Addressing modulational instability in anti-resonant hollow-core fibers for pulse compression

Abstract

When pulses propagate in gas-filled anti-resonant hollow-core fibers (AR-HCFs) modulational instability (MI) can lead to pulse break-up and loss of coherence. In pulse broadening and compression schemes, MI is a parasitic effect that induces significant shot-to-shot fluctuations of the peak power of compressed pulses and increases rapidly over a narrow range of input pulse energies. In this work we use experimental studies and supporting numerical simulations to compare two AR-HCFs that are chosen to enhance or suppress MI. We demonstrate that judicious selection of the wall thickness of the anti-resonant elements (AREs) can drastically reduce the MI gain, thereby increasing the limit of pulse energy scaling of stable ultrafast pulse compression.

Figures (5)

• Figure 1: a) Experimental setup. The output of a Yb-doped fiber laser oscillator/amplifier system is focused with a lens (L) into an anti-resonant hollow-core fiber filled with $\approx 20$ bar argon. The output is then recollimated and compressed with chirped mirrors (CM). A half wave plate (HWP) and a pair of thin film polarizers (TFP) act as power control before two diagnostics: second harmonic generation (SHG) FROG and an optical spectrum analyzer (OSA). b), c) Fiber cross section images of Fiber A and Fiber B, respectively, where $T$ is the anti-resonant element (ARE) wall thickness.
• Figure 2: Pulse energy scaling of spectral broadening in each AR-HCF. a) Fiber A with a capillary wall thickness of 300 nm showing clear evidence of MI sidebands near 900 and 1200 nm. The inset shows a linear fit to the peak values (dB scale) of the MI sideband near 1200 nm, plotted versus the input pulse energy (linear scale), indicating that the gain is exponential. b) Fiber B with a capillary wall thickness of 820 nm displaying no MI sidebands.
• Figure 3: Experimental single shot spectra of Fiber A using dispersive Fourier transform (DFT) spectroscopy. Individual spectra are shown in gray while the ensemble average is shown in black. An independently measured average spectrum from an optical spectrum analyzer (OSA) is shown in red. The wavelength axis is calibrated in two separate regions denoted by the black vertical dashed line (see text for details).The region from 1120-1300 nm is multiplied by 200 to show detail of the sideband.
• Figure 4: Pulse propagation simulations with experimental measurements for T=300 nm ( Fiber A, left column) and T=820 nm ( Fiber B, right column) wall thickness AR-HCFs. One thousand simulations were conducted for each fiber and are shown in gray. The average of each set of simulations is shown in black. Experimental results (when available) are shown in coloured dashed lines. a) and b) output spectra; c) and d) output temporal intensity; e) and f) temporal intensity after numerical compression along with experimental FROG retrievals. Insets show zoomed-in views of the peak fluctuations.
• Figure 5: a) Standard deviation of the peak power distribution for 1000 simulations as a function of the product of input pulse energy and argon pressure in Fiber A. Insets show the peak power ($P_p$) distributions normalized to pressure ($P$) for the points framed by black rectangles. b) Simulated normalized average spectra corresponding to the rightmost points in a). Also shown in blue is a spectrum for Fiber B under comparable broadening conditions with corresponding peak power standard deviation represented by the blue square in a).