2
recorded the first results: some observations of Campbell, Walker and Yang [162]
suggested the presence of planetary masses around some near stars. However, Gordon
et al. [164] were extremely cautious as planets in orbit was just one of the possible
interpretations of the data. Not many took in consideration their studies.
One Year Later, David W. Latham (Harvard Smithsonian Center for Astrophysics) et
al. [163] found strong evidences of what could have been a planet around a star, called
HD 114762. Since Latham‘s planet had a mass at least 10 times the one of Jupiter, it
was hypothesized that actually it was a Brown dwarf, so also this news did not have a
big impact.
In 1992, Alexander Wolszczan (Pennsylvania State University) and Dale A. Frail
(National Radio Astronomy Observatory) used a very precise timing method and
discovered Earth-mass planets in orbit around PSR 1257+12, as pulsar planets, they
surprised many astronomers who expected to find planets only around main sequence
stars. Not many believed that a Solar System planet-like was discovered; anyway this
was the first clue suggesting that planetary formation is an ordinary process.
Then, in 1995 during a conference in Florence, two Swiss astronomers, Michel
Mayor and Didier Queloz (Observatory of Geneva), astonished the entire world by
announcing the first detection of a planet orbiting a sun-like star, 51 Pegasi [56]. After
its discovery, many teams confirmed the planet's existence and obtained more
observations of its properties. It was found that the planet orbits the star in 4.2 Earth
days, and is much closer to it than Mercury is to our Sun, yet has a minimum mass
about half that of Jupiter (about 150 times that of the Earth). At the time, the presence
of a gas giant planet so close to its star was not compatible with theories of planet
formation and this 'hot Jupiter' was considered an anomaly; actually it was not even
believed that this body, named ―51 Peg b‖, could survive there. Computational models
suggested that it is sufficiently massive that its thick atmosphere is not blown away by
the star's solar wind [165].
The discovery of this strange and unpredicted object gave a boost to this
revolutionary astronomical branch: the study of planetary systems beyond our own.
3
Nowadays, the search for exoplanets is imposing itself like one of the most
interesting field of astronomy. As a matter of fact, we can count more than 80 ground-
based and 20 space-based ongoing programs and future projects. More than 370
exoplanets have been discovered with detection rates escalating [3].
Figure 1.1: Number of extrasolar planet discoveries per year as of August 2009.
A decade ago, a handful of giant planets, had been discovered by the radial velocity
method – a revolutionary step. Today Jovian-mass planets and considerably smaller
ones (down to 1.9 Earth-masses in the case of GJ 581e [122]) are known by the
hundreds, and even planetary systems similar to our own have been found.
Gravitational lensing has been used to find a true analog to the Solar System,
somewhat scaled down, consisting of both a Jupiter-like and Saturn-like planet on
circular orbits a few astronomical units from their star [41]. Coronagraphic and related
high-contrast direct-imaging techniques have produced images of nine planetary
systems [3]. Precision photometry has led to the discovery of planets transiting their
parent stars. The transit geometry allows direct detection of a planet‘s emergent light in
some cases, and characterization of its atmosphere. New methods of finding and
characterizing planets -astrometry and optical/infrared imaging-are developing rapidly.
Application of these methods, which will greatly increase not just the inventory of
4
planets but also begin to characterize them, is now limited more by resources than
technology.
Having entered the phase of extrasolar planets characterization, we can constrain
their horizontal and vertical temperature profiles and estimate the contribution of
clouds and hazes [4]. Most importantly it is possible to use the wavelength dependence
of this extinction to identify key chemical components in the planet‘s atmosphere. This
achievement opens up enormous possibilities in terms of exoplanet characterization. In
particular, in this thesis the probing of the atmosphere of HD 209458b in primary
transit in the four IRAC bands at 3.6, 4.5, 5.8 and 8 μm will be presented [5]. We find
our photometry data to be consistent with the presence of water, while the limitations
imposed by the accuracy of our data prevent us from commenting on the possible
additional presence of methane, CO and/or CO2.
As well as detecting the atmospheres of exoplanets, part of the characterization
process undoubtedly involves the search for extrasolar moons. In this work, we
explore the motivations for undergoing such a search, review some of the proposed
detection techniques and we investigate the detectability of a habitable-zone exomoon
around various configurations for exoplanetary systems with the Kepler Mission or
photometry of approximately equal quality. We find that Saturn-like planets offer the
best opportunity for detecting a single large satellite and that habitable exomoons
down to 0.2 M ⊕ should be detectable with the expected performance of Kepler [6].
5
C h a p t e r 2
EXTRASOLAR PLANET DETECTION METHODS
Extrasolar planets are incredibly difficult to detect. This is because any planet is an
extremely faint light source compared to its parent star. For this reason, only few
extrasolar planets have been observed directly [3].
Instead, indirect methods have to be resorted to find extrasolar planets. Here we
present seven different indirect techniques that are currently used:
1. Transit Photometry
2. Radial Velocity
3. Microlensing
4. Astrometry
5. Circumstellar Disks
6. Pulsar Timing
7. Magnetospheric Emission
Many of these methods rely on the measure of the interaction between the exoplanet
and its parent star. By observing changes in the parent star, the existence of the planet
can be deduced. Since the changes become larger as the planet becomes more massive,
it is always easier to detect Jovian planets rather than terrestrial ones.
6
Table 2.1: Basic quantities for planets [7].
Sun Jupiter Earth HD
209458b
Mass (kg) 1.99.1030 1.9.1027 5.98.1024 1.31.1027
MV (mag) 4.85 25.5 27.8 -
Radius (km) 696000 71474 6378 94346
P (days) - 4329 365 3.52
Semimajor
Axis (AU)
- 5.2 1 0.045
RV
semiamplitude
of reflex
motion (m/s)
- 12.5 0.09 86.52
Projected
semimajor
axis at 10pc
(mas)
- 520 100 4.5
Contrast
1 1.82.108 1.5.109 -
Transit
lightcurve
depth (%)
- 1.01 0.0084 1.7
7
2.1 Radial Velocity
Radial velocity (RV) method detects planets by measuring a parent star‘s periodic line
of sight velocity change due to its orbit around a common centre of mass, using the
Doppler shift of the star‘s spectral lines.
Beginning with the detection of a planet half the mass of Jupiter around 51 Pegasi, by
Meyor & Queloz in 1995, the Doppler spectroscopy technique has been the most
successful so far in finding extrasolar planets. In fact, this method is responsible for the
initial detection of over 80% of the exoplanets known today.
Figure 2.1: Schematic view of the wobble of a star due to an orbiting planet as observed from Earth.
The star moves around the barycentre of the planetary system and its spectrum appears blue-shifted as it
approaches the observer and red-shifted when it moves away.
The semi‐amplitude of the stellar radial velocity signal K resulting from an
exoplanetary system is given by
(2.1)
8
where Mp is in Jupiter masses, M∗ is in solar masses, P is in years, and the orbit is
assumed to be circular. From this equation, we see immediately that radial velocity
detection gives the largest signal for massive planets in short period orbits. For
instance, 51 Peg has a 4.2‐day orbital period and Mp sin i of about half of a Jupiter
mass. Its velocity semi‐amplitude is K = 59 m s–1, which is significantly larger than
Mayor & Queloz‘s velocity measurement precision of 13 m s–1. In contrast, the radial
velocity signal of the Sun due to Jupiter‘s orbit is only about 13 m s–1. We note also
that the radial velocity technique is sensitive to Mp sin i, rather than Mp. Thus, RVs
allow only to measure a lower limit to the mass of a planet, unless the inclination i of
the system can be determined by some independent means (e. g. the transit method).
Figure 2.2: Schematic view Orbital parameters of a planet-star system. The star s and the planet p are in
circular orbit around the centre of mass cm of the system. The orbital radii are as for the star and ap for
the planet, these are plotted along the orbital plane. The angle i between the normal to the orbital plane
and the line of sight determines the orbital inclination angle. The radial velocity Vs of the star as
measured along the line of sight (from the upper right in the diagram) depends on the sine of the orbital
inclination angle [8].
Radial velocity measurement precision has continually improved from the initial ~15
m s–1 regime [9] to the current state‐of‐the‐art of 1 m s–1 or better. The attainment of
very high measurement precision has mostly been a matter of controlling systematic
measurement errors. The use of a gas absorption cell or the continual illumination with
9
a separate reference spectrum measures intrinsic spectrograph drifts. Enclosing the
spectrograph in a vacuum and stabilizing the temperature to mK tolerances can result
in the achievement of residuals to planetary orbit fits as low as 10-20 cm s–1. This will
be the case of HARPS-NEF, a high-resolution optical spectrograph born from the
collaboration between Harvard and the Geneva Observatory that will be operational in
2010 on the 4.2-m Herschel Telescope (La Palma, Canary Islands) for follow-up
studies of transit candidates from the Kepler mission [2].
Figure 2.3: Development of Doppler techniques during last decades.
Reaching this level of measurement precision opens up new possibilities for planet
detection. While the radial velocity signal of an Earth‐mass planet in a 1‐AU orbit
around a solar‐mass star is only 9 cm s–1 (Table 2.1), we do have the measurement
precision to detect habitable Earths around low‐mass M stars, and to detect Super
Earths (planets of a few Earth masses) in short‐period orbits around solar‐mass stars.
These Super Earths are objects in the transition region between the masses of
ice‐giants like Uranus and Neptune and the masses of the rocky terrestrial planets of
our inner Solar System. An extremely wide variety of compositions of such planets is
possible, with mixtures of various fractions of metals (iron and nickel), silicates, ―ices‖
(water, methane, ammonia, etc.) and H2‐He. The detection of these objects, and the
measurement of their mass and radius (for those that undergo transits), will open up
exciting new areas of planetary astrophysics.
10
The main factors currently limiting the precision of the RVs, in addition to photon
noise, are stellar noise, the wavelength calibration, telescope guiding, stability in the
illumination of the spectrograph, and detector‐related effects. The only solution to
photon noise is more photons. This may come from either larger telescopes equipped
with high‐precision spectrographs, or more telescope time on existing facilities
permitting longer exposures without compromising the size of the samples of stars
surveyed for planets. Indications are that the remedy to the problem of stellar noise
may be similar. Longer exposures or binning over suitable timescales (again implying
access to more telescope time) can reduce astrophysical jitter to some extent by
averaging out those intrinsic variations (see, e.g., [10]).
Planets often occur in multi‐planet systems. The first of these found was the system
around υ And. Intensive follow‐up observations often reveal the presence of additional
planets in the system. Many times, these planets will be gravitationally interacting with
each other. The study of such interactions helps us to understand the dynamics of
planetary system formation and evolution.
Figure 2.4: Temporal radial-velocity measurements obtained with CORALIE for the 2-planet systems
HD82943 [11].
Doppler studies have revealed that approximately 10% of all F, G, and K stars have
at least one planet in the mass range 0.3−10 M
Jup with periods between 2 and 2000
11
days [12]. Extrapolating to longer periods, it is estimated that 17–19% of stars have a
gas giant within 20 AU. We have also learned that gas giant formation is much more
efficient around stars with super‐solar metallicity (e.g. [13]).
2.2 Transits
The transit method met its first success in 1999 with the transit observation of the
RV planet HD 209458b ([14], [15] and Fig. 2.5, left). It then became popular for two
reasons: 1) the detection of a planetary transit around a bright star requires a telescope
as small as 20-cm in diameter, and many transit surveys were initiated after this first
success; 2) studies of a planetary system seen edge-on (i ≈ 90°) is much richer than for
systems at any inclination, because the radius and the mass are directly measured [16].
After the current success of ground-based transit searches like OGLE, major transit
discoveries are expected from space, or already obtained (Fig. 2.5, right). Several
NASA and ESA missions have been dedicated to exoplanet transits in large part or in
total. The Kepler mission aims to determine the frequency of occurrence of Earth
analogs, by monitoring ~100,000 solar‐type stars (a field in Cygnus) for four years [17].
The CoRoT mission is similarly monitoring numerous solar‐type stars to determine the
frequency of occurrence of Super Earths, and to find and characterize exoplanets
ranging in size from giant planets, down to Super Earths. CoRoT has already identified
new giant transiting planets (e.g., [18]), and the CoRoT team is currently working to
extend the detections to much smaller planets [2].
Figure 2.5: Left: the light curve of HD 209458, showing the first observed planetary transit [14]. Right:
the same system observed by HST/STIS [19].
12
2.2.1 Principles
The transit method consists in detecting the shallow dip in a stellar light curve when a
planet crosses the line of sight towards its host star during its revolution. It thus
requires an almost perfect alignment between the observer, the planet, and the star.
The transit appears periodically, with a period equal to the revolution period of the
planet. The probability P
tr
for a planetary system to show a transit is a direct relation to
the star radius R and the semi-major axis a:
trP R a .
Applying this equation to the solar system, it follows that the Earth in front of the
Sun has a 1/214 probability to produce a transit for a distant observer; for Jupiter the
value is 1/1100. With the discovery of the ―hot Jupiter‖ class of planets, the prospects
of detecting exoplanets by looking for transits improved dramatically. For 51 Peg b,
the transit probability is about 10% and transit events recur every ~ 4 days [16].
A planetary system observed edge-on, as in the case of transiting systems, also offers
the geometry for the strongest radial velocity (RV) signal to be observed; moreover,
the ―sin i‖ indetermination for the mass disappears, so studying the planet transit light
curve we can determine its physical parameters: radius, orbital inclination i, density,
surface gravity, orbital distance a.
2.2.2 Measured parameters
The planetary transit measured in the stellar light curve is mainly described by three
parameters: depth, duration, shape. Depending on the latitude of the transit on the
stellar disk, the transit light curve will be U-shaped (central occultation) or V-shaped
(grazing occultation). Quantitatively, the related parameter is the duration of the ingress
and egress (alternatively, the duration of the flat bottom of the transit) [16].
Let us calculate these parameters in the simplified case of a circular orbit and a stellar
disc of uniform brightness. The sketch of a planetary transit is given in Figure 2.6.
Ingress (resp. egress) is defined as the phase from contact 1 (resp. 3) to contact 2 (resp.
4). The ―flat bottom‖ corresponds to phases 2 to 3.
Ricevi informazioni sui nostri servizi, sulle offerte e non perdere news e consigli su università e lavoro.
