Revised Reference Notes
Revised Reference Notes
Pulsar Planets
In each section of the course, we will provide reference notes. These summarise the main points from the video lessons in that section. They are
intended to be a reference source - allowing you to quickly pick out key facts
and equations without having to re-watch all the videos.
1.1
INTRODUCTION
Exoplanets are planets that orbit other stars. They are hard to see because
other stars are far away - much too far to visit with space-probes, given our
current technology.
We can therefore only study them with telescopes. The problem here is
that the light from any exoplanets is overwhelmed by the light from their
stars, which are a billion times brighter.
Four techniques have been used to study them:
Reflex motion - look not at the planet but at the wiggle it induces in
the star.
Transits - look for the drop in brightness when the planet goes in front
of the star
Gravitational micro-lensing - look for the increase in brightness of a
distant star when a planet passes in front of it and focusses its light on
us.
Direct imaging - use clever techniques to spot the planet against the
glare of its star.
1
1.2
PULSARS
Pulsars are neutron stars left over from supernova explosions. Millisecond
pulsars spin hundreds of times per second and are thought to be pulsars
originally in binary systems, which were spun up by gas from their companion
star.
Because of the regular pulses these objects put out, changes in the distance to them can be measured with exquisite precision. This allows us to
measure the reflex motion caused by any planets very accurately - and in one
pulsar, the reflex motion due to three planets was found.
1.3
If you measure the period P of the reflex motion, you can work out how far
away from the star the planet must be r using the equation
s
r=
GM P 2
4 2
2r Mp
P M
where Mp is the mass of the planet. Note that v is the true velocity of
the star: what we actually observe vo is v0 = v sin i where i is the inclination
angle of the orbit.
If you want to measure the radius of the orbit of the star (due to its reflex
motion), it is given by
r
Mp
=
M
r
(assuming the mass of the planet is much less than that of the star). Once
again, what we measure is not r but r sin i.)
1.4
EXPLANATION?
How can we explain the existance of pulsar planets? One possibility is that
they somehow survived the supernova explosion that formed the pulsar, but
2
that is unlikely for many reasons. A second possibility is that some of the
gas fired out by the explosion falls back in, forms an accretion disk and then
planets form in this disk, in a similar way to protoplanetary disks. Lack of
angular momentum may be a problem for this theory.
We believe that there must have been a second star in the system originally (to spin the pulsar up to millisecond periods) which somehow has gone
missing. A third possibility is that this second star was disrupted and the
planets formed from its debris. Or finally, the neutron star might have sunk
into the core of this second star forming a giant Thorne-Zytkow object, which
might have helped keep pre-existing planets and brought them in closer.
1.5
Two more have been found recently. One is in a globular cluster, which
was probably formed by some sort of capture process as stars in this dense
environment come close to each other.
The second is so close to the pulsar and so massive that it could not
survive unless it were made of something incredibly dense. Perhaps it is the
core of a carbon White Dwarf star which has been stripped of its outer layers,
leaving a core of diamond?
2.1
mp
r
m r
1 + mp
m
This motion is very small - the change in angle as observed from the
Earth, a distance D away is
=
where is measured in radians.
r
D
This equation is explained in much more detail in the first section of our
preceding course Greatest unsolved mysteries of the universe - check that
out for more details if you wish.
In practice this reflex motion is unobservably small, due to telescope imperfections, atmospheric seeing, diffraction and the random arrival of photons
(Poisson noise).
2.2
A second alternative is to look for the reflex motion back and forth along
the line of sight. This would cause a Doppler shift in the wavelengths of
any features in the spectrum of the star. The Doppler effect, and the whole
concept of spectra were also introduced in the first section of our preceeding
course Greatest unsolved mysteries of the universe - check that out for
more details if you wish.
The star will be moving at a velocity
v =
2r
P
which will cause a shift in the wavelength of any spectral line given
by the equation
v
c
where c is the speed of light.
2.3
OBSERVING TECHNIQUES
Once again, this shift is tiny. It is hard to measure because of many possible
distortions in the spectrograph, and in particular the uncertainty and rapidly
moving position of the light on the slit of the spectrograph.
Two approaches have been used to overcome this. One puts the incoming
light through a fibre-optic cable to scramble its position, and then compares
it to reference light from an arc lamp. The second technique puts an iodine
cell in front of the spectrograph slit, to imprint reference absorption lines all
over the spectrum.
2.4
RESULTS
Planets were found - but very strange ones. Hot Jupiters - planets with
roughly the mass of Jupiter, but incredibly close to their stars - in orbits
of only a few days, and 70 Virginis type systems - giant planets in highly
eccentric orbits.
Neither are anything like our solar system, and both were thought to be
impossible because of the snow-line theory, which suggests that giant planets
can only form far enough out for ice to stay frozen.
3
3.1
As time has gone on, radial velocity searches have become more sensitive,
allowing them to detect smaller planets with their smaller wobbles. As the
surveys continue for longer, they also become sensitive to larger planets further out, in orbits of many years in length.
3.2
Where do the hot Jupiters come from? If they form inside a protoplanetary disk, they may set up ripples in the disk which will cause them
to migrate. This could bring them from out beyond the snow-line into the
regions close to the star.
If anything, this mechanism is too effective, and should move all planets
in - there should be nothing except hot Jupiters.
An alternative is planetary billiards: the planets warp each others orbits,
collide, get flung out, and in general end up in confusing elliptical orbits.
3.3
TRANSIT SEARCHES
If a planet is in a nearly edge-on orbit, it may pass in front of the star once
per orbit and cause a slight dip in brightness. If the normal brightness of the
start is B and the change in brightness is B, then the radius of the planet
rp is given by:
s
rp = r
5
B
B
3.4
PLANET RADII
It turns out that transit searches mostly find hot Jupiters, but most of them
are larger than predicted by standard models. The discrepancy is worse for
the hottest planets.
The equilibrium temperature T of the surface of a planet can be calculated
(this was done in our previous course Greatest unsolved mysteries of the
universe) and comes out as
s
T =
L
16D2
3.5
ROSSITER-MCLAUGHLIN EFFECT
As a planet moves across the disk of a star, it blocks the light from different
parts. If the star is spinning, this will cause the apparent velocity of the
star to shift. This is called the Rossiter-McLauchlin effect and allows us to
determine the relative direction of the stars spin and the planets orbit.
The two are often mis-aligned, which is troublesome for disk migration
theories.
4
4.1
It is hard to measure the brightness of stars with much better than 0.1
6
Kepler spent three years staring at one part of the sky. Due to gyroscope
failures it now has to divide its attention between several regions of the sky
for a few months each.
4.2
SUPER EARTHS
In our own solar system, planets fall into two quite distinct categories, the
rocky planets and the giant planets. There is nothing between the largest
rocky planet (Earth) and the smallest giant planet (Neptune).
Kepler shows that in other solar systems, there are large numbers of
planets intermediate between the Earth and Neptune in both radius and
mass. These are called Super Earths, though mini-Neptunes would be a
good name too.
It is not clear what they are like. Some may be giant rocky planets, some
are probably gas planets and some may even be water worlds.
4.3
To find out the true population of exoplanets, we need to remove the various biasses which affect surveys. For example, planets in close-in orbits are
more likely to transit, and their transits will be larger and hence easier to
detect. Thus you would expect close-in planets to be over-represented in any
exoplanet surveys.
One way to correct for this is Monte-Carlo simulations. You guess
what the population of exoplanets might be like, simulate this population,
then put the simulated results through the discovery process and see which
simulated exoplanets would actually have been discovered. Compare this to
the planets actually discovered and if the agreement is good, your guessed
population is consistent with the data.
A second approach is to divide planets up into categories (e.g those with
orbital periods of 2-4 days and masses 0.5 - 0. 8 Jupiter masses). For each
category, you can calculate the probability of a planet with those properties
being seen. Use this to correct the observed numbers and recover an estimate
of the true population.
These methods show that for planets with periods of less than 50 days,
the numbers steadily increase towards smaller masses and larger periods.
4.4
SECONDARY TRANSITS
The great precision of space-based light curves allows for more to be learned
about planets from the details of the brightness changes.
Secondary transits occur when the planet moves behind the star, and
their depth tells us how bright the day side of the planet is. A slope in
the brightness between primary and secondary transits tells us the difference between the day- and night-side brightness of the planet. At optical
wavelengths the planet just reflects the sunlight, so there is no light from
the night-side of the planet. At infra-red wavelengths, the planet will glow
depending on its temperature, so the day-to-night brightness difference tells
us the temperature difference from the hot to the dark side of the planet.
Most exoplanets measured using this technique are quite dark at optical
wavelengths, presumably due to a lack of clouds. Close-in exoplanets have
large temperature differences between day and night, while further out ones
seem to have less of a difference.
4.5
By subtracting a spectrum taken during the secondary transit from one taken
just out of the secondary transit, the spectrum of the day side of the planet
8
5
5.1
Gravitational Lensing
GRAVITATIONAL LENSING
4GM
rc2
where G is the gravitational constant and c is the speed of light (note this is an approximation to the full equation, only valid for small angles).
5.2
4GM
Dc2
where D is the distance from the Earth to the background star. Note
that this is actually just a special case of the true Einstein Radius equation
valid when the lens is half way to the background star.
5.3
MICROLENSING
Unfortunately, these angles are much too small to see when stars or planets
lens other stars in our own galaxy. But while we cant see the multiple images,
9
we can detect the amplification of the background star - its apparent increase
in brightness. This case (gravitational lensing with angles too small to see
but causing increases in brightness) is called gravitational microlensing.
It is most commonly calculated using ray-tracing. Simulated light rays
are sent out from the background star and bent as defined by the equations
above, and the position they land at is recorded in the image plane, an
imaginary plane passing through the Earth.
A single mass causes single spikes in brightness, but when you have multiple masses within a few Einstein Radii of each other you get complex patterns,
of short spikes in brightness superimposed on the overall trend.
5.4
RESULTS
5.5
Recently, a number of short brightness spikes have been found which seem
to come from planets which are not in orbit around stars (or at least are a
long way out). These free-floating planets seem to outnumber stars in our
galaxy. Perhaps they were flung out in the game of planetary billiards which
may have taken place in young solar systems?
10
6
6.1
Debris Disks
INFRARED EXCESS
The IRAS astronomy satellite discovered that many stars, most famously
Vega, are brighter at far-infrared wavelengths than youd expect from their
stellar emission.
For Vega, this excess emission peaked at around 60 m wavelength. We
can use this to estimate the temperature of the disk, via Wiens displacement
law:
peak T = b = 2.9 103 m K
where peak is the wavelength at which the emission peaks, and T is the
temperature. Using this, we find that the excess emission is coming from
something very cold, and hence presumably far out.
We can estimate how far out the emission is coming from by using the
equation for the equilibrium temperature of a planet derived in the first
course:
s
L
4
T =
16D2
where L is the luminosity of the star, D the distance between star and planet,
and is the Stefan-Boltzmann constant.
To produce enough radiation to explain the infra-red excess, given this
low temperature, a planet would have to be unfeasibly large. The fraction
of the stars light which is intercepted by a planet of radius r and distance
from the star D, and hence available to re-radiate is
r2
4D2
But luckily, if the same mass is divided up into smaller chunks of radius
r, the fraction of starlight intercepted and re-radiated is proportional to 1/r.
6.2
DEBRIS DISKS
So it seems that the best way to explain the infra-red excesses is by clouds of
small particles far out from the star. Observations indicate that these clouds
are typically in disk-like geometries.
11
7
7.1
Adaptive Optics
LIGHT GOING THE WRONG WAY
A small fraction of the light in any optical system does not go the right way.
This is a problem when you are trying to image a faint object (like a planet)
close to a bright object (like a star).
The ratio of the brightness of a planet (of radius r ) to its star is
1 r
4 D
2
A second problem is atmospheric seeing. The air you are looking through
is not of uniform temperature. Hot and cold air has a slightly different
refractive index, so the hot and cold bubbles of air distort your image.
The best observatory sites on Earth are at high altitude and have very
smooth air flow, but even there, light is typically blurred over an angle of
around 0.6 arcseconds.
7.2
DIFFRACTION LIMIT
Even from space, however, the light entering your telescope is bent by diffraction. Any wave (like light) usually travels in a straight line as that is the
only direction in which all the waves add up in phase (Huygens principle).
But when light has to pass through a finite aperture, light can add up in
phase over a wider range of angles, blurring the image.
This diffraction blurring places a fundamental limit on how sharp an
image you can get with any telescope, even in space. The blurring angle is
roughly
=
D
where is the wavelength of the light and D is the diameter of the telescope.
7.3
ADAPTIVE OPTICS
13
8
8.1
Direct Imaging
OBSERVE IN THE INFRA-RED
Early attempts at direct imaging could only achieve contrast ratios (the ratio
of the brightness of the star to a detectable planet) of a few hundred to one.
So to have any hope of seeing anything, they needed to make observations
in the infra-red. If planets are hot, they will emit their own light at infra-red
wavelengths which might be quite bright.
As mass is assembled to form a giant planet in the first place, gravitational
potential energy (when the fragments are far apart) will be converted into
kinetic energy (as they fall towards each other) and then into thermal energy
(once they collide).
A rough estimate of the gravitational energy U available in forming a
planet of mass M and radius r is
U=
GM 2
4r
To work out how much this might heat up a planet, you can compare it
to the specific heat capacity times mass. And even a very rough estimate
will show you that newborn giant planets may well be very very hot.
A hot planet will radiate energy. This can be estimated using the StefanBoltzmann equation: P = AT 4 , and it seems that planets will remain hot
for periods on the order of ten million years.
The first successful exoplanet image was of a very massive and hot object
near a very small and cool star. Almost more like a binary star system than
a solar system.
8.2
BROWN DWARFS
There is no clear distinction between a small star and a very large planet. A
star with a mass of less than about 10
8.3
MORE DISCOVERIES
14
allowed a slew of new discoveries, mostly around stars with debris disks. The
planets found are huge, hot, and far out.
One planet was found at optical wavelengths by the Hubble Space
Telescope, orbiting Fomalhaut. But it reflects too much light to be a planet
- it is probably a cloud of dust orbiting a smaller planet.
8.4
MODELLING
8.5
SURVEYS
More recently, systematic surveys for exoplanets have failed to find many. It
appears that very hot, massive and bright planets a long way out are rare.
The microlensing results must therefore come from free-floating planets and
not planets in the outer parts of solar systems.
The technology is improving fast and we can expect more results soon!
Earth-like Planets
9.1
This is possible because such planets are much closer in, and because the low
mass and radius of red-dwarf stars makes reflex motion and transit signals
bigger.
Kepler has found a number of potentially habitable planets, but because
the Kepler stars are so faint and far away it is currently impossible to work
out their mass and density, or to image them.
Some potentially habitable planets have been found about nearby red
dwarf stars. In some ways, red dwarfs look like the ideal place for habitable life - they are common, long lived, and seem to have suitable planets.
Although it is possible that their lack of UV and their flares might make
evolution difficult.
9.2
PLANET COMPOSITION
There are unlikely to be planets with new elements, or even with very different ratios of elements, as the same processes of nuclear fusion are responsible
for elements everywhere.
One possible exception: there may be stars with more carbon than oxygen
(though this is controversial). If true, all the oxygen would combine with
carbon to form carbon monoxide, which would blow away. The remaining
carbon would form planets of graphite, diamond and silicon carbide (rather
than the oxygen dominated rocks in our own solar system).
A second possibility is that some planets might have much higher levels of
radioactive elements than our own, due to enrichment from a supernova just
before they formed. This might change the internal heat, volcanic activity
and plate tectonics.
9.3
KEEPING AN ATMOSPHERE
Volatile elements (like our atmosphere and ocean) are probably delivered to
the Earth by either comets or icy asteroids. But small planets cannot hold
on to volatiles like these.
The typical speed v of an atom or molecule in an atmosphere is given by:
s
v=
3kT
m
v=
2GM
r
where M is the mass of the planet, and r the radius of the planet.
Lighter molecules will move faster and hence have a greater chance of
escape. The typical speed of a hydrogen atom in Earths atmosphere is
roughly three times less than the escape velocity. But a small fraction of the
atoms will move much faster than the typical (mean) speed and so slowly
over time, hydrogen will escape.
9.4
MOUNTAINS
The height of a mountain cannot be too large, or its weight would cause the
rocks at its base to flow like a liquid. On Earth, the practical limit is roughly
10 km (depending somewhat on the type of rock), so Mt Everest is pretty
close to the limit.
The limiting height h should be proportional to:
h
r2
M
where r is the radius of a planet and M its mass. If we assume that all
planets have the same density (i.e. are all made of rock), then M r3 , and:
h
1
r
so smaller planets (like Mars) can have bigger mountains, which seems to
fit.
9.5
FUTURE TELESCOPES
18