1.
Direct detection
By reflected starlight, need
to see object 109 times fainter than star (for Jupiter size); using
infrared radiation from the planet, this could be reduced to ~105,
but that is still too large for any present telescope.
Need
“nulling interferometer” or advanced coronograph (+ excellent
adaptive optics if from the ground).
There are claims that Jupiters will be directly detected from adaptive optics ground-based telescopes (Keck interferometer).
Long-term future space missions to
directly image Earth-like planets and get spectra: Terrestrial Planet Finder (TPF), Darwin (Europe), Planet
Imager (PI)
Spectra could give direct signature
of “biomarkers”, especially ozone (since shouldn’t be much oxygen before
photosynthetic life).
Even without spectra, just
photometric “light curve” (as the planet rotates) could in principle give
evidence on fraction of surface covered by clouds, land mass, oceans, ice,
vegetation.
2. Center of mass wobble of parent star: Favors detection of
massive planets (can get “Jupiters”, but won’t get terrestrial planets this
way). Three completely different
approaches are used:
A. Periodic
change in stellar position:
called “astrometric” (means
“measuring positions) or “angular perturbation” method.
This method has the longest history, but several false detections in
past; reason is it takes decades to
use this method (see below).
See “wavy” motion of star because center of mass is wobbling. (Illustrations)
Most sensitive to distant
(from parent star) planets [Understand why]Þ long periods Þ have to wait many years!
Also, need astrometric (position) accuracy of milli-arcseconds to get “Jupiters,” micro-arcseconds to get “Earths.”
To date no discoveries (one
verification of r.v. method), but that is a selection effect. If there are more giant planets far from
their stars, the should start showing up within a few more years.
Many programs at work from the ground: PTI gets 50 micro-arcsec, Keck interferometer
will get 10-30. But will need future
ambitious space missions (SIM [2007?], GAIA [2010?;5 year
lifetime]—hundreds of thousands of solar-like (FGK) stars will be
searched to detect thousands of Jupiter-like planets, and maybe some Earth-like
planets.
B. Periodic
change in stellar radial velocity (r.v.):
Most sensitive to planets close to parent star. But only gives lower limit to masses because
orbital inclination matters .
Can get orbital eccentricity (see figure in text).
Need to detect r.v. variation of ~10 meters/sec to see
Jupiter. Can now get down to ~2-3
meters/sec (optimistic), but 1 meter/sec seems like absolute limit. So will probably never detect Earth-like planets this way.
But it is
fantastically successful at detecting massive (Jupiter-like) planets: about 80
detected so far! In fact, with one or
two possible exceptions, it is the only
method that has so far (because of its nature) been successful.
Detailed results summarized in class and in text. Lots of big surprises: “hot Jupiters”, large
orbital eccentricities (completely unlike our solar system) in many cases. At least one 3-planet system. Planet in a binary system.
C. Timing
methods –periodic light travel time variations in a stellar “clock”.
1992: Pulsar planets discovered
(from delays in the light arrival time of pulsar pulses, because pulsar’s
distance from us is changing.)
PSR
1257+12: 3 planets; distance ratios almost identical to those of innermost 3
planets in our solar system! But how
could these planets have formed??
Surely couldn’t have survived the supernova that preceded the pulsar!
Current searches: using white dwarf
oscillations as clocks (UT program)
3. Photometric methods
A. Transits: Searches for signs of
eclipse of star by orbiting planet (similar to how eclipsing binary stars are
discovered).
Must look for very small (< 0.1%)
dip in light curve, because planet is so small compared to star.
Need nearly edge-on planetary orbit,
so chances of detection are only ~1%.
But if you monitor many thousands of stars… Rewards are very large: can
get mass and size of planet as well
as information about the planet’s atmosphere.
[In 2001 HST observed starlight filtered through a planet’s atmosphere
during a transit.]
With simultaneous transit and r.v. data, can get even more
detailed information.
However, difficult to get such photometric accuracy.
Jupiters: require milli-magnitude (~
0.1%) Þ can do from ground. STARE, VULCAN,… See web site “Extrasolar
Planet Encyclopedia” for amazingly large list.
Earths: require <micro-magnitude Þ must go to space: COROT (Europe), KEPLER
(U.S., approved for 2007; 105 stars!). Kepler will be able to detect transits that reduce star’s light
by only about 0.01%, so will surely detect lots more massive planets, but also
(if they exist) some terrestrial-mass planets.
B. Gravitational microlensing: This
method uses stars in our Galactic bulge as sources of light rays which are bent
by the gravitational fields of the “lens” stars in the foreground, between us
and the Galactic bulge. This gives a
“microlensing light curve” that rises and falls. Planets that orbit these
“lens” stars can be detected when the light rays from one of the lensed images
pass close to a planet orbiting the lens star.
The gravitational field of the planet distorts the light curve: the
deviation is typically about 10%, and duration is a few hours to a day
(compared to 1-2 months for the lensing due to the star).
Unique advantage: Strength of signal is nearly independent of planetary
mass! Microlensing signals of low-mass
planets have shorter duration and lower detection probability compared to
high-mass planets, but not a weaker signal.
So microlensing surveys with frequent observations of large number of stars should be able to
detect terrestrial planets with good confidence.
The big challenge is that microlensing events are rare, so have to
monitor millions of stars, and even
of those that lens, only about 2% of earth-mass planets orbiting these stars
will be in right position to be detected (if all the stars have earth-mass
planets). Also need very good angular
resolution and fairly accurate (~1%) photometry. Several other problems, but these are being addressed.
GEST (Galactic Exoplanet Survey Telescope)—1.5m
space telescope with large field of view.
Will survey about 100 million stars.
Could detect planets down to Mars mass, should find ~100 Earth-mass
planets at 1AU (if all stars have such planets). “Free-floating” planets will also be detected! (Only method that
can do that.) Will also be able to detect ~50,000 giant planets by
transits. Sensitive to planets at
nearly all distances from star, unlike other methods.
