Lithium abundance and surface magnetic fields: new constraints in magnetic models of M dwarfs

Precision modeling of M dwarfs has become worthwhile in recent years due to the increasingly precise values of masses and radii which can be obtained from eclipsing binary studies. Torres (2013) has identified 4 prime M dwarf pairs with the most precise empirical determinations of masses and radii. The measured radii are consistently larger than standard stellar models predict. We have previously modeled M dwarfs in the context of a criterion due to Gough & Tayler in which magnetic fields inhibit the onset of convection according to a physics-based prescription. New constraints on the models of M dwarfs are now provided by measurements of lithium abundances. The key aspect of Li in terms of setting constraints on magnetic modeling is that Li burning starts at T = 2.5 MK, and temperatures of just such magnitude are associated with the base of the convection zone: magnetic inhibition of convective onset can shift this base slightly closer to the surface, i.e. to slightly lower temperatures, thereby reducing the amount of Li depletion compared to a non-magnetic model. In the present paper, we consider how our magneto-convection models handle the new test of stellar structure provided by Li measurements. Among the prime systems listed by Torres, we find that plausible magnetic models work well for CM Dra and YY Gem but not for CU Cnc. (The fourth system in Torres's list does not yet have enough information to warrant magnetic modeling.) For CU Cnc, we suggest that the observed lithium may have been accreted from a circumstellar disk. We find that our magneto-convection models of CM Dra, YY Gem and CU Cnc yield results which are consistent with the observed correlation between magnetic flux and X-ray luminosity.