Promising single-frequency laser for next-generation gravitational wave detectors

The immediate goal for the next generation of gravitational wave detectors is the development of reliable single-frequency laser sources. It is necessary for these laser sources to have a wavelength of 1μm, an output power of several hundred watts, and a spectral linewidth in the kilohertz (kHz) range. The main challenge to this development, however, is the low output power of existing single-frequency laser sources. At present, these sources have output powers that are limited to the 1W region because of thermal and mechanical problems.¹

During the early phase of fiber technology development, injection-locked solid-state oscillators (ILOs) were found to be a promising solution (in terms of power scalability and reliability) for overcoming the problem of low output power single-frequency laser sources.² Based on this technology, a high-power solid-state ILO has been developed at Laser Zentrum Hannover e.V.¹ in collaboration with the Albert Einstein Institute in Hannover, Germany. This ILO, which can be operated 24 hours a day and seven days a week with an output power of 160W and a wavelength of 1064nm, is currently being used for gravitational wave detection at the advanced Laser Interferometer Gravitational-Wave Observatories.³

The continuous development of active double-clad (DC) fibers and high-brightness diode lasers has also paved the way for high-power single-frequency fiber-based master oscillator power amplifiers (MOPAs). Output power levels in the range of 500W, at a spectral linewidth of ≤1MHz,^4–6 have thus been achieved in laboratory setups. In these cases, the power level is considered as below the threshold of stimulated Brillouin scattering (SBS).⁷ To minimize Brillouin scattering levels, large mode area (LMA) fibers with mode field areas (MFAs) of up to about 700μm² have been used. The use of this new generation of LMA fibers (with complex glass structures), however, often involves beam quality issues (e.g., transversal mode instabilities). Furthermore, the lack of suitable fiber components inhibits progress in the development of rugged monolithic DC fiber amplifiers. All these drawbacks (including SBS) thus hamper the development of reliable, continuously operating, fiber-based single-frequency laser systems with output powers of several hundred watts. Current commercially available laser systems (with spectral laser linewidths below 10MHz) can only be used to deliver output power levels in the range of 50W. In addition, the highest output power for monolith single-frequency DC fiber amplifiers that has yet been achieved in a laboratory is 200–300W.^{8, 9}

For the first time in this context, we therefore propose a simple and robust core-pumped ytterbium-doped fiber amplifier to address the above-mentioned problems. With this device we can achieve high-power amplification of single-frequency laser sources at a wavelength of 1μm. For our high-power amplification approach we only require a commercial off-the-shelf single-mode fiber (MFA of about 100μm²) to obtain a single-frequency output power of several hundred watts, with a kHz linewidth.

The experimental setup for our core-pumped MOPA is depicted in Figure 1. The setup consists of a seed laser, a nonplanar ring oscillator (NPRO) with an output power of 2W, as well as a pre- and main amplifier. The spectral linewidth of the NPRO is specified as 1kHz. We use an ytterbium-doped polarization-maintaining single-mode fiber, with a core diameter of 10μm (numerical aperture of 0.075), in the main amplifier stage. The main amplifier is pumped by a single-mode fiber laser (developed in-house) that emits at an output power of up to 200W and at a wavelength of 1018nm.¹⁰

Figure 1. Experimental setup of the core-pumped master oscillator fiber amplifier (MOPA). The nonplanar ring oscillator (NPRO) has an output power of 2W and a spectral linewidth (Δf) of 1kHz, and a wavelength (λ) of 1064nm. λ/2: Half-wave plate. FPI: Fabry-Pérot interferometer. FSS: Fused silica sheet. PA: Pre-amplifier. WDM: Wavelength-division multiplexer. Yb-SM-fiber: Ytterbium-doped single-mode fiber.

With the use of this experimental setup we have thus characterized the output power properties of our core-pumped MOPA. Our results (see Figure 2) indicate that the maximum output power (158W) occurs at an absorbed pump power of 171W. Furthermore, the linear slope of the backward propagating power (up to the maximum output power of 158W), as shown in Figure 2, confirms the SBS-free operation⁷ of our amplifier. In our experiments, the output power was thus only limited by the available pump power and not by SBS.

Figure 2. Experimentally measured output power characteristics (blue squares) and relative backward power (black circles) of the core-pumped MOPA, plotted as functions of absorbed pump power.

To gain a more detailed understanding of the Brillouin scatter processing, we used a Fabry-Pérot interferometer (FPI) to measure the evolution of the Brillouin gain spectra (BGS) with respect to the amplifier output power. The relevant spectral range (i.e., within one free spectral range) contains the BGS at an output power level of 67 and 105W (see Figure 3). The full width at half-maximum of the two BGS is approximately 40 and 100MHz, respectively (one-tenth of the FPI's free spectral range corresponds to 200MHz). The strong broadening of the BGS that we observe with increasing amplifier output power is caused by the large temperature gradient along the short active fiber. This temperature-induced broadening of the BGS—in addition to the short active fiber length—is key for achieving efficient SBS mitigation.

Figure 3. Brillouin gain spectra of the high-power fiber amplifier at three output power levels. RT: Center frequency of the Brillouin gain spectrum at room temperature (estimated). a.u.: Arbitrary units.

In summary, we have developed a new core-pumped ytterbium-doped fiber amplifier to achieve high-power amplification of single-frequency laser sources at a wavelength of 1μm. Our monolithic amplifier setup is simple, cost-effective, and rugged. It is thus a promising approach for reliable continuous operation at output power levels of several hundred watts and is suitable for use in the next generation of gravitational wave detectors. In the next stages of our research activities we will focus on the development of further power scaling and fiber integration.