The First High-Resolution Spectra of 1.3 L Subdwarfs

We present the first high-resolution (R ~ 31,000) spectra of the cool sdL 2MASS0532, and what was originally identified as an early-type L subdwarf (sdL) LSR1610-0040. Our work, in combination with contemporaneous work by Cushing and Vacca, makes it clear that the latter object is more probably a mid-M dwarf with an unusual composition that gives it some sub-dwarf spectral features. We use the data to derive precise radial velocities for both objects and to estimate space motion; both are consistent with halo kinematics. We measure the projected rotational velocities, revealing very slow rotation for the old sd?M6 object \lsr. \twom exhibits rapid rotation of vsini = 65 +- 15km/s, consistent with the behavior of L dwarfs. This means that the braking time for L dwarfs is extremely long, or that perhaps they never slow down. A detailed comparison of the atomic Rb and Cs lines to spectra of field L dwarfs shows the spectral type \twom is consistent with being mid- to late-L. The Rb I and K I lines of \lsr\ are like an early-L dwarf, but the Cs I line is like a mid-M dwarf. The appearance of the Ca II triplet in absorption in this object is very hard to understand if it is not as least as warm as M6. We explain these effects in a consistent way using a mildly metal-poor mid-M model. M subdwarfs have weak metal-oxides and enhanced metal-hydrides relative to normal M dwarfs. \lsr\ exhibits metal-hydrides like an M dwarf but metal-oxides like a subdwarf. The same explanation that resolves the atomic line discrepancy explains this as well. We identify atomic lines of Ti around 9600 \AA and a small contribution of FeH, but we cannot confirm a detection of TiH in the spectra of cool L subdwarfs. High resolution spectroscopy has aided in beginning to understand the complex molecular chemistry in metal-deficient and ultralow-mass objects.