Oi

s tn

0.01

0.001

0.0001

0.01 0.1 1 10 100 1000 io4 106

Figure 3. Companions mass as a function of separation for solar-type stars within 25 parsecs of the Sun. Crosses mark stellar and brown dwarf secondaries, while the solid squares identify planetary-mass companions. The dotted lines mark the effective limits of the planetary radial velocity surveys. The scarcity of systems with masses between 0.01 and 0.1 Mq is the brown dwarf desert.

There is no obvious direct correlation between the mass of the primary star and the masses of planetary companions; for example, the ^0.35 M0 M3 dwarf, Gl 876, has two planets with masses comparable to Jupiter. On the other hand, one might expect an upper limit to the mass distribution to emerge, simply because lower mass stars are likely to have lower mass protoplanetary disks.

3.2. Metallicities and the thick disk

Chemical abundance, particularly individual elemental abundance ratios, serves as a population discriminant for the ESP host stars. Halo stars have long been known to possess a-element abundances (Mg, Ti, O, Ca, Si) that are enhanced by a factor of 2-3 compared to the Sun. This is generally attributed to the short formation timescale of the halo (Matteucci & Greggio 1983): a-elements are produced by rapid a-capture, and originate in Type II supernovae, massive stars with evolutionary lifetimes of 107 to 108 years. In contrast, Type I supernovae, which are produced by thermal runaway on an accreting white dwarf in a binary system, have evolutionary timescales of 1-2 Gyrs. These systems produce a much higher proportion of Fe; thus, their ejecta drive down the [a/Fe] ratio in the ISM, and in newly forming stars.

Recent high-resolution spectroscopic analyses of nearby high velocity stars provide evidence that the thick disk is also a-enhanced (Fuhrmann 1998, 2004; Prochaska 2000).

Figure 4. a-element abundances as a function of metallicity. The lower panel plots data for nearby stars from Fuhrmann (1998), where the open circles are disk dwarfs, the solid squares mark stars identified as members of the thick disk. The upper panel plots data from Valenti & Fischer's (2005) analysis of stars in the Berkeley/Carnegie planet survey; the solid points mark stars known to have planetary companions. Three stars (identified in the text) are almost certainly members of the thick disk.

Figure 4. a-element abundances as a function of metallicity. The lower panel plots data for nearby stars from Fuhrmann (1998), where the open circles are disk dwarfs, the solid squares mark stars identified as members of the thick disk. The upper panel plots data from Valenti & Fischer's (2005) analysis of stars in the Berkeley/Carnegie planet survey; the solid points mark stars known to have planetary companions. Three stars (identified in the text) are almost certainly members of the thick disk.

The thick disk is the extended population originally identified from polar star counts by Gilmore & Reid (1983); current theories favor an origin through dynamical excitation by a major merger early in the history of the Milky Way (Bensby 2004). With abundances in the range —1 < [m/H] < -0.3 (Figure 4, lower panel), the thick disk clearly formed after the Population II halo, but before Type I supernovae were able to drive up the iron abundance. Thus, the thick disk population almost certainly comprises stars from the original Galactic disk, which formed within the first 1-2 Gyrs of the Milky Way's history.

What is the relevance of these observations to planet formation? Valenti & Fischer (2005) have recently completed abundance analysis of the high-resolution spectroscopic data acquired by the Berkeley/Carnegie radial-velocity survey. Their analysis includes measurement of the abundance of Ti, an a-element. The upper panel of Figure 4 shows the distribution of the full sample as a function of [Fe/H], identifying stars known to have planets. Three of the latter stars have a-abundances consistent with thick disk stars (HD 6434, 0.48 MJ planet; HD 37124, 0.75 MJ; HD 114762, 11 MJ), while three other stars (HD 114729, 0.82 MJ; p CrB, 1.04 MJ; HD 168746, 0.23 MJ) have intermediate values of [a/Fe]. These results strongly suggest that, even though planets may be extremely rare among metal-poor halo stars (Gilliland et al. 2000), planetary systems have been forming in the Galactic disk since its initial formation.

Figure 5. The velocity distribution of ESP host stars (right hand panels) compared with the velocity distribution of solar-type stars within 40 parsecs of the Sun.

3.3. Kinematics

Stellar kinematics are usually characterized using the Schwarzschild velocity ellipsoid; probability plots (Lutz & Upgren 1980) allow one to compare stellar samples that include multiple components with Gaussian velocity distributions (see, for example, Reid, Gizis & Hawley 2002). With the completion of the Geneva-Copenhagen survey of the Solar Neighborhood (Nordstrom et al. 2004), we have distances, proper motions, radial velocities and abundances! for most solar-type stars within 40 parsecs of the Sun. There are 1,273 stars within that distance that have 0 < (b — y) < 0.54 and MV ^ 4.0; analyzing their kinematics, we have

(U, V, W; au, aV, aW) = (—9.6, —20.2, —7.6; 38.8, 31.0, 17.0 km s-1) , (3.1)

with an overall velocity dispersion, atot = 52.7 km s-1.

There are 129 ESP hosts with accurate distances and space motions, and their mean kinematics are

(U, V, W; au, aV, aW) = (—4.0, —25.5, —20.4; 37.7, 22.9, 20.4 km s-1) , (3.2)

with an overall velocity dispersion, atot = 48.6 km s-1. All three stars identified as likely members of the thick disk (and two of the possible members) have high velocities (60 to 110 km s-1) relative to the Sun.

f One should note that the abundances in this catalog are tied to the Schuster et al. uvby-based metallicity scale, which has color-dependent systematic errors (Haywood 2002). Those systematic errors can be corrected.

The two velocity distributions are very similar. The total velocity dispersion of the ESP hosts is slightly lower than the field, mainly reflecting the higher proportion of old, metal-poor stars in the latter population (this also accounts for the lower value of <rV for the ESP hosts). Interpreted in terms of the standard stellar diffusion model, <rtot « t 1/3, the observed difference formally corresponds to a difference of only 20% in the average age. The relatively high mean motion perpendicular to the Plane is somewhat surprising, and may warrant further investigation. With that possible exception, however, the velocity distribution of ESP host stars is not particularly unusual, given the underlying metallicity distribution.

Was this article helpful?

0 0

Post a comment