Welcome to Smart Optical Fiber System Lab

News Center

Location:Home > News Center >Spectrum broade...

Spectrum broadening and regulation of intelligent seed source——exploring spectrum optimization to improve the power limit of high-power lasers

06
03
2023

Research Background


The Master Oscillator Power Amplifier (MOPA) structure is easy to adjust and has significant power boost performance, and is considered to be the preferred structure for high-power fiber lasers. Ideally, the laser output power has a nearly linear relationship with the pump power or gain coefficient. However, when the output power continues to increase, it is affected by various nonlinear effects in the fiber, such as cross-phase modulation, stimulated Raman scattering effect, Due to the stimulated Brillouin scattering (Stimulated Brillouin Scattering, SBS) effect and four-wave mixing, the output power curve enters the nonlinear region and presents a gain compression phenomenon. Among them, the threshold of the SBS effect is the lowest, which becomes the bottleneck that limits the maximum output power of the laser.

Considering the complexity of the manufacturing process, the degree of increase in the threshold of the SBS effect, and operability, the scheme of using phase modulation to widen the bandwidth of the seed source has the characteristics of simple operation, high controllability, and high power boosting efficiency within the effective bandwidth. Research hotspots of SBS effect in high-power fiber lasers. However, the controllability of the spectrum after the current broadening is poor, the spectrum is not ideal, and the randomness of the signal causes the self-pulsation phenomenon, which limits the maximum output power.


Research content


In view of the above problems, this research is dedicated to the ideal spectrum design of the seed source and intelligent spectrum flexible control, and explores the output power limit of high-power fiber lasers obtained through spectrum optimization. The key research contents include:

1. Establishment of arbitrary seed source broadening spectrum and laser power threshold model

The physical model between the wide-spectrum seed source of arbitrary spectrum type and the laser power threshold mainly includes three parts: the high-order phase modulation model of the wide-spectrum drive signal and the seed source spectrum, the multi-level power amplification evolution model of the seed source after broadening, and the laser nonlinearity Effect Threshold Model. The research idea is shown in Figure 1. Starting from the multi-frequency drive spectrum broadening seed source, the relationship between the broadening bandwidth, spectral type and other parameters and the system SBS threshold is grasped through simulation, and the laser spectrum and return light are analyzed in time domain to improve the spectrum. shape, further optimize the seed source phase modulation technology, and provide a theoretical basis for reducing the sensitivity of the nonlinear effect of the system and increasing the maximum output power.


Figure 1   Research scheme of seed source broadening spectrum and laser power threshold model

2. Establishment of self-defined signal design for seed source spectrum regulation and laser power testing

In this study, the high-precision signal controllability of the arbitrary waveform generator is used to control the spectrum shape and in-band power distribution with high precision, and obtain a high-order modulation spectrum with a near-rectangular distribution, and then study the realization method of the arbitrary spectrum shape. Compare the degree of output power improvement. Taking the spectrum of the rectangular seed source as an example, starting from the time domain and frequency domain respectively, the P-tuning sequence and the multi-frequency driving signal are designed, and the P-tuning and multi-frequency driving signals obtained by simulation are shown in Figure 2. The P adjustment sequence has the advantages of simple implementation and convenient adjustment. By changing the signal rate, P value and amplitude, the adjustment of the widening bandwidth and flatness can be directly realized. Multi-frequency signals have more dimensions and higher-precision spectrum adjustment space, which can be extended to signal design of any spectrum.


Figure 2 (a) Simulation diagram of spectrum envelope of P-tuned sequence


Figure 2  (b)  Multi-frequency driving signal spectrum simulation diagram

Using the MOPA structure laser amplification test platform as shown in Figure 3(a), firstly, the P-adjustment sequence is used as the driving electrical signal to obtain relatively flat 10GHz and 20GHz bandwidth broadening spectra, and the laser output power and reverse power curves are obtained experimentally As shown in Figure 3(b), compared with the cascaded white noise modulation broadening spectrum under the same bandwidth, the output power of the laser has an increase of 300W. Subsequently, using multi-frequency driving electrical signals, when the broadening bandwidth is 10 GHz, the laser output power and reverse power curves at different frequency intervals are shown in Figure 3(c). When the frequency interval is 500kHz, the highest output power is 2234W. Compared with the cascaded white noise scheme under the same bandwidth, the power threshold is increased by nearly 500W. This experiment fully verified that the optimization of the spectral shape and frequency interval of the driving signal can improve the power threshold of the laser, laying the foundation for further optimization of the spectrum.


Figure 3 (a) MOPA structure laser amplification system.


Figure 3 (b)   P-adjustment sequence drive, laser output power and reverse power curve


Figure 3 (c) Multi-frequency drive, laser output power and reverse power curve

3. Spectral precise measurement and deep reinforcement learning algorithm to optimize the spectrum

To establish an intelligent optimization model for broadening the spectrum of high-order phase modulation seed sources, first of all, accurate spectral measurement information is required as the input of the deep reinforcement learning algorithm to generate and feed back the training sequence until the modulation spectrum corresponding to the best output spectrum is obtained. As shown in Figure 4. At present, the highest resolution of a 1-micron spectrometer is only 0.02nm. In this study, the frequency-sweeping heterodyne beat frequency power acquisition method is used to improve the spectral measurement accuracy from 1.5GHz to about 50MHz. A commercial spectrometer (Yokogawa, AQ6370D) was used to measure the same spectrum as this scheme, and the comparison chart of the obtained measured spectrum is shown in Figure 4(a). The measurement spectrum of a commercial spectrometer is a broadband flat-top envelope, which cannot observe in-band information. However, this scheme can more accurately capture the sag and jitter information in the band, and measure the bandwidth more accurately. Establishment provides more accurate spectral information.


Figure 4 (a) Comparison of measured spectra between commercial spectrometers and this solution
(b) Based on depth Spectral Optimization Models for Reinforcement Learning Algorithms. (c) Deep reinforcement learning algorithm to optimize the spectrum

In summary, this study proposes a seed source spectral broadening method based on intelligent spectral control and optimization, and establishes arbitrary signal high-order phase management, seed source spectrum, and laser power threshold models. Through custom signal design and single-stage high-order Phase modulation, combined with high-precision spectral measurement based on swept-frequency heterodyne beat frequency power detection and the spectral type optimization model of deep reinforcement learning algorithm, can realize the controllable seed source of 5-30GHz spectral shape, and improve the SBS effect threshold of the amplification link , to achieve intelligent and flexible regulation for different laser systems and parameters.

Laboratory of Intelligent Fiber Ecosystem (LIFE, Laboratory of Intelligent Fiber Ecosystem), affiliated to the State Key Laboratory of Regional Optical Fiber Communication Networks and New Optical Communication Systems of Shanghai Jiaotong University, is dedicated to the research of intelligent fiber laser systems, aiming to solve the problem of ultrafast pulsed fiber lasers. The problem of precise pulse control in , and the problem of precise control of spectrum in high-power CW fiber lasers.

Relevant research was supported by the National Natural Science Foundation of China Youth Fund Project. For detailed principles and results analysis, please refer to the original text: https://164968.saaas.com/goods/942781