1.4
Current Status of Extra-Solar Planet Searches
As of March 20th 2009, a total of 331 detections of extra-solar planets have been published8. Fig. 1.15 shows the number of planets detected with each of the methods described in the previous section.
Figure 1.15: Number of planets detected with each method.
The most successful technique is the RV method which has provided 75% of all detections. Note that since only a minimum mass of the candidates can be derived, some might actually be too massive to be a planet. However, in a statistical view, most of the candidates detected by the RV method are indeed planets. Radial velocity searches will continue with focus on the detection of very low-mass planets, multi-planetary systems, planets with longer periods and planets around low-metallicity stars and giants.
A total of 57 transiting planets are known to date9. Remarkably, more than half of the transiting planets have been found in the past two years, making the transit method similarly successful in that period compared to the radial velocity method (see Fig. 1.15). The majority of the recently detected transiting planets were found in wide-angle surveys which target bright stars such as WASP, HAT, TrES or XO (Pollacco et al., 2006; Noyes et al., 2008; O’Donovan et al., 2007; McCullough et al., 2005). Also the space mission CoRoT has contributed by adding six new discoveries (Aigrain et al., 2008). Deep surveys such as OGLE which are targeting highly crowded regions of the Milky Way disk have not been able to keep up with the increased detection rate of all-sky monitoring programs
8 for a daily update see www.exoplanet.eu 9 five of which were detected with the RV method
mainly due to limited amount of observation time and a lower number of target stars. However, upcoming projects such as Pan-Planets or OmegaTranS will presumably give rise to a revival of the deep surveys in the next years. The first detection of a transiting Earth-mass planet in the habitable zone is expected from the Kepler space mission that has been launched recently.
Although each of the other methods (imaging, timing and microlensing) has its own specific aspects and although each of the detected objects is very interesting, a large number of detections is not expected in the near future. This is mainly due to the fact that these methods are not applicable in large scale surveys but require detailed analyses of single objects. An exception might be a microlensing space mission that would have both the photometric precision and the time sampling that is needed to detect a large number of planets in a small field of view (Bennett et al., 2004). Note also that the microlensing technique is currently the only method capable of detecting free-floating planets or planets outside the Milky Way (e.g. in M31).
Figure 1.16: Comparing the number of planets detected with the RV and transit methods. In the last years the detection rate of transiting planets increased significantly whereas the RV detections flattened out at a rate of∼30 detections per year.
The high number of extra-solar planets that have been found over the last decade allows a detailed statistical analysis of their parameters (see e.g. Udry & Santos (2007)). In the following we discuss the most important results that have been found.
1.4. CURRENT STATUS OF EXTRA-SOLAR PLANET SEARCHES 25
The minimum mass distribution of RV planets
The analysis of the mass distribution of extra-solar planets (Fig. 1.17) indicates that although the RV method is more sensitive to massive companions, the frequency of planets decreases as a function of mass (Jorissen et al., 2001). The distribution falls to a value close to zero for masses above 10 MJup. In the Brown Dwarf regime (between 13 and 80 MJup), there is a deficiency of companions to solar-type stars which is called the Brown Dwarf desert (Halbwachs et al., 2000). Together with the shape of the mass distribution, this suggests a different formation mechanism between low-mass stars and planetary systems.
0 2 4 6 8 10 12 M sin i [M_Jup] 0 10 20 30 40 50 60 N
Figure 1.17: Minimum mass distribution of planets detected with the RV method.
The period distribution of RV planets
Fig. 1.18 shows the period distribution of planets detected with the RV method. Although the RV method has a selection effect towards finding short period planets (K∼a−pl1, see eq. 1.11), the resulting bias in the period distribution of Jupiter-sized planets is assumed to be low, because the precision of most RV surveys is high enough to detect most of the massive planets, even with longer periods (Cumming et al., 2008).
Giant-planets with periods below 3 days have been named Very Hot Jupiters (VHJ) in contrast to Hot Jupiters (HJ) with periods between 3 and 5 days10.
0 1000 2000 3000 4000 period [days] 0 20 40 60 80 100 120 N 0 20 40 60 80 100 120 140 period [days] 0 10 20 30 40 50 N
Figure 1.18: Period distribution of planets detected with the RV method.
transiting planets RV planets transiting planets HD80606b HD17156b HAT−P−2b 0.01 0.1 1.0 0.0 0.5 1.0
semi−major axis [AU]
eccentricity
1.4. CURRENT STATUS OF EXTRA-SOLAR PLANET SEARCHES 27
The eccentricity distribution of RV and transiting planets
Unlike the planets in our Solar System, extra-solar planets have been found to exhibit signifi- cant eccentricities. Fig. 1.19 shows the eccentricity distribution of all planets detected with the RV and transit method as a function of semi-major axis. For distances larger than 0.1 AU, the values range from 0 to more than 0.9 with a distribution that is similar compared to eccentricities found in binary star systems (Halbwachs et al., 2005). However, the high eccentricities cannot be explained by the standard giant planet formation models. Several eccentricity pumping mechanisms have been proposed such as interactions in multiple systems (Juri´c & Tremaine, 2008; Rasio & Ford, 1996) or the interactions between the planet and the disk of planetesimals (Murray et al., 2002).
For orbital distances smaller than 0.1 AU, the average eccentricity drops significantly which is a result of a circularization process caused by tidal interactions of the planet with the star (see§1.2.3).
Multi-planet systems
Among the 254 planets detected with the RV technique, there are 79 planets in multi-planet systems. 23 systems with 2 planets, 8 systems with 3 planets, 1 system with 4 planets (HD160691) and 1 system with 5 planets (55 Cnc) have been detected. However, this is considered to be only a lower limit. Ongoing monitoring of most planet systems is continuously increasing the number of known multi-planet systems. Many RV curves of stars with a single planet show long term trends that indicate the presence of a second planet with a longer period (see e.g. Fischer et al. (2001)).
Taking these considerations into account and given the higher rate of multi-planet systems in the older long-running RV surveys, it seems likely that most stars form systems of planets rather than isolated, single planets (Udry & Santos, 2007).
The frequency of giant planets and its metallicity dependence
Analyzing the results from several RV surveys, Fischer et al. (2005) found a strong correlation between the frequency of extra-solar planets and the host star metallicity. Fig. 1.20 shows the frequency of planets with orbital periods less than 4 years with a RV amplitude larger than 30 m/s. The frequency rises from 4% for solar metallicity to 27% for [Fe/H]=0.5. The frequency for sub-solar metallicity is consistent with being constant (see also recent results by Sozzetti et al. (2009)).
There are two theories that explain the metallicity dependent frequency of Jupiter-sized planets. The first scenario is favored by the CAGC planet formation model (see§1.2.1) and is assuming that the overabundance of heavy elements in planet-hosting stars is primordial (Ida & Lin, 2004). A collapsing molecular cloud with high metallicity has a larger number of dust particles which can build more planetesimals in the agglomeration and accretion phase. Naturally, the chance for the formation of a giant planet will be higher if more planetesimals are present in the proto-planetary disk.
The second scenario assumes that the planet-hosting stars are actually not metal-rich but only the outer layers (that are observed to derive the metallicity of a star) are polluted by the debris of the planetary system (Laughlin & Adams, 1997; Gonzalez, 1998). This would be in agreement with the DI planet formation model (see §1.2.2) which predicts a very weak metallicity dependency of the giant planet frequency.
Figure 1.20: Planet frequency as a function of host star metallicity.
If the overabundance is coming from pollution of the outer layers, mixing processes in the convection zone of the stars should diminish the effect and since the depth of the convection zone depends on the spectral type of the star, there should be a trend with effective temperature. No such trend has been observed, most groups are therefore in favor of the primordial metallicity enhancement (Valenti & Fischer, 2008; Ecuvillon et al., 2006; Pinsonneault et al., 2001). However, Pasquini et al. (2007) analyzed the metallicity distribution of 14 planet-hosting giant stars and found no evidence for a metallicity dependent frequency. Since giant stars have a much larger mass of the convection zone they conclude that planet hosting main-sequence stars might indeed be polluted by infall of debris whereas the pollution disappears as soon as the convective mixing increases during the transition to the giant phase.
The masses and radii of transiting planets
Fig. 1.21 shows the radii and masses of all transiting extra-solar planets with M < 5 MJup. Most planets have densities in the range from 0.4 to 1.5 g/cm3(shown as dashed lines). Three planets have significantly larger radii than predicted by models for core-less planets densities (Fortney et al., 2007; Burrows et al., 2007). The origin of the large radii might be explained by additional heat sources such as more significant core heating, tidal dissipation due to a non-zero eccentricity and/or irradiation (for a detailed discussion see Liu et al. (2008)).
1.4. CURRENT STATUS OF EXTRA-SOLAR PLANET SEARCHES 29 TrES−4b WASP−12b WASP−15b SWEEPS−04 WASP−10b HD80606b HD17156b ρ=0.4 g/cm3 ρ=0.8 g/cm3 ρ=1.5 g/cm3 0 2 4 0.0 0.5 1.0 1.5 2.0