3.96 kW All-Fiberized Linearly Polarized and Narrow Linewidth Fiber Laser with Near-Diffraction-Limited Beam Quality

In this paper, we realize a 3.96 kW all-fiberized and polarization-maintained (PM) amplifier with narrow linewidth and near-diffraction-limited beam quality. Based on a master oscillator power amplifier (MOPA) configuration seeded with phase-modulated single-frequency laser, a 3.96 kW signal laser is achieved with a 3 dB linewidth of 0.62 nm at the pump power of 5.02 kW. At the maximum output power, the polarization extinction ratio (PER) is ~13.9 dB, and the beam quality (M2 factor) is M2x = 1.31, M2y = 1.41. As far as we know, this is the maximum output power of PM narrow linewidth fiber laser with near-diffraction-limited beam quality and all-fiber format.


Introduction
Due to the advantages of compact structure, convenient thermal management and high electro-optic conversion efficiency, high-power narrow linewidth all-fiberized lasers with near-diffraction-limited beam quality have been widely used in the fields of spectral and coherent beam combinations [1][2][3], nonlinear frequency conversion [4], and remote communication [5]. Increasing brightness of such fiber lasers have always been the research focus. However, limited by nonlinear effects such as stimulated Brillouin scattering (SBS), stimulated Raman scattering (SRS) and thermally induced mode instability (TMI) effect [6][7][8], the increase of output power without degradation of beam quality of narrow linewidth fiber lasers still faces many challenges.

Experimental Setup
The experimental setup of 3.96 kW PM narrow linewidth fiber amplifier using the MOPA configuration is shown in Figure 1. The laser seed was a 50 mW, single-frequency linearly polarized, laser working at~1064 nm. The phase modulator was driven by an amplified WNS to widen the linewidth of single frequency laser for suppressing SBS effect. Subsequently, the modulated signal was power-amplified to~20 W by PM pre-amplifiers (PM AMPs). A PM circulator (AFR, Zhuhai, China) was used to detect the backward signal power for diagnosing the SBS effect. The pre-amplified signal entered the main amplifier through a PM mode field adaptor (PM MFA). The main amplifier adopted the bidirectional pumped configuration and used stable wavelength laser diodes (LDs) (BWT, Beijing, China) centering at 976 nm as the pump source. The pump light was injected into a double clad and large mode area PM Yb-doped fiber (PM YDF) (Self-made) with core/cladding diameter of 20/400 µm through two (6 + 1) × 1 PM beam combiners. The length of the YDF was~8.5 m, and the absorption coefficient of the YDF was~1.5 dB/m at 976 nm. The PM cladding power strippers (PM CPSs) at the front and rear ends of the main amplifier were employed to remove the residual cladding light in the system. Finally, the signal laser was transmitted to free-space through a PM quartz block holder (PM QBH) (Self-made).

Experimental Results
The formation of TMI was mainly due to the interaction between fundamental mode and higher order modes (HOMs), which can be effectively mitigated by increasing the relative loss of HOMs. At present, many HOMs loss methods have been proposed, among

Experimental Results
The formation of TMI was mainly due to the interaction between fundamental mode and higher order modes (HOMs), which can be effectively mitigated by increasing the relative loss of HOMs. At present, many HOMs loss methods have been proposed, among which coiling gain fiber with a certain diameter may be a simple and economical technique. Therefore, we first studied the feasibility of increasing TMI threshold by coiling gain fiber in the linearly polarized fiber amplifier using unidirectional pumped configuration. To begin with, the YDF was loosely coiled in the water-cooled plate as a circle spiral shape with a minimum diameter of 45 cm. Figure 2a,b show the temporal signals and corresponding power spectral density (PSD) of the output laser when only co-pumped power was injected. As we can see from Figure 2a,b, the temporal signal was basically stable with the standard deviation of 1.83% and the slight noise-like protuberances began to appear in the PSD when the output power was 688 W. As the output power increased to 744 W, the temporal signal demonstrated a sudden instability with the standard deviation of 3.78%, which was more than twice that at 688 W. There also existed an obvious frequency component in the PSD ranging from 0 to 10 kHz. We confirmed that the TMI threshold with loosely coiled YDF was 744 W when using the co-pumped scheme.

Experimental Results
The formation of TMI was mainly due to the interaction between fundamental mode and higher order modes (HOMs), which can be effectively mitigated by increasing the relative loss of HOMs. At present, many HOMs loss methods have been proposed, among which coiling gain fiber with a certain diameter may be a simple and economical tech nique. Therefore, we first studied the feasibility of increasing TMI threshold by coiling gain fiber in the linearly polarized fiber amplifier using unidirectional pumped configu ration. To begin with, the YDF was loosely coiled in the water-cooled plate as a circle spira shape with a minimum diameter of 45 cm. Figure 2a,b show the temporal signals and corresponding power spectral density (PSD) of the output laser when only co-pumped power was injected. As we can see from Figure 2a,b, the temporal signal was basically stable with the standard deviation of 1.83% and the slight noise-like protuberances began to appear in the PSD when the output power was 688 W. As the output power increased to 744 W, the temporal signal demonstrated a sudden instability with the standard devi ation of 3.78%, which was more than twice that at 688 W. There also existed an obviou frequency component in the PSD ranging from 0 to 10 kHz. We confirmed that the TM threshold with loosely coiled YDF was 744 W when using the co-pumped scheme. The YDF was then tightly coiled in the water-cooled plate as a racetrack spiral shape with a minimum diameter of 8.5 cm at both ends according to our prior theoretical anal ysis [42]. Subsequently, co-pumped and counter-pumped schemes were used, respec tively. The properties of the output laser are shown in Figure 3a-c. The red square and The YDF was then tightly coiled in the water-cooled plate as a racetrack spiral shape with a minimum diameter of 8.5 cm at both ends according to our prior theoretical analysis [42]. Subsequently, co-pumped and counter-pumped schemes were used, respectively. The properties of the output laser are shown in Figure 3a-c. The red square and blue round in Figure 3a represent the output power-scaling curve with co-pumped and counterpumped schemes, respectively. We found that the output power showed a linear growth with the increase of pump power when the unidirectional pumped scheme was applied. As shown in Figure 3b,c, when the co-pumped scheme was adopted, there existed a significantly enhanced frequency component in the PSD ranging from 0 to 10 kHz as the output power was from 1.712 kW to 1.761 kW. Based on the counter-pumped scheme, the frequency component in the PSD ranging from 0 to 10 kHz became stronger as the output power increased from 2.297 kW to 2.392 kW. Thus, the TMI thresholds of co-pumped and counter-pumped schemes were considered to be 1.761 kW and 2.392 kW, respectively. The above research proved that the TMI threshold can be remarkably increased by coiling the YDF in an appropriate way in the PM fiber amplifier. Simultaneously, theoretical research shows that, compared with unidirectional pumped configuration, bidirectional pumped configuration has more advantages in increasing the TMI threshold, and that there is an optimal proportion of counter-pumped power, which can maximize the TMI threshold of the amplifier [43]. Thus, in the following experiments, the coiling method of the YDF remained unchanged. Considering the mitigation of TMI effects, the main amplifier employed bidirectional pumped configuration, and counter-pumped power was appropriately proportioned to explore the power breakthrough ability of the PM fiber amplifier. tion, bidirectional pumped configuration has more advantages in increasing the TMI threshold, and that there is an optimal proportion of counter-pumped power, which can maximize the TMI threshold of the amplifier [43]. Thus, in the following experiments, the coiling method of the YDF remained unchanged. Considering the mitigation of TMI effects, the main amplifier employed bidirectional pumped configuration, and counterpumped power was appropriately proportioned to explore the power breakthrough ability of the PM fiber amplifier.   Figure  4a, with the increase of pump power, the output power showed a linear growth. A record power of 3.96 kW was realized when the pump power was 5.02 kW. The slope efficiency of the amplifier was ~79.5%. At the maximum output power, the co-pumped power was 2.167 kW, and the counter-pumped power was 2.853 kW. This co-pumped and counterpumped power distribution not only alleviated the power-handling capability of the PM combiner, but also effectively increased the TMI threshold of the amplifier. We can see from Figure 4a that the backward signal power mainly caused by the SBS effect also exhibited a linear increase trend as the amplification of the signal laser. The backward signal   Figure 4a, with the increase of pump power, the output power showed a linear growth. A record power of 3.96 kW was realized when the pump power was 5.02 kW. The slope efficiency of the amplifier was~79.5%. At the maximum output power, the co-pumped power was 2.167 kW, and the counter-pumped power was 2.853 kW. This co-pumped and counterpumped power distribution not only alleviated the power-handling capability of the PM combiner, but also effectively increased the TMI threshold of the amplifier. We can see from Figure 4a that the backward signal power mainly caused by the SBS effect also exhibited a linear increase trend as the amplification of the signal laser. The backward signal power at maximum output power was 311 mW, which was only 0.08‰ of the output power. This means that the SBS effect did not occur in the system during power scaling. Therefore, SBS was effectively suppressed by the WNS phase modulation. Figure 4b shows the PER of the signal laser. The PER changed between 16.3 dB and 13.8 dB during power amplification, and the PER was~13.9 dB at 3.96 kW. Nanomaterials 2022, 12, x FOR PEER REVIEW 5 of 8 power at maximum output power was 311 mW, which was only 0.08‰ of the output power. This means that the SBS effect did not occur in the system during power scaling. Therefore, SBS was effectively suppressed by the WNS phase modulation. Figure 4b shows the PER of the signal laser. The PER changed between 16.3 dB and 13.8 dB during power amplification, and the PER was ~13.9 dB at 3.96 kW. The spectral characteristics of the amplifier are shown in Figure 5a,b. Figure 5a is the spectra of signal laser at different output power. The 3 dB linewidth of signal laser at 3.96 kW was ~0.62 nm centering at ~1064.4 nm. The Raman stokes-shifted light around 1113 nm was observed at 3.006 kW, and its intensity gradually strengthened with the increase of the output power, but the signal-to-noise ratio (SNR) was as high as 49 dB compared with the spectral component around 1113 nm when the signal laser power was 3.96 kW. The spectral characteristics of the amplifier are shown in Figure 5a,b. Figure 5a is the spectra of signal laser at different output power. The 3 dB linewidth of signal laser at 3.96 kW was~0.62 nm centering at~1064.4 nm. The Raman stokes-shifted light around 1113 nm was observed at 3.006 kW, and its intensity gradually strengthened with the increase of the output power, but the signal-to-noise ratio (SNR) was as high as 49 dB compared with the spectral component around 1113 nm when the signal laser power was 3.96 kW. Therefore, SRS was effectively suppressed in our system. The evolution of the spectral linewidth of the signal laser through the power amplification was also described. As shown in Figure 5b, with the output power increase from 20 W (seed laser power) to 3.96 kW, the 3 dB linewidth of the signal laser changed from 0.48 nm to 0.62 nm, and the 10 dB linewidth of the signal laser varied from 1.08 nm to 1.40 nm. Overall, distinct spectral broadening did not occur during the power amplification. The spectral characteristics of the amplifier are shown in Figure 5a,b. Figure 5a is the spectra of signal laser at different output power. The 3 dB linewidth of signal laser at 3.96 kW was ~0.62 nm centering at ~1064.4 nm. The Raman stokes-shifted light around 1113 nm was observed at 3.006 kW, and its intensity gradually strengthened with the increase of the output power, but the signal-to-noise ratio (SNR) was as high as 49 dB compared with the spectral component around 1113 nm when the signal laser power was 3.96 kW. Therefore, SRS was effectively suppressed in our system. The evolution of the spectral linewidth of the signal laser through the power amplification was also described. As shown in Figure 5b, with the output power increase from 20 W (seed laser power) to 3.96 kW, the 3 dB linewidth of the signal laser changed from 0.48 nm to 0.62 nm, and the 10 dB linewidth of the signal laser varied from 1.08 nm to 1.40 nm. Overall, distinct spectral broadening did not occur during the power amplification.  Figure 6a,b show the temporal signals and corresponding PSD of the output laser at 3.878 kW and 3.96 kW. As shown in Figure 6a, when the output power was 3.878 kW, the temporal signal of the output laser was relatively stable, but the PSD exhibited slight noise-like protuberances from 0 to 10 kHz in Figure 6b. However, the frequency component became significantly enhanced when the output power increased to 3.96 kW. At the output power of 3.878 kW and 3.96 kW, the standard deviations of the corresponding temporal signals were 0.65% and 1.78%, respectively, which was increased nearly 3 times.  Figure 6a,b show the temporal signals and corresponding PSD of the output laser at 3.878 kW and 3.96 kW. As shown in Figure 6a, when the output power was 3.878 kW, the temporal signal of the output laser was relatively stable, but the PSD exhibited slight noise-like protuberances from 0 to 10 kHz in Figure 6b. However, the frequency component became significantly enhanced when the output power increased to 3.96 kW. At the output power of 3.878 kW and 3.96 kW, the standard deviations of the corresponding temporal signals were 0.65% and 1.78%, respectively, which was increased nearly 3 times. Therefore, the TMI effect can be considered to occur when the output power exceeds 3.878 kW.
Nanomaterials 2022, 12, x FOR PEER REVIEW 6 of 8 Therefore, the TMI effect can be considered to occur when the output power exceeds 3.878 kW.  Figure 7 shows the measured beam quality at the maximum output power. The M 2 factor of 3.96 kW is M 2 x = 1.31, M 2 y = 1.41. That means that the PM amplifier worked at near-diffraction-limited beam quality. Meanwhile, the measured M 2 factor of seed laser was M 2 x = 1.21, M 2 y = 1.33. Although the TMI was observed in the temporal signal of the output laser at 3.96 kW, there was no obvious degradation of beam quality compared with that of seed laser. The degradation of the beam quality usually lagged slightly behind that of temporal signal stability, which is common in the fiber amplifiers using gain fiber with a core/cladding diameter of 20/400 μm.  Figure 7 shows the measured beam quality at the maximum output power. The M 2 factor of 3.96 kW is M 2 x = 1.31, M 2 y = 1.41. That means that the PM amplifier worked at near-diffraction-limited beam quality. Meanwhile, the measured M 2 factor of seed laser was M 2 x = 1.21, M 2 y = 1.33. Although the TMI was observed in the temporal signal of the output laser at 3.96 kW, there was no obvious degradation of beam quality compared with that of seed laser. The degradation of the beam quality usually lagged slightly behind that of temporal signal stability, which is common in the fiber amplifiers using gain fiber with a core/cladding diameter of 20/400 µm. factor of 3.96 kW is M 2 x = 1.31, M 2 y = 1.41. That means that the PM amplifier worked at near-diffraction-limited beam quality. Meanwhile, the measured M 2 factor of seed laser was M 2 x = 1.21, M 2 y = 1.33. Although the TMI was observed in the temporal signal of the output laser at 3.96 kW, there was no obvious degradation of beam quality compared with that of seed laser. The degradation of the beam quality usually lagged slightly behind that of temporal signal stability, which is common in the fiber amplifiers using gain fiber with a core/cladding diameter of 20/400 μm.

Conclusions
In conclusion, we demonstrated a 3.96 kW all-fiberized narrow linewidth PM amplifier with 3 dB linewidth of 0.62 nm based on the MOPA structure. The slope efficiency of the main amplifier was ~79.5%. The PER was ~13.9 dB, and the M 2 factor was M 2 x = 1.31, M 2 y = 1.41 at the maximum output power. The backward signal power satisfied a linear growth, which revealed that there was no occurrence of SBS. SRS was observed during power scaling, but the SNR was up to 49 dB at the maximum power. Besides, the PSD of the signal laser showed obvious frequent components in the range of 0 to10 kHz at the maximum power, indicating that further power scaling is limited by TMI. New modulation signals to narrow the linewidth and TMI mitigation strategy will be the focus of future research.

Conclusions
In conclusion, we demonstrated a 3.96 kW all-fiberized narrow linewidth PM amplifier with 3 dB linewidth of 0.62 nm based on the MOPA structure. The slope efficiency of the main amplifier was~79.5%. The PER was~13.9 dB, and the M 2 factor was M 2 x = 1.31, M 2 y = 1.41 at the maximum output power. The backward signal power satisfied a linear growth, which revealed that there was no occurrence of SBS. SRS was observed during power scaling, but the SNR was up to 49 dB at the maximum power. Besides, the PSD of the signal laser showed obvious frequent components in the range of 0 to10 kHz at the maximum power, indicating that further power scaling is limited by TMI. New modulation signals to narrow the linewidth and TMI mitigation strategy will be the focus of future research.  Data Availability Statement: 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.

Conflicts of Interest:
The authors declare no conflict of interest.