The CARMENES radial velocity (RV) survey is observing 324 M dwarfs to search for any orbiting planets. In this paper, we present the survey sample by publishing one CARMENES spectrum for each M dwarf. These spectra cover the wavelength range 520-1710 nm at a resolution of at least R > 80, 000, and we measure its RV, Hα emission, and projected rotation velocity. We present an atlas of high-resolution M-dwarf spectra and compare the spectra to atmospheric models. To quantify the RV precision that can be achieved in low-mass stars over the CARMENES wavelength range, we analyze our empirical information on the RV precision from more than 6500 observations. We compare our high-resolution M-dwarf spectra to atmospheric models where we determine the spectroscopic RV information content, Q, and signal-to-noise ratio. We find that for all M-type dwarfs, the highest RV precision can be reached in the wavelength range 700-900 nm. Observations at longer wavelengths are equally precise only at the very latest spectral types (M8 and M9). We demonstrate that in this spectroscopic range, the large amount of absorption features compensates for the intrinsic faintness of an M7 star. To reach an RV precision of 1 m s −1 in very low mass M dwarfs at longer wavelengths likely requires the use of a 10 m class telescope. For spectral types M6 and earlier, the combination of a red visual and a near-infrared spectrograph is ideal to search for low-mass planets and to distinguish between planets and stellar variability. At a 4 m class telescope, an instrument like CARMENES has the potential to push the RV precision well below the typical jitter level of 3-4 m s −1 .

The theory of summation of electromagnetic line transitions is used to tabulate the Taylor expansion of the refractive index of humid air over the basic independent parameters (temperature, pressure, humidity, wavelength) in five separate infrared regions from the H to the Q band at a fixed percentage of Carbon Dioxide. These are least-squares fits to raw, highly resolved spectra for a set of temperatures from 10 to 25 • C, a set of pressures from 500 to 1023 hPa, and a set of relative humidities from 5 to 60%. These choices reflect the prospective application to characterize ambient air at mountain altitudes of astronomical telescopes. The paper provides easy access to predictions of the refractive index of humid air at conditions that are typical in atmospheric physics, in support of ray tracing [5] and astronomical applications [4,16,47,50] until experimental coverage of the infrared wavelengths might render these obsolete. The approach is in continuation of earlier work [46] based on a more recent HITRAN database [60] plus more precise accounting of various electromagnetic effects for the dielectric response of dilute gases, as described below.The literature of optical, chemical and atmospheric physics on the subject of the refractive index of moist air falls into several categories, sorted with respect to decreasing relevance (if relevance is measured by the closeness to experimental data and the degree of independence to the formalism employed here):1. experiments on moist air in the visible [3,9,18,53 [42,61,79] and eventually in the static limit [24], the refractive index plotted as a function of wavelength is more and more structured by individual lines. Since we will not present these functions at high resolution but smooth fits within several bands in the infrared, their spiky appearance sets a natural limit to the far-IR wavelength regions that our approach may cover. II. DIELECTRIC MODEL A. MethodologyThe complex valued dielectric function n(ω) of air n = 1 +χ (1) is constructed from molecular dynamical polarizabilitiesN m are molecular number densities, S ml are the line intensities for the transitions enumerated by l. ω 0ml are the transition angular frequencies, Γ ml the full linewidths at half maximum. c is the velocity of light in vacuum, and i the imaginary unit. The line shape (2) adheres to the complex-conjugate symmetry χ m (ω) = χ * m (−ω), as required for functions which are real-valued in the time domain. The sign convention of Γ ml merely reflects a sign choice in the Fourier Transforms and carries no real significance; a sign in the Kramers-Kronig formulas is bound to it. The integrated imaginary part is [28] whereν = k/(2π) = ω/(2πc) = 1/λ is the wavenumber.

The HITRAN 2000 database of infrared line transitions has been used to calculate the dispersion coefficient of water vapor at room temperature in the atmospheric window up to 25 microm, confirming an equivalent earlier compilation [Infrared Phys. 26, 371 (1986)]. I complement this line set by using an previously published ultraviolet pseudospectrum [J. Chem. Phys. 68, 1426 (1978)] to improve coverage of the near infrared. The effect of admixtures of abundant nitrogen, oxygen, and carbon dioxide is predicted on the same calculational basis to synthesize the air representative of the mountain that hosts the Very Large Telescope Interferometer and is found to be small compared with the dominant role of water at wavelengths above 3 microm.

Context. The main goal of the CARMENES survey is to find Earth-mass planets around nearby M-dwarf stars. Seven M dwarfs included in the CARMENES sample had been observed before with HIRES and HARPS and either were reported to have one short period planetary companion (GJ 15 A, GJ 176, GJ 436, GJ 536 and GJ 1148) or are multiple planetary systems (GJ 581 and GJ 876). Aims. We aim to report new precise optical radial velocity measurements for these planet hosts and test the overall capabilities of CARMENES. Methods. We combined our CARMENES precise Doppler measurements with those available from HIRES and HARPS and derived new orbital parameters for the systems. Bona-fide single planet systems were fitted with a Keplerian model. The multiple planet systems were analyzed using a self-consistent dynamical model and their best fit orbits were tested for long-term stability. Results. We confirm or provide supportive arguments for planets around all the investigated stars except for GJ 15 A, for which we find that the post-discovery HIRES data and our CARMENES data do not show a signal at 11.4 days. Although we cannot confirm the super-Earth planet GJ 15 Ab, we show evidence for a possible long-period (Pc = 7030-630+970 d) Saturn-mass (mcsini = 51.8-5.8+5.5M⊕) planet around GJ 15 A. In addition, based on our CARMENES and HIRES data we discover a second planet around GJ 1148, for which we estimate a period Pc = 532.6-2.5+4.1 days, eccentricity ec = 0.342-0.062+0.050 and minimum mass mcsini = 68.1-2.2+4.9M⊕. Conclusions. The CARMENES optical radial velocities have similar precision and overall scatter when compared to the Doppler measurements conducted with HARPS and HIRES. We conclude that CARMENES is an instrument that is up to the challenge of discovering rocky planets around low-mass stars.

No abstract