We analyse medium-resolution spectra (Rsim 18000) of 19 late type dwarfs in order to determine vsini-s using synthetic rather than observational template spectra. For this purpose observational data around 2.2 $mu$m of stars with spectral classes fro
m G8V to M9.5V were modelled. We find that the Na I (2.2062 and 2.2090 $mu$m) and $^{12}$CO 2-0 band features are modelled well enough to use for vsini determination without the need for a suitable observational template spectra. Within the limit of the resolution of our spectra, we use synthetic spectra templates to derive vsini values consistent with those derived in the optical regime using observed templates. We quantify the errors in our vsini determination due to incorrect choice of model parameters Teff, log $g$, $v_{rm micro}$, [Fe/H] or FWHM and show that they are typically less than 10 per cent. We note that the spectral resolution of our data(sim 16 km/s) limited this study to relatively fast rotators and that resolutions of 60000 will required to access most late-type dwarfs.
We present an analysis of high resolution spectra in the J band of five ultra cool dwarfs from M6 to L0. A new ab initio water vapour line list and existing line lists of FeH and CrH were used for spectra modelling. We find a good fit for the Mn I 12
899.76 A line. This feature is one of the few for which we have a reliable oscillator strength. Other atomic features are present but most of the observed features are FeH and water lines. While we are uncertain about the quality of many of the atomic line parameters, the FeH and CrH line lists predict a number of features which are not apparent in our observed spectra. We infer that the main limiting factor in our spectral analysis is the FeH and CrH molecular spectra.
