We have seen that there are a number of ways of detecting the existence of extrasolar planets, most indirect. All the techniques have their own advantages and disadvantages and the different selection effects of these detection methods are summarised in Fig. 1.10, on which are plotted the mass and orbital radii of known exoplanets, together with characteristics of the Solar System planets.
Currently employed detection methods are biased towards close-orbiting heavy planets and thus very few lighter terrestrial-like planets have so far been found, with the lowest mass for planet orbiting an active star so far being estimated as 5.5MEarth (Sect. 1.3). Three earth-mass extrasolar planets have actually been discovered, but these do not orbit a main sequence star, but instead have been observed orbiting the pulsar PSR 1257+12 (Wolszczan and Frail, 1992; Wolszczan, 1994). Although no terrestrial planets have so far been discovered, there is no reason to think that they are not present and as measurement techniques improve, it is widely hoped that terrestrial planets may soon start being detected.
As can be seen, the radial velocity technique is best for detecting heavy, close orbiting planets, and thus the planets found so far are clustered in the top left corner of Fig. 1.10. The limit of detectability of existing measurements is shown, together with the expected improvement due to ever increasing sensitivity and longer observation runs. Radial velocity programmes currently under way include the Anglo-Australian Planet Search (e.g. Carter et al., 2003), the California and Carnegie Planet Search,
Fig. 1.10. Selection effects of different exoplanet detection programmes. Solar System planets are indicated by letter.
ELODIE (Naef et al., 2004) and CORALIE (Mayor et al., 2004). Analysis methods are rapidly becoming more sophisticated and Charbonneau (2006) note that the radial velocity method may soon start turning up terrestial-like planets (Sect. 1.3). Transit observations also favour shorter periods, but can also detect lighter planets, especially the planned space-based missions (Sect. 1.4.1). The selection limits of astrometric observations (Sect. 1.2.2), both for ground-based programmes such as at the Keck Observatory and the VLT, and for space-based missions such as the forthcoming Space Interferometry Mission (SIM), are expected to start probing into the terrestrial planet region of the diagram. This technique is complemented by the microlensing technique (Sect. 1.2.4) which similarly is more sensitive for space-based proposals (such as GEST), than terrestrial ones. We will now consider future observations using these methods and what may reasonably be expected of them.
A number of transit surveys have been planned for the next few years - both ground- and space-based. They may be conveniently split into two categories: 1) Deep surveys, which have small pixel size, can see faint stars, but do not cover a wide area of sky; and 2) Wide surveys, which cover a wide area of the sky with large pixel size, but cannot see fainter stars. The a priori probability of transit detection is the ratio of the diameter of a star to the diameter of a planet's orbit (Lewis,
^icrolensing - GEST
2004). For the Sun-Jupiter system this is 1.4 x 1011/5.2 x 1013 = 1.8 x 10~3. Hence, transit surveys need to observe lots of star systems to have sufficient probability of planetary detection and so both survey methods attempt to view many stars simultaneously, taking special care not to confuse real planetary transits with other phenomena (e.g. Mandushev et al., 2005).
One ground-based wide survey is TrES (Transatlantic Exoplanet Survey), a network of three 99-mm aperture field-flattened Schmidt telescopes based at Palomar Observatory, Lowell Observatory and the STARE transit camera at the Observatorio del Teide, Canary islands. STARE made the first observation of a planetary transit (HD 209458 b) (Sect. 1.2.3) and the TrES network has now discovered two further planets, TrES-1 (Alonso et al., 2004) and TrES-2 (O'Donovan et al., 2006a). The infrared radiation of TrES-1 has since been observed directly with the Spitzer telescope by Charbonneau et al. (2005) who report a surface temperature of over 1000K. Another ground-based wide-survey is SuperWASP, which has two facilities: one based in the Canary Islands and another soon to be operating in South Africa. Both instruments are comprised of five 11-cm aperture wide-angle cameras and are developments of the WASP0 prototype instrument (Kane et al., 2004; Pollacco et al., 2006). Both SuperWASPs began operations in 2004 and are able to monitor 10,000 stars simultaneously over a 15o x 15° field of view. So far, two planets have been discovered, WASP-1b and WASP-2b (Collier Cameron et al., 2007).
Space-based wide-survey observations are predicted to be both more sensitive and less prone to false identifications of planetary transits. One such project was the Sagittarius Window Eclipsing Extrasolar Planet Search (SWEEPS), which used the Hubble Space Telescope in 2004 to search for transit events (Sahu et al., 2006). A current mission is COROT, which is a French-led project to place a small 0.27-m telescope into orbit to study astroseismology and also detect extrasolar planet transits. COROT was launched on 27 December 2006 and started its first observing run on 8 February 2007. It will observe an area of the sky of size 2.8° x 2.8° for 22 years. The US-led Kepler mission is a Schmidt telescope with 1.4-m primary mirror and a 0.95-m aperture, due for launch in October 2008. Kepler will continuously and simultaneously monitor the brightness of approximately 100,000 A-K dwarf (main-sequence) stars brighter than 14th magnitude in the Cygnus-Lyra region along the Orion arm, for a period of 4 years.
Deep transit surveys will also be conducted by microlensing programmes, which are outlined in Sect. 1.2.4.
Almost all of the currently known exoplanets have been discovered through indirect methods. However, four extrasolar planetary-mass objects have now been directly imaged about brown dwarfs, although their masses are towards the top end and in some cases exceed what might really be classified as planets and instead might be better described as brown dwarfs 13Mj). The objects are: 2M1207 b (5 — 8Mj) (Song et al., 2006; Chauvin et al., 2005a), GQ Lup b (10 — 40MJ) (McElwain et al., 2007), AB Pic b (13 — 14MJ) (Chauvin et al., 2005b) and SCR 1845 b (9 — 65MJ)
(Biller et al., 2006). All objects except SCR 1845 b orbit at great distance from their parent stars. In addition, we have also seen that methods have successfully been developed to study the spectra of exoplanets though differencing methods (Sect. 1.2.3). In this section we will look at other methods of directly detecting the reflected starlight or thermal emission of exoplanets about nearby stars.
As a planet orbits a star, part of the starlight will be reflected by the planet towards the observer. This component will be Doppler-shifted by an amount depending on the planet's orbital velocity 100 km/s), rather than the star's 10 m/s) and can be extracted using very high-resolution ground-based spectroscopy and correlation techniques. Collier Cameron and Leigh (2004) review the current status of a number of direct planetary detections achieved by this technique. Assuming the planetary radius is known, these observations may be used to estimate the visible planetary albedo. For example, the albedo of t Bootis b (Collier Cameron et al., 1999; Charbonneau et al., 1999) is estimated by Leigh et al. (2003a) to be less than 0.39. Similarly, the albedoes of v Andromeda b and HD 75289 b are estimated to be less than 0.3 (Collier Cameron et al., 2002) and 0.14 (Leigh et al., 2003b) respectively. These albedoes are much less than Jupiter's (0.5).
Models of the expected spectra of extrasolar giant planets (e.g. Sudarsky et al., 2003) show that the reflected sunlight from such planets will be significantly affected by absorption of atmospheric constituents such as sodium and carbon monoxide, whereas the stellar spectrum is expected to be smoothly varying. Hence, these absorption features provide a possible means of discriminating between the light reflected by a planet and the direct stellar light. Wiedemann et al. (2001) report just such a detection of the 3-yU,m methane absorption of t Bootes b. There are other programmes in development, such as TRIDENT (Marois et al., 2005) on the 3.6m Canada-France-Hawaii-Telescope (CFHT), which observes the edge of a methane absorption band between 1.5 and 1.8 ¡m.
Using two telescopes, separated by a long baseline of precisely controlled optical length D, the beams may be combined with a phase difference of n to completely eliminate the light from the central star. Constructive interference will then occur at a number of angles 0 where D sin 0 = (2n + 1) A/2 and n is an integer. By varying the baseline D (assuming fixed wavelength A), a range of constructive interference angles can be examined to attempt to detect either the weak stellar reflection or thermal emission of an extrasolar planet. Both the Keck Observatory and VLT have long baseline interferometric observation programmes in development. Another interesting project is the Large Binocular Telescope Interferometer (LBTI) in Arizona, which achieved 'First Light' in October 2005. LBTI uses adaptive optics and a beam combiner including a dielectric material to correct for colour dependence of light interference. LBTI will operate in the infrared (3-5 ¡m) and should be able to detect planets further than 0.03" from their stars. Nulling interferometry is also the planned mode of the proposed ESA Darwin space mission, and a possible mode of the proposed NASA Terrestrial Planet Finder (TPF-IR). In these mission plans, a fleet of large telescopes would fly in formation and the light combined in a central hub using precisely controlled phase delays. Due to their very long baselines, low temperatures and no atmospheric absorption, these missions will be able to not only directly detect extrasolar planets in the infrared, but also measure their emission spectra, allowing the composition of their atmospheres to be determined.
Finally, in this technique, which is only suitable for space missions, light from the central star is eliminated using a mask in the focal plane. The method is used in solar studies to study the corona and prominences of the Sun's atmosphere, from which its name is derived. The technique may be used by a version of the proposed NASA Terrestrial Planet Finder (TPF-C). In addition, the new James Webb Space Telescope (scheduled for launch in 2013) will house the NIRCam instrument, which includes a coronagraphic module, operating from 2-5 ¡m. This system will be capable of 108-109 high contrast imaging for separations > 0.1 arcsecond. In addition, tunable narrow-band filters will allow the measurement of spectra from 2.5-4.5 ¡m at low resolution.
Was this article helpful?