Mitigation of Polarization Effects in Single-mode Fiber Spectrographs

• Published in: Publications of the Astronomical Society of the Pacific Vol. 132; no. 1015; pp. 1 - 12
• Main Authors: Bechter, Andrew J., Bechter, Eric B., Crepp, Justin R., Ketterer, Ryan, Crass, Jonathan
• Format: Journal Article
• Language: English
• Published: Philadelphia The Astronomical Society of the Pacific 01.09.2020; IOP Publishing Limited; IOP Publishing
• Subjects: Astronomical Instrumentation, Telescopes, Observatories, and Site Characterization; Diffraction; Doppler effect; Light sources; Mathematical models; Noise; Numerical models; Optics; Polarimetry; Polarization; Radial velocity; Refractive index; Spectroscopy
• DOI: 10.1088/1538-3873/ab9cc6
• ISSN: 0004-6280; 1538-3873

The use of single-mode fibers (SMFs) to illuminate radial velocity (RV) spectrographs shows promise to achieve extremely precise Doppler measurements. Due to their small core diameter, SMFs only propagate a single spatial mode which allows for diffraction-limited optical performance while simultaneously eliminating fiber modal noise. The single spatial mode however consists of two orthogonal polarization modes. In circular core fiber with a non-isotropic refractive index profile or asymmetries in the cross-sectional geometry, the two polarization modes propagate with different relative speeds inducing birefringence. Conditions at a telescope observatory will subject the fiber to mechanical (bending and twisting) and thermal stresses, inducing birefringence that varies in time. The interaction of variable birefringence combined with with polarization sensitive optics, such as diffraction gratings, results in an intensity modulation that causes unwanted Doppler shifts via "polarization noise." In this paper, we characterize variable fiber birefringence both in the laboratory and at the Large Binocular Telescope using a Stokes parameters. We then combine the measured Stokes vector through a numerical model of a SMF spectrograph to understand the impact of variable polarization on RV precision. We find that polarization noise is a tens of cm s−1 to several m s−1 effect, which is exacerbated by the degree of polarization of the light source and the polarization response of the spectrograph optics. Finally we show experimentally mitigating the RV offset using polarization averaging methods and in-line fiber depolarizers and can reduce a several m s−1 polarization noise to ≤10 cm s−1.