PRESUPUES TO MINIMO

By looking at Figure2.3, where I plot the metallicity distribution of stars with planets 3_{, one} realizes immediately that the Sun, with [Fe/H] ≡ 0.0, falls in the low-metallicity tail. Indeed, the distribution peaks at [Fe/H] ' 0.3, showing evidence of moderate metal-enrichment with respect to the average metallicity ([Fe/H] ' −0.1) of field dwarfs in the solar neighborhood. The possibility that super-solar metallicity could correspond to a higher likelihood of a given star to harbor a planet was investigated since the first detections by precision radial- velocity surveys (Gonzalez 1997, 1998a, 1998b; Fuhrmann et al. 1997, 1998; Laughlin & Adams 1997). A number of studies have been performed throughout these years, using increasingly larger sample sizes, employing both spectroscopic and photometric techniques for metallicity determination, and adopting control samples of field stars without detected planets (Santos et al. 2000, 2001, 2003, 2004a; Reid 2002; Laughlin 2000; Gonzalez & Laws 2000;Gonzalez et al. 2001;Israelian et al. 2001;Queloz et al. 2000a;Smith et al. 2001; Gim´enez 2000; Martell & Laughlin 2002; Heiter & Luck 2003;Sadakane et al. 2002;Pinson- neault et al. 2001;Murray & Chaboyer 2002; Laws et al. 2003; Fischer et al. 2003b).

The global trend is that planet-harboring stars are really more metal rich than stars without known planets. Based on observationally unbiased stellar samples, the strong de- pendence of planetary frequency on the host star metallicity has been clearly demonstrated by Santos et al. (2001), and confirmed by Fischer et al. (2003b) and by Santos et al. (2004a), who showed a sharp break in frequency at [Fe/H] ' 0.0, albeit in the presence of very small-number statistics at the low-metallicity end of the distribution.

The absence of short-period transiting planets in the globular cluster 47 Tucanæ is also

3_{Using Iron (Fe) as the reference element, then [Fe/H] = log[NF e/NH]?− log[N}

F e/NH]¯, with NF e, and

Figure 2.3 Metallicity distribution of the planet-host stars

used by Gilliland et al. (2000) and by Weldrake et al. (2005) to argue that low-metallicity stars are less likely to harbor giant planets. The claims by these authors suffer however from some ambiguity, as in the cluster core investigated by Gilliland et al. (2000) with HST transit photometry crowding could play a significant role in giant planet formation, migration, and survival. The outer regions of the cluster monitored by Weldrake et al. (2005) are less affected by crowding, however the lower occurrence rate of hot Jupiters in a metal-poor environment does not rule out the existence of a population of giant planets at wider radii4_{, and other mechanisms could be called into question to explain this result (see} Section2.3).

Many authors have debated whether the observational evidence is an indicator of pri- mordial high metallicity in the planet host stellar sample, or if the trend with [Fe/H] could be due to selection effects or pollution by ingested planetary material. The recent analyses performed by Santos et al. (2001, 2004a) and Fischer et al. (2003b) are almost conclusive

4_{Sigurdsson et al. (2003) have recently inferred a mass of a few Jupiter masses for the third, long-} period component orbiting the white dwarf - pulsar system B1620-26 in the globular cluster M4, five times more metal-poor than 47 Tuc, providing the first evidence for planet formation in extremely metal-poor environments

arguments that observational selection biases do not play a major role (see Gonzalez (2003) for a thorough review of the subject of biases).

The idea of pollution is also losing credit among the scientific community, primarily based on the evidence of no correlation between [Fe/H] and effective temperature Teff, or

convective envelope mass Mconv, for the planet host sample (e.g. Fischer et al. 2003b;Santos

et al. 2003, 2004a, and references therein). Results on this specific issue are not conclusive yet, however. For example, Vauclair ( 2004) has recently pointed out how the absence of a [Fe/H]-Mconv correlation does not automatically imply that stars with planets have not been

polluted.

Furthermore, theoretical calculations (Montalb´an & Rebolo 2002;Boesgaard & King 2002) suggest that detection of anomalous abundances of rare elements such as lithium (Li) or beryllium (Be) could be interpreted as evidence for accretion of planets into the atmosphere of a star. The abundances of Li isotopes in the spectral region around the 6707˚A line in planet-host stars have been investigated in the recent past by several authors (Gonzalez & Laws 2000; Ryan 2000; Israelian et al. 2001, 2003, 2004; Reddy et al. 2002; Mandell et al. 2004), and similar studies have been conducted for the BeII lines at 3130˚A and 3131˚A. (Garc´ıa L´opez & P´erez de Taoro 1998; Deliyannis et al. 2000;Santos et al. 2002,2004c).

While evidence for Li excesses in some planet-harboring stars has been reported in the literature (Israelian et al. 2001, 2003; Laws & Gonzalez 2001), clearly suggesting that ac- cretion of planetary material can actually take place in some stars, in general stars with planets have normal light-element abundances, typical of field stars. It thus seems unlikely that pollution effects can be responsible for the overall metallicity enhancement of the planet host stellar sample. Finally, analyses of over a dozen other elements have been carried out in the recent past (Santos et al. 2000; Gonzalez et al. 2001; Smith et al. 2001; Takeda et al. 2001; Sadakane et al. 2002; Bodaghee et al. 2003; Ecuvillon et al. 2004a, b), and the general evidence is that the abundance distributions in stars with planets are the extension of the observed behavior for [Fe/H], a result quantified by trends of decreasing [X/Fe] with decreasing [Fe/H].

As of today, the best explanation for the metallicity excess in stars with planets is that the enhanced [Fe/H] is primordial in nature. I defer to Section 2.3 a discussion of what this

Figure 2.4Left: minimum mass histogram of the extrasolar planet sample. Center: the distribution of orbital periods. Right: the eccentricity distribution

might imply from the point of view of theories of giant planet formation.