3 Difference frequency generation of high-power mid-infrared ultrafast light sources  3.3 Differential frequency generation of 2-5 μm high-power tunable mid-infrared ultrashort pulses To obtain a 2-5 μm high-power tunable mid-infrared ultrashort pulse, a 1.55 μm high-energy ultrashort pulse is used to broaden the spectrum to 1.3-1.9 μm via a SESS (Sequencing Energy Sequencing System) and replaces the signal light in Figure 7(a) in the difference frequency generation system. All components of the difference frequency system are identical to those in Figure 7(a). Adjusting the input pulse energy of the SESS allows the signal light spectrum to be tuned from 1.3-1.9 μm, and laterally moving the PPLN crystal (www.wisoptic.com) matches the polarization period with the wavelengths of the pump and signal lights. Figure 11 shows the spectrum and power corresponding to a pump power of 15 W. The highest power is located at an idler wavelength of 3.28 μm (corresponding to a signal wavelength of 1.5 μm), with an average output power of 1.87 W and a single pulse energy of 56 nJ. As the idler wavelength increases, its average output power decreases, with an average power of 1.02 W at a center wavelength of 4.8 μm. Obvious carbon dioxide and water absorption peaks are observed in the spectra at wavelengths of 4.2 μm and 2.7 μm. The spectral peaks at 2.7 μm and 2.25 μm are the result of the difference frequency generation of the pump light and the signal light at 1.65 μm and 1.9 μm, respectively. Due to the lack of a bandpass filter to separate these two signal light components, the idler frequencies of both bands are simultaneously output to the power meter. Adjusting the PPLN (www.wisoptic.com) polarization period to achieve quasi-phase matching in one band can suppress the generation efficiency of the other spectral component, but the effect is limited. The final output power is 1.1 W, which includes the total power of both spectral peaks. The pump light pulse power was increased to 30 W, and frequency difference was used with signal pulses with center wavelengths of 1.35 μm, 1.4 μm, 1.45 μm, 1.55 μm, and 1.6 μm to obtain idler light with wavelengths of 4.2 μm, 3.9 μm, 3.58 μm, 3.06 μm, and 2.9 μm, respectively, with powers of 1.98 W, 2.48 W, 2.73 W, 2.58 W, and 3.02 W. When the signal wavelength was tuned to 1.3 μm, the SESS generated a wider spectral sidelobe bandwidth. Using the same 50 nm bandpass filter, the frequency bandwidth corresponding to the shorter center wavelength of 1.3 μm was even wider, resulting in a narrower pulse width. After amplification by the DFG process, the peak power was higher, causing the pulse to self-focus in the crystal, ultimately damaging the crystal. fig.11. output spectrum and power of the 2 5μm high power tunable short wave mid infrared Fig.11. Output spectrum and power of the 2-5μm high power tunable short-wave mid-infrared laser source.

3 Difference frequency generation of high-power mid-infrared ultrafast light sources  3.2 Difference frequency generation for producing high-power mid-infrared ultrashort pulses at 3μm The time delay line was adjusted to synchronize the pump pulse and signal pulse in time, and the variation of mid-infrared output power with pump power and signal power was measured respectively. The results are shown in Figure 9.   Fig.9. In the experimental output of the idle light energy changes with the pump light and signal light energy after optimized delay During the measurement, the delay line was optimized to ensure maximum output mid-infrared power. In Figure 9(a), when the signal energy is 0.3 nJ (black squares), the idler energy initially increases exponentially with the pump energy, reaching saturation after the pump energy exceeds 600 nJ. This trend is consistent with the trend in Figure 3(a) where the signal energy is 0.1 nJ and the pump energy varies between 100 and 3.5 μJ. As the signal energy increases to 9 nJ (green triangles), the nonlinear interaction significantly enhances, and the pump energy required to reach saturation decreases to 240 nJ. When the signal energy is 120 nJ (black hexagonal star), the idler energy enters the saturation region after the pump energy exceeds 120 nJ, consistent with the curve corresponding to the signal pulse energy of 100 nJ in Figure 3(a). Due to the delay optimization, the idler energy curve converges to its maximum value in an orderly manner, consistent with the trend of the curve in Figure 3(a). Judging from the idler energy curve, the DFG operates in the saturation region when the pump energy is 900 nJ and the signal energy is 120 nJ, consistent with the corresponding energies in Figure 3(a). Figure 9(b) shows the idler energy versus signal energy curve for different pump energies (60-900 nJ). Compared to the growth pattern of the curve in Figure 3(b), the pulse energy corresponding to the parameters in Figure 9 is higher, resulting in no region where the idler energy increases linearly with the signal energy. The idler growth rate gradually slows down at all pump energies. When the pump energy is only 60 nJ (black squares), no obvious saturation is observed. When the pump energy is 540 nJ (green triangles), the signal energy required to reach the saturation region is 45 nJ. When the pump energy is 900 nJ, the signal energy required to reach the saturation region is 9 nJ, consistent with the simulation pattern in Figure 3(b). When the pump light energy is 900 nJ and the signal light energy is 120 nJ, the output mid-infrared idler light energy is 92 nJ, with a repetition rate of 33.3 MHz, corresponding to an average power of 3.06 W. The idler light spectrum is shown in Figure 10. The spectrum has a central wavelength of 3.06 μm and a full-width at half-maximum of approximately 70 nm. The red dashed line is the mid-infrared spectrum calculated using the output parameters of the dual-wavelength laser. The experimental and simulated spectrum widths are nearly identical, with a slight shift in the central wavelength due to a slight difference in the effective polarization period of the PPLN (www.wisoptic.com) crystal in the experiment and the simulation. Fig. 10. The final output of the mid-infrared spectrum