiber integration. Using GeO2-doped silica fiber as the Raman medium, a variety of silica-fiber-based RFLs have been developed by researchers [11–15]. In 2008, Rakich et al. demonstrated the first all-silica-fiber cascaded Raman pulse laser pumped by 1.53-µm gain-modulated nanosecond pulses [11]. They achieved 5th-order cascaded Raman shifts, with the maximum average power of 24 mW at 2.41 µm, corresponding to a 5th-order Raman conversion efficiency of 16%. In 2015, Jiang et al. observed 2nd-order Raman shifts at >2.4 µm in an ultra-high numerical aperture (NA) silica fiber (UHNA7, Nufern) pumped by 2-µm actively Q-switched pulses [12]. The 2nd-order Raman conversion efficiency is 16.5% at 2.43 µm (pumped at 2.008 µm) and 7.9% at 2.48 µm (pumped at 2.04 µm). In 2023, Wang et al. proposed a Raman fiber laser utilizing noise-like pulses (NLPs) for pumping [15]. The 1st and 2nd Raman stokes lights, centered at 2148 nm and 2359 nm, were obtained separately with low conversion efficiencies (specific values were not provided in the paper). It is worth noting that all the aforementioned silica-fiber-based Raman laser systems exhibit notable pump residuals in the total output, which affects the Raman spectral purity as well as the pump-to-Raman conversion efficiency.
Actually, compared to other types of pulses, the square-wave pulse is more conducive to transferring energy from pump light to long-wave Raman light through the SRS process [16,17]. Since the peak power of the square-wave pulse is nearly constant over the entire time profile, it is beneficial to provide constant Raman gain. The generation of square-wave pulses typically involves the use of modulators or other pulse-shaping components [18,19], leading to increased complexity and cost in the laser system. Fortunately, in 2008, Chang. et al. introduced the concept of dissipative soliton resonance (DSR) [20], enabling the direct generation of pulse-breaking-free square-wave pulses in mode-locked fiber lasers. With the elevation of pump power, the DSR pulse maintains an almost constant peak power while its width can be broadened linearly [21–23]. More importantly, the DSR pulse eliminates the need for an additional pulse-shaping system in the pulse oscillator or amplifier stage [24]. Recently, our group demonstrated an efficient 1st-order Raman conversion from 1987 to 2177 nm by employing DSR pulses for pumping [25]. The potential of DSR pulses in realizing highly efficient cascaded RFLs remains to be explored. In addition, employing a simple method to achieve tunability in Raman wavelengths will significantly enhance the versatility of the Raman laser system.
In this paper, we present an efficient tunable all-silica-fiber 2nd-order cascaded Raman pulse laser, utilizing DSR pulses for pumping. A Tm-doped fiber laser mode-locked by nonlinear amplifying loop mirror (NALM) is constructed to generate DSR square-wave pulses. When pumped at 1966.5 nm, the maximum optical conversion efficiency of the 1st-order Raman light at 2153 nm is 64.4%, with a spectral purity of 92.4% and an average power of 0.39 W. Additionally, the maximum optical conversion efficiency of the 2nd-order Raman light at 2370 nm reaches 19.3%, with a spectral purity of 39.2% and an average power of 0.25 W. The repetition frequency of the output Raman pulse is 656.3 kHz. By tuning the central wavelength of the DSR seed pulse, the 2nd-order Raman light can be adjusted within a 41-nm range (2354∼2395 nm).
2. Experimental setup
The schematic of the all-silica-fiber 2nd-order cascaded pulsed Raman source is shown in Fig. 1. It consisted of a Tm-doped seed source, two-stage Tm-doped fiber amplifier (TDFAs), and a ∼30-m-long Raman gain fiber (UHNA7, Thorlabs). The Tm-doped seed source is composed of a NALM and a unidirectional ring (UR), interconnected through a 2 × 2 50/50 fiber coupler. The NALM acts as an artificially saturable absorber with a reverse saturable absorption (RSA) effect, playing a crucial role in the generation of DSR pulses. Besides, NALM mode-locked DSR fiber lasers offer superior wavelength tunability [26] compared to nonlinear polarization rotation mode-locked ones. The NALM comprises a 2-m-long Tm-doped double-clad fiber (TDF1, CorActive DCF-TM-10/128, with a dispersion coefficient of -84 ps2/km @1950nm), a polarization controller (PC1), a ∼180-m-long standard single-mode fiber (SMF1, Corning SMF-28e, with a dispersion coefficient of -80 ps2/km @1950nm). The 180-m SMF1 is employed to enhance the nonlinear phase shift difference between pulses traveling in opposing directions within the NALM, thereby reducing the mode-locking threshold of the DSR fiber laser. The TDF1 is pumped by a 793-nm multi-mode laser diode (LD1) through a 793/1950-nm pump/signal combiner. The UR contains a PC (PC2), a 70/30 fiber output coupler (OC1), a polarization-independent isolator (PI-ISO), and a ∼120-m SMF2. The role of the PI-ISO is to guarantee unidirectional propagation of the pulse in the UR, and the OC1 is employed to extract 30% of the intracavity pulse energy as a seed for subsequent applications. Additionally, intracavity PCs, i.e., PC1 and PC2, are utilized to fine-tune the net-cavity birefringence. The total cavity length is about 309.3 m and the net-cavity dispersion is calculated to be -24.8 ps2. Subsequently, the seed pulse is amplified by two-stage homemade TDFAs. The 1st-stage TDFA is fabricated by a 793-nm multi-mode LD2, another pump/signal combiner, and a 1.5-m TDF2 (Nufern, SM-TDF-10P/130-HE). The TDF2 has a core/cladding diameter of 10/130 µm, a core/cladding NA of 0.15/0.46, and a cladding peak absorption factor of ∼3.6 dB/m at 793 nm. The structure of the 2nd-stage TDFA is similar to the 1st-stage TDFA, while the length of TDF3 is 3 m. The PI-ISO used in each TDFA stage is to prevent harmful feedback light. At the output port of the 2nd-stage TDFA, a 0.3-m single-clad SMF3 (Nufern, SM-1950) is fused behind the amplifier to eliminate the unabsorbed 793-nm pump light. Lastly, a ∼30-m UHNA7 fiber is spliced to the SMF3. The UHNA7 fiber features a normal dispersion below 2.6 µm [12], which is beneficial for avoiding significant spectral broadening. The two-stage TDFAs and UHNA7 are water-cooled to 16 °C on an aluminum plate for efficient heat dissipation.
Fig. 1. Experimental setup of our proposed all-silica-fiber cascaded pulsed Raman laser system. LD: laser diode, TDF: Tm-doped fiber, SMF: single-mode fiber, PC: polarization controller, OC: output coupler, PI-ISO: polarization-independent isolator.
The waveform of the seed pulse is monitored using a 2-GHz digital oscilloscope in combination with a 12.5-GHz InGaAs photodetector (EOT, ET-5000F). The radio frequency (RF) spectrum is captured with an RF spectrum analyzer (Rohde & Schwarz, FSV3007). Both the seed spectrum and Raman output spectrum are recorded by an optical spectrum analyzer with a measurement range of 1900∼5500 nm (Yokogawa, AQ6377). The output power is measured employing a MIR power meter (Thorlabs).
3. Results and discussion
3.1 Generation and amplification of DSR pulses
Due to the higher loss of the SMF-28e fiber in the 2-µm band (∼24.3 dB/km) [27] and the 30% output coupling ratio, the mode-locking threshold of the thulium-doped seed source is elevated, approximately 2.5 W. By finely adjusting PC1 and PC2, a stable mode-locked pulse with an envelope closely resembling a rectangle can be attained. Figure 2(a) shows an oscilloscope trace of the output pulse train. The temporal interval between successive pulses is approximately 1.52 µs, corresponding to 309.3-m cavity length. In the inset of Fig. 2(a), a magnified view of an individual pulse reveals a rectangular profile with a width of ∼8.2 ns. In our experiments, the profile of the rectangular pulse is stable, with only slight fluctuations in its flat-topped portion during long-term operation. Figure 2(b) displays the pulse spectrum, centered at 1966.5 nm, with a 3-dB spectral bandwidth of 6.89 nm. The inset of Fig. 2(b) shows the autocorrelation trace measured using a second-harmonic autocorrelator (FR-103XL). No coherent peak is observed on a broad pedestal, thus ruling out the possibility of rectangular noise-like pulses [21]. In Fig. 2(c), The RF spectrum illustrates a signal-to-noise ratio (SNR) of about 64.6 dB for the output pulses at a fundamental frequency repetition rate of 656.3 kHz, which suggests that the mode-locking operation has good stability. In the inset of Fig. 2(c), the RF spectrum within the 450-MHz range is depicted, revealing a modulation period of approximately 122 MHz, corresponding to a pulse width of 8.2 ns. To study the temporal evolution of the pulse with pump power, the orientations of PC1 and PC2 are kept unchanged and LD1 pump power is incrementally raised at fixed intervals. As shown in Fig. 2(d), the output power and pulse width are linearly related to the LD1 pump power, while the peak pulse power remains constant (∼2.5W) during the evolution. The slope of pulse width versus the LD1 pump power is ∼17 ns/W. The experimental results mentioned above align with the typical characteristics of DSR pulses [21].
Fig. 2. Characterization of the DSR seed pulse. (a) Pulse sequence. Inset: single pulse; (b) Optical spectrum. Inset: autocorrelation trace; (c) RF spectrum. Inset: RF spectrum covers a range of 450 MHz; (d) Pulse output power, width, and peak power versus LD1 pump power.
In our experiments, it is found that the width of the DSR seed pulse is an important parameter for realizing efficient cascaded Raman conversion. At fixed pump powers of amplifiers, a wide DSR-pulse seed would lead to low peak power in amplified DSR pulses, preventing the initiation of cascaded Raman shifts. Conversely, if the DSR-pulse seed is too narrow, the peak power of amplified DSR pulses becomes excessively high, rendering it prone to generating a supercontinuum in UHNA7 fiber. In pursuit of efficient cascaded Raman conversion, the width of the DSR seed pulse is carefully optimized, and a width of 8.2 ns is ultimately chosen. The DSR seed pulse then enters the 1st-stage TDFA for pre-amplification. It is worth noting that 1st-stage TDFA operates at a low LD2 pump power level, which is used to amplify the average power of the DSR pulse from ∼13.6 to ∼121.3 mW. If the LD2 pump power is set too high, the final output Raman spectrum would undergo significant broadening. This phenomenon may be related to the saturation effect of the amplifier under excessive LD2 pump power [28]. In such instances, the leading edge of the DSR pulse gains more than the trailing edge, causing pulse distortion into an ‘h’ shape. In this scenario, the higher peak power pulse front triggers deleterious nonlinear effects, ultimately resulting in spectral broadening.
Figure 3(a) depicts the average and peak power of amplified DSR pulses at the output of the 2nd-stage TDFA versus the LD3 pump power. When the LD3 pump power is increased from 2 to 7 W, the average power of amplified DSR pulses rises from 331.6 to 1720.2 mW, the corresponding peak power increases from 61.6 to 319.6 W, with a maximum single-pulse energy of 2.62 µJ. In addition, the power stability of amplified DSR pulses is measured for one hour, as seen in Fig. 3(b). The standard deviation (STD) of the output power is 7.2 mW, corresponding to a power fluctuation of ∼0.42%, indicating high power stability of amplified DSR pulses. In addition, the pulse-to-pulse variations in the peak power of the amplified DSR pulses are measured. The results show a ∼1.4% peak-to-peak fluctuation of the output pulse train. To examine the impacts of the two-stage TDFAs on the temporal waveform of the DSR pulse, the waveforms at the input and output ports of the amplifier are compared, as shown in the inset of Fig. 3(b). Notably, no evident deformation or broadening is observed in the amplified DSR pulses. Previous studies have revealed that the square-wave pulses generated through the active modulation techniques tend to experience distortion after amplification due to the saturation effect. Therefore, pre-compensation is necessary for the seed pulse [18,28]. In contrast, our approach to preventing pulse distortion involves optimizing the pump powers of the TDFAs, ensuring they stay away from the saturation region. The resultant amplified DSR pulses still retain the initial pulse profiles, which would facilitate the efficient cascaded Raman shifts.
Fig. 3. Characterization of amplified DSR pulses at the output of the 2nd-stage TDFA. (a) Output power and peak power of amplified DSR pulses versus LD3 pump power; (b) Output power stability of amplified DSR pulses in one hour with the LD3 pump power of 7 W. STD: standard deviation. Inset: a comparison of DSR pulse waveforms at the input and output ports of the amplifier. Red line: waveform of the amplified DSR pulse, black line: waveform of the DSR seed pulse.
3.2 Generation of efficient cascaded Raman shifts
In our experiments, a coupling loss of approximately 18.9% occurs between the SMF3 and the UHNA7 fiber due to the mode-field mismatch. Consequently, the maximum peak power of the amplified DSR pulse launched into the UHNA7 fiber decreases from 319.6 to 259.2 W. The classic SRS threshold formula is given by Pth≈16Aeff/gRLeff, where Pth denotes the SRS threshold power, Aeff signifies the effective fiber mode field area, gR represents the Raman gain coefficient, and Leff stands for the effective fiber length [29]. This formula implies an inverse proportionality between Pth and Leff. In our laser system, a ∼30-m UHNA7 fiber serves as the Raman gain medium, allowing us to achieve Raman shifts at a relatively low incident peak power. At the output of the UHNA7 fiber, the optical spectra evolution of total output (including the 1st-order Raman light, 2nd-order Raman light, residual pump, and noise) with the peak power Pin of the incident DSR pulse is displayed in Fig. 4(a). As Pin increases from 50 to 259.2 W, the pump component is gradually transformed into 1st- and 2nd-order Raman signals. Moreover, a distinct boundary distinguishes the 1st- and 2nd-order Raman signals, indicating effective suppression of Raman spectrum broadening. In comparison to the utilization of Gaussian-type pulses as the pump [12], DSR square-wave pulses are employed for pumping in our cascaded pulsed Raman laser, ensuring constant Raman gain over the entire temporal profile. As a result, more pump components could be converted into Raman signals through the SRS process. To enhance clarity, the output spectra at Pin of 112.4 W and 244.9 W are plotted separately in Fig. 4(b). The black curve represents the case at a Pin of 112.4 W, revealing a high 1st-order Raman spectral purity (defined as the energy ratio of the Raman spectral component in the total output spectrum) of approximately 92.4% at ∼2153 nm. The blue curve corresponds to the case at a Pin of 244.9 W, signifying a significant increase in the 2nd-order Raman spectral purity, reaching about 39.2% at ∼2370 nm. Subsequently, the spectral integration method is utilized to determine the energy ratio of the Raman components in the overall output. This approach enables the separate calculation of the average power for the 1st- and 2nd-order Raman lights. Figure 4(c) illustrates the laser output power versus the average power Pavg of incident DSR pulses. It can be seen that the output power of the 1st-order Raman light initially increases to a maximum value, followed by a decrease. This can be attributed to the transformation of a portion of the 1st-order Raman light into the 2nd-order Raman light at higher Pavg. At a Pavg of ∼604.9 mW, the average power of the 1st-order Raman light is ∼389.7 mW, corresponding to a 1st-order pump-to-Raman conversion efficiency of 64.4% (with a high 1st-order Raman spectral purity of 92.4%). The average power of the 2nd-order Raman light increases with Pavg. However, the maximum pump-to-Raman conversion efficiency for the 2nd-order Raman light is not achieved under the highest Pavg due to increased fiber losses at longer wavelengths. At a Pavg of ∼1317.9 mW, the 2nd-order Raman light attains an average power of ∼254.4 mW, which corresponds to a maximum 2nd-order pump-to-Raman conversion efficiency of 19.3% (with a 2nd-order Raman spectral purity of 39.2%). To the best of our knowledge, we have achieved the highest 1st- and 2nd-order Raman conversion efficiencies and spectral purities in a MIR all-silica-fiber cascaded Raman laser. In addition, to investigate the long-term stability of the Raman laser system, the total output is recorded for one hour at the maximum Pavg, as depicted in Fig. 4(d), The STD of the output power is 5.1 mW, corresponding to a power fluctuation of ∼0.79%. The results indicate that our proposed all-silica-fiber cascaded Raman pulse laser maintains high power stability.
Fig. 4. Output characteristics of the cascaded pulsed Raman laser. (a) Output spectra versus peak power of the DSR pulse (8.2-ns pulse width) launched into UHNA7 fiber; (b) Output spectra at incident peak powers of 112.4 and 244.9 W, respectively. Black curve: 112.4 W, blue curve: 244.9 W; (c) The average powers of the total output, the 1st-order Raman light, the 2nd-order Raman light, as well as residual pump and noise versus average power of incident DSR pulses. Note: the noise refers to the amplified classical pump noise and the amplified quantum noise; (d) Power stability of total output in one hour at the maximum incident average power of DSR pulses. Note: the spectra of (a) and (b) are in linear coordinates.
3.3 Wavelength tunability of the cascaded Raman fiber laser
Without inserting any additional tunable filter inside the Tm-doped DSR seed source, the central wavelength of the DSR seed pulse can be adjusted from 1956.1 nm to 1984.1 nm (covering a range of 28 nm) by fine-tuning the intracavity PCs. As shown in Fig. 5(a), the central wavelengths of the DSR pulses are centered at 1956.1 nm, 1962.4 nm, 1966.5 nm, 1972.2 nm, 1978.6 nm, and 1984.1 nm for different intracavity PC states. The mechanism of wavelength tunability may be attributed to the fiber birefringence-induced filtering effect [26,30]. Generally, SMF and TDF possess relatively weak inherent birefringence. However, when these fibers are coiled into multiple loops, they introduce additional birefringence into the laser cavity. Consequently, the artificial birefringence filter effect within the cavity becomes non-negligible. Furthermore, the manipulation of intra-cavity PCs, leading to fiber bending and squeezing, could also induce a change in birefringence strength. This dynamically influences the position of the filter transmission peak. It's noteworthy that the fiber laser attains its maximum gain at the transmission peak of the spectral filter. Since it is easier for the laser to establish mode-locking at the gain peak, adjusting the intra-cavity PCs provides a simple approach to achieving wavelength-tunable DSR pulses. As shown in Fig. 5(b), different cascaded Raman lasing wavelengths can be achieved by employing DSR seed pulses with different central wavelengths. The shortest wavelength for 2nd-order Raman light is 2354 nm while the longest wavelength reaches 2395 nm, corresponding to a tuning range of 41 nm. This is the first report of a tunable all-silica-fiber cascaded Raman laser in the 2.4-µm band. Our proposed tunable cascaded Raman laser system eliminates the need for additional tunable filters, reducing the system complexity. Furthermore, it exhibits an advantage in wavelength tunability compared to conventional fixed-wavelength cascaded pulsed Raman lasers, thereby extending its applications in the 2.4-µm band.
Fig. 5. (a) The wavelength tuning operation of the DSR seed source; (b) The wavelength tuning range for cascaded pulsed RFL. Note: the spectra of (b) are in linear coordinates.
4. Conclusion
In conclusion, we have demonstrated an efficient tunable all-silica-fiber 2nd-order cascaded Raman pulse laser utilizing DSR pulses for pumping. When pumped at 1966.5 nm, the 1st-order Raman light at 2153 nm has achieved a maximum optical conversion efficiency of 64.4%, corresponding to a spectral purity of 92.4% and an average power of 0.39 W. Furthermore, the 2nd-order Raman light at 2370 nm has reached a maximum optical conversion efficiency of 19.3%, with a spectral purity of 39.2% and an average power of 0.25 W. As far as we know, these conversion efficiencies and spectral purities represent the highest levels achieved in an all-silica-fiber MIR cascaded pulsed Raman laser. By tuning the central wavelength of the DSR seed pulse, the 2nd-order Raman light can be adjusted over a range of 41 nm (2354∼2395 nm). We believe this work provides a prospective solution for realizing efficient tunable cascaded pulsed Raman lasers in the 2.4-µm band. Further power scaling, efficiency enhancement, or wavelength expansion of cascaded Raman pulse fiber lasers based on DSR pulse pumping can be achieved through the utilization of multistage amplifiers or innovative MIR Raman fibers with wider transmission windows.
Funding
National Natural Science Foundation of China (U20A20210, 61927821, 62005040, 61775031); Fundamental Research Funds for the Central Universities (ZYGX2021YGCX014, ZYGX2019Z012, ZYGX2020KYQD003); Sichuan Province Science and Technology Support Program (2023NSFSC0033, 2023NSFSC1964).
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. C. R. Petersen, U. Møller, I. Kubat, et al., “Mid-infrared supercontinuum covering the 1.4–13.3 µm molecular fingerprint region using ultra-high NA chalcogenide step-index fibre,” Nat. Photonics 8(11), 830–834 (2014). [CrossRef]
2. M. Yumoto, N. Saito, T. Lin, et al., “High-energy, nanosecond pulsed Cr: CdSe laser with a 2.25–3.08 µm tuning range for laser biomaterial processing,” Biomed. Opt. Express 9(11), 5645–5653 (2018). [CrossRef]
3. S. Wolf, T. Trendle, J. Kiessling, et al., “Self-gated mid-infrared short pulse upconversion detection for gas sensing,” Opt. Express 25(20), 24459–24468 (2017). [CrossRef]
4. H. Luo, F. Liu, J. Li, et al., “High repetition rate gain-switched Ho-doped fiber laser at 2.103 µm pumped by h-shaped mode-locked Tm-doped fiber laser at 1.985 µm,” Opt. Express 26(20), 26485–26494 (2018). [CrossRef]
5. Y. Wang, H. Luo, H. Wu, et al., “Tunable pulsed dysprosium laser within a continuous range of 545 nm around 3 µm,” J. Lightwave Technol. 40(14), 4841–4847 (2022). [CrossRef]
6. H. Luo, J. Yang, J. Li, et al., “Widely tunable passively Q-switched Er3+-doped ZrF4 fiber laser in the range of 3.4–3.7 µm based on a Fe2+: ZnSe crystal,” Photonics Res. 7(9), 1106–1111 (2019). [CrossRef]
7. V. Fortin, M. Bernier, D. Faucher, et al., “3.7 W fluoride glass Raman fiber laser operating at 2231 nm,” Opt. Express 20(17), 19412–19419 (2012). [CrossRef]
8. M. Bernier, V. Fortin, M. El-Amraoui, et al., “3.77 µm fiber laser based on cascaded Raman gain in a chalcogenide glass fiber,” Opt. Lett. 39(7), 2052–2055 (2014). [CrossRef]
9. F. Wang, X. Zhou, X. Zhang, et al., “Mid-infrared cascaded stimulated Raman scattering and flat supercontinuum generation in an As-S optical fiber pump at 2 µm,” Appl. Opt. 60(22), 6351–6356 (2021). [CrossRef]
10. Y. Jiao, Z. Jia, X. Guo, et al., “Third-order cascaded Raman shift in all-solid fluorotellurite fiber pumped at 1550 nm,” Opt. Lett. 47(3), 690–693 (2022). [CrossRef]
11. P. T. Rakich, Y. Fink, and M. Soljačić, “Efficient mid-IR spectral generation via spontaneous fifth-order cascaded-Raman amplification in silica fibers,” Opt. Lett. 33(15), 1690–1692 (2008). [CrossRef]
12. H. Jiang, L. Zhang, and Y. Feng, “Silica-based fiber Raman laser at> 2.4 µm,” Opt. Lett. 40(14), 3249–3252 (2015). [CrossRef]
13. H. Jiang, L. Zhang, X. Yang, et al., “Pulsed amplified spontaneous Raman emission at 2.2 µm in silica-based fiber,” Appl. Phys. B 122(4), 1–4 (2016). [CrossRef]
14. T. Du, Y. Li, H. Wang, et al., “2166 nm all-fiber short-pulsed Raman laser based on germania-core fiber,” Opt. Express 27(24), 34552–34558 (2019). [CrossRef]
15. M. Wang, H. Wu, D. Ouyang, et al., “High power noise-like pulse at 2 µm and its applications in mid-IR Raman light and flat supercontinuum,” Infrared Phys. Technol. 131, 104635 (2023). [CrossRef]
16. K. K. Chen, S.-U. Alam, P. Horak, et al., “Excitation of individual Raman Stokes lines in the visible regime using rectangular-shaped nanosecond optical pulses at 530 nm,” Opt. Lett. 35(14), 2433–2435 (2010). [CrossRef]
17. L. Xu, S. Alam, Q. Kang, et al., “Raman-shifted wavelength-selectable pulsed fiber laser with high repetition rate and high pulse energy in the visible,” Opt. Express 25(1), 351–356 (2017). [CrossRef]
18. K. Vu, A. Malinowski, D. Richardson, et al., “Adaptive pulse shape control in a diode-seeded nanosecond fiber MOPA system,” Opt. Express 14(23), 10996–11001 (2006). [CrossRef]
19. A. Malinowski, P. Gorman, C. Codemard, et al., “High-peak-power, high-energy, high-average-power pulsed fiber laser system with versatile pulse duration and shape,” Opt. Lett. 38(22), 4686–4689 (2013). [CrossRef]
20. W. Chang, A. Ankiewicz, J. M. Soto-Crespo, et al., “Dissipative soliton resonances,” Phys. Rev. A 78(2), 023830 (2008). [CrossRef]
21. Y. Wang, L. Li, J. Zhao, et al., “Unusual evolutions of dissipative-soliton-resonance pulses in an all-normal dispersion fiber laser,” IEEE Photonics J. 11(1), 1–9 (2019). [CrossRef]
22. J. Zhao, J. Zhou, L. Li, et al., “Dissipative soliton resonance and its depression into burst-like emission in a holmium-doped fiber laser with large normal dispersion,” Opt. Lett. 44(10), 2414–2417 (2019). [CrossRef]
23. D. J. Li, D. Y. Tang, L. M. Zhao, et al., “Mechanism of dissipative-soliton-resonance generation in passively mode-locked all-normal-dispersion fiber lasers,” J. Lightwave Technol. 33(18), 3781–3787 (2015). [CrossRef]
24. Z. Zheng, D. Ouyang, X. Ren, et al., “0.33 mJ, 104.3 W dissipative soliton resonance based on a figure-of-9 double-clad Tm-doped oscillator and an all-fiber MOPA system,” Photonics Res. 7(5), 513–517 (2019). [CrossRef]
25. F. Liu, J. Li, H. Luo, et al., “Efficient Raman pulse fiber laser pumped by a dissipative soliton resonance pulse near 2 µm,” Opt. Express 31(4), 6741–6749 (2023). [CrossRef]
26. D. Zhao, B. Zhang, S. Liu, et al., “Wavelength tunable dissipative soliton resonant holmium-doped fiber laser beyond 2.1 µm,” Opt. Commun. 537, 129395 (2023). [CrossRef]
27. M. Belal, M. Petrovich, N. Wheeler, et al., “First demonstration of a 2-µm OTDR and its use in photonic bandgap CO2 sensing fiber,” IEEE Photonics Technol. Lett. 26(9), 889–892 (2014). [CrossRef]
28. D. N. Schimpf, C. Ruchert, D. Nodop, et al., “Compensation of pulse-distortion in saturated laser amplifiers,” Opt. Express 16(22), 17637–17646 (2008). [CrossRef]
29. C. Jauregui, J. Limpert, and A. Tünnermann, “Derivation of Raman threshold formulas for CW double-clad fiber amplifiers,” Opt. Express 17(10), 8476–8490 (2009). [CrossRef]
30. Y. Xu, Y. Song, G. Du, et al., “Dissipative soliton resonance in a wavelength-tunable thulium-doped fiber laser with net-normal dispersion,” IEEE Photonics J. 7(3), 1–7 (2015). [CrossRef]
