Towards a general illumination framework for extreme-precision radial velocities

Abstract

Extreme-precision radial-velocity (EPRV) spectrographs now operate near the ∼10 cm s−1 level, and the next generation of instruments targets < 10 cm s−1 in order to detect terrestrial planets in the habitable zones of Sun-like stars. At this level, once vacuum operation, thermal control, and detector characterization have suppressed the classical instrumental terms, one of the least standardized and least portable contributions that remains is the physics of how light is delivered to the spectrograph: coupling, propagation through optical fibers, focal-ratio degradation (FRD), scrambling, modal noise, and the coupling of the calibration sources. We refer to this collectively as illumination physics. A survey of the literature shows that, although a great deal is known about each of these effects individually, the field still lacks a transferable framework that predicts their combined impact on measured radial velocity (RV); as a result, successive instruments (among them CARMENES, ESPRESSO, EXPRES, HARPS, HPF, KPF, MAROON-X, NEID, NIRPS, SOPHIE, and SPIRou, with others not treated in detail here) have rediscovered the same illumination sensitivities one at a time. We identify three structural gaps in the present literature: scrambling is not characterized in a way that maps onto RV; FRD is physically well understood but operationally non-portable between instruments; and the dependence of fiber behavior on the temporal coherence (linewidth) of the light source has no unified predictive model. We then describe a coordinated program designed to close these gaps, comprising controlled fiber-launch experiments, structured FRD measurements aimed at transfer functions, a Fourier-transform-spectrograph (FTS) testbed that pairs nearfield/ far-field (NF/FF) imaging with line-shape measurement, dedicated Fabry-Pérot experiments, and hollow-cathode- lamp coupling studies, with the goal of producing community-wide illumination metrics, sensitivity functions, and design rules. This contribution presents the framework and its current status; full experimental results will follow.