Investigation of Noise Sources Down to the Shot-Noise Limit in Yb-Doped Fiber Amplifers for TMI Investigations

Summary form only given. Ytterbium-doped fiber laser amplifiers are known for their high single-pass gain and average powers up to the kilowatt range, while maintaining single-mode output. An upper limitation for the achievable output power is given by transverse mode instabilities (TMI). Due to the interference with higher order modes that create a thermally induced long period grating, chaotic power transfer happens above a certain threshold, which degrades the beam quality. It has been shown that this effect can be influenced by manipulating the grating strength itself. It has also been shown that a certain phase relation of the grating to the guided modes is necessary to efficiently transfer power or hinder the effect by phase disturbance [1]. A natural phase disturbance is given by noise, which is introduced by the pump or seed source in a fiber amplifer. In this contribution, we experimentally investigate the amount of intensity noise in a fiber-pre-amplifer, which is typically used in kW experiments and relate it to the fundamental shot noise limit (SNL). In the experiments, we used a single-frequency external cavity diode laser as a low noise source and amplify it by 30 dB from 10 mW to an output power of 10 W. This is a typical pre-amplifier configuration, before such sources can be amplified to the kW level. The fiber amplifier is a double-clad fiber with a 10/125 μm geometry, pumped by a wavelength stabilized pump diode at 976 nm via a monolithic pump coupler. In order to prevent stimulated Brillouin scattering, the seed laser is phase modulated by a combination of a sinusoid and white noise and thus broadened to 50 GHz linewidth. The noise measurements are done by balanced self-homodyne detection including optimized photodiode readout circuits spanning different frequency ranges. The recorded spectra are given in Fig. 1. The SNL was verified through an attenuation measurement for the subtracted detector signals. Finally, the measured excess noise values are fitted by their corresponding second-order polynomial power dependency and extrapolated to the full power of the amplifier. The seed source, phase modulator and fiber amplifier have been characterized in this setup separately to investigate their contributions. For the fiber amplifier, both co- and counter-pumping configuration, are analysed. Due to detector limitations, all given spectra are measured at an optical power in the mW regime corresponding to approx. 30 dB of attenuation in the amplifiers.