The stars of the Galactic halo have [Fe/H] abundances mostly less than -1.0. Their kinematics are very different from the rotating thick and thin disks: the mean rotation of the stellar halo is close to zero, and it is supported against gravity primarily by its velocity dispersion. It is now widely believed that much of the stellar halo comes from the debris of small accreted satellites (Searle & Zinn 1978). There remains a possibility that a component of the halo formed dissipationally during the Galaxy formation process (Eggen et al. 1962, Samland & Gerhard 2003). Halo-building accretion events continue to the present time: the disrupting Sgr dwarf is an example in our Galaxy, and the faint disrupting system around NGC 5907 is another example of such an event (Martínez-Delgado et al. 2010). The metallicity distribution function (MDF) of the major surviving satellites around the Milky way is not like the MDF in the stellar halo (e.g. Venn & Hill 2008) but the satellite MDFs may have been more similar long ago. We note that the fainter satellites are more metal-poor and are consistent with the Milky Way halo in their [α/Fe] behaviour.
Is there a halo component that formed dissipationally early in the Galactic formation process? Hartwick (1987) showed that the metal-poor RR Lyrae stars delineate a two-component halo, with a flattened inner component and a spherical outer component. Carollo et al. (2010) identified a two-component halo and the thick disk in a sample of 17,000 SDSS stars, mostly with [Fe/H] < -0.5. They described the kinematics well with these three components:
Thick disk: (V, σ, [Fe/H]) = (182, 51, -0.7)
Inner halo: (V, σ, [Fe/H]) = (7, 95, -1.6)
Outer halo: (V, σ, [Fe/H]) = (-80, 180, -2.2)
Here [Fe/H] is the mean abundance for the component, V and σ are its mean rotation velocity relative to a non-rotating frame, and velocity dispersion, in km s-1. The outer halo appears to have retrograde mean rotation. As we look at subsamples at greater distances from the Galactic plane, we see that the thick disk dies away and the retrograde outer halo takes over from the inner halo. With the above kinematic parameters, the equilibrium of the inner halo is a bit hard to understand. It may not yet be in equilibrium. From comparison with simulations, Zolotov et al. (2009) argue that the inner halo has a partly dissipational origin, while the outer halo is made up from debris of faint metal-poor accreted satellites.
Recently Nissen & Schuster (2010) studied a sample of 78 halo stars with [Fe/H] > -1.6 and find that they show a variety of [α/Fe] enhancement. Their sample shows high and low [α/Fe] groups, and the low [α/Fe] stars are mostly in high energy retrograde orbits. The high [α/Fe] stars could be ancient halo stars born in situ and possible heated by satellite encounters. The low-alpha stars may be accreted from dwarf galaxies.
How much of the halo comes from accreted structures? An ACS study by Ibata et al. (2009) of the halo of NGC 891 (a nearby edge-on galaxy like the Milky Way) shows a spatially lumpy metallicity distribution, indicating that its halo is made up largely of accreted structures which have not yet mixed away. This is consistent with simulations of stellar halos by Font et al. (2008), 2Gilbert et al. (2009) and Cooper et al. (2010).
To summarize this section on the Galactic stellar halo: the stellar halo is probably made up mainly of the debris of small accreted galaxies, although there may be an inner component which formed dissipatively.