Metallicity determination of M dwarfs^⋆

Expanded parameter range in metallicity and effective temperature

Observational Astrophysics, Department of Physics and Astronomy, Uppsala University, Box 516, 751 20 Uppsala, Sweden
e-mail: sara.lindgren@physics.uu.se

Received: 1 March 2017
Accepted: 23 May 2017

Abstract

Context. Reliable metallicity values for late K and M dwarfs are important for studies of the chemical evolution of the Galaxy and advancement of planet formation theory in low-mass environments. Historically it has been challenging to determine the stellar parameters of low-mass stars because of their low surface temperature, which causes several molecules to form in the photospheric layers. In our work we use the fact that infrared high-resolution spectrographs have opened up a new window for investigating M dwarfs. This enables us to use similar methods as for warmer solar-like stars.

Aims. Metallicity determination with high-resolution spectra is more accurate than with low-resolution spectra, but it is rather time consuming. In this paper we expand our sample analyzed with this precise method both in metallicity and effective temperature to build a calibration sample for a future revised empirical calibration.

Methods. Because of the relatively few molecular lines in the J band, continuum rectification is possible for high-resolution spectra, allowing the stellar parameters to be determined with greater accuracy than with optical spectra. We obtained high-resolution spectra with the CRIRES spectrograph at the Very Large Telescope (VLT). The metallicity was determined using synthetic spectral fitting of several atomic species. For M dwarfs that are cooler than 3575 K, the line strengths of FeH lines were used to determine the effective temperatures, while for warmer stars a photometric calibration was used.

Results. We analyzed 16 targets with a range of effective temperature from 3350−4550 K. The resulting metallicities lie between −0.5< [M/H] < +0.4. A few targets have previously been analyzed using low-resolution spectra and we find a rather good agreement with our values. A comparison with available photometric calibrations shows varying agreement and the spread within all empirical calibrations is large.

Conclusions. Including the targets from our previous paper, we analyzed 28 M dwarfs with high-resolution infrared spectra. The targets spread approximately one dex in metallicity and 1400 K in effective temperature. For individual M dwarfs we achieve uncertainties of 0.05 dex and 100 K on average.