The field of extrasolar planets has exploded over the last couple decades, providing an entirely new suite of information on planetary systems in general and perspective on our own solar system in particular. One of the most exciting advances came when astronomers were at last able to see a few of these planets directly. Even those who maintain "I'll believe it when I see it" now have visual proof that our solar system is not alone in the universe. However, science advances rapidly, and the discoveries announced only a few years ago are already ripe for an update. Revisiting three direct detections provides confirmation, archival data, and some spirited scientific discussion.
"Hubble's Universe" is a recurring broadcast from HubbleSite, online home of the Hubble Space Telescope. Astrophysicist Frank Summers takes viewers on an in-depth tour of the latest Hubble discoveries.
Welcome to Hubble's Universe Unfiltered.
I'm Dr. Frank Summers of the Space Telescope Science Institute.
In this episode, we're going to revisit some of the discoveries I told you about in
a previous episode.
Those discoveries were about extrasolar planets.
And here we have an artist's illustration of a planet around a red dwarf star.
That's the idea, extrasolar, not the sun, a planet around a star that's not the sun.
And the discoveries of extrasolar planets have been fast and furious for the past couple
of decades.
Let me illustrate that for you.
Here is a plot of stars across the entire night sky.
On the x-axis here is called right ascension, and on the y-axis is declination.
These correspond to roughly longitude and latitude on Earth, but this is on the night
sky.
So you can see all the stars we have in the night sky, how many of them have extrasolar
planets?
Well, the answer, back in 1987, was none.
And what I'm going to plot for you here is the year of the first publication of the
orbit of a planet around another star.
And the first ones actually come earlier than I had expected.
The first one came in 1988, you can see that one dot up there, and then another in 1989,
and nothing in 1990, 91, or 92, but we got a third one in 1993, nothing again in 1994,
and then the watershed one came in 1995.
This one right here, this was the one that was not only published, but was also publicized
because it was a firm discovery.
This one is 51 Pegasi, and it was announced in September 1995, and quickly confirmed by
another group about a week later.
This is when we as astronomers, and you as the public, first knew that there were planets
around other stars.
And for the rest of the 1990s, 1996, 97, 98, and 99, there were just a few discoveries
each year, because there were only a couple groups that were capable of finding these
planets.
However, once astronomers know that it can be done, well, everyone's going to jump
on the bandwagon.
Between 1999 and 2000, you can see there's a lot more discoveries covering a lot more
of the sky, because other groups had gotten into the act, had gotten enough data that
they could announce their own discoveries.
And then it accelerates, and you watch in 2001, 2002, 2003, 2004, 2005, 2006, 2007,
2008, 2009, and 2010, that entire decade, we filled out the sky with planets around
other stars.
The other thing you might notice is that there are a couple clumps.
You can see a clump here, and a clump over here.
These were targeted surveys looking in one area of the sky for extrasolar planets.
And the biggest such survey was the Kepler mission.
And the Kepler results came out in 2011, and it's quite obvious where they are.
You can see this big clump of stars here from the Kepler mission, and we have continued
in 2012 and up through April of 2013.
This is what the graph of the stars and the night sky that have extrasolar planets.
There are over 700 stars with planets around them.
And that's just from one by one measure.
The website that I got this from, other ones say there are about 800.
Furthermore, Kepler has over 2,600 candidates that it can't confirm yet.
So there are almost over 3,000 planets that are confirmed and candidates across the night
sky.
When you look at just the planets that are confirmed, you can see that in the 1990s,
they were just dribbling and drabbling in.
They really picked up in the 2000s, and it continues to accelerate.
You can see the bump here from Kepler.
Now, of course, 2013 is down, but that's only because it's one third of the year.
If you add in two more thirds, then you're going to get pace still accelerating.
Unfortunately, very recently, Kepler had a problem.
Its second reaction wheel failed, and it needs three reaction wheels and able to point carefully
on the sky.
It has four.
Two of them have failed.
It only has two reaction wheels left, so it can't point as precisely as possible.
Unfortunately, Kepler's primary mission looks like it's at an end.
That doesn't mean its discoveries are at an end, because Kepler has a lot of data that's
already been taken and is down on the ground being processed.
New discoveries will come from Kepler in the coming years, even if Kepler's not taking
more data in its primary mission.
You can see the great growth of discoveries in extrasolar planets.
That's why, even though it's only a couple years later, it's appropriate time to revisit
some of those discoveries I talked about in an earlier episode.
Those discoveries were about, first, the planet star Fomalhaut.
Hubble took the first image of a planet in visible light around the star Fomalhaut.
This isn't a picture of Fomalhaut.
It's actually everything but Fomalhaut, because we're using a technique we call coronography
to block the light of the star.
This white dot indicates where the star would be.
We block the light of the star in order to see the faint material around it.
And Fomalhaut has this dust ring around it, and the geometry of the dust ring indicated
that would be a good place to find a planet.
There might be a planet hiding, and the planet is actually this small little red dot there.
Hubble was able to observe it at this position in 2004.
It had moved to this position in 2006, calculating the orbit.
It was confirmed that this was a planet.
In that same episode, I also talked about the star HR 8799.
And again, we're using coronagraphy to block the light of that star.
And they found three planets around HR 8799.
So actually, it was announced the same day the Fomalhaut announcement was made.
The first day, we had four direct observations of planets.
And then a week later, around the star Beta Pictoris, they released this image.
Now, this is a dust disk around Beta Pictoris, this bright stuff along here.
And in close, they were able to pull out an image of a planet.
But this one was not able to be confirmed at the time.
Well, we've got the updates for you, and we're going to go in reverse order.
So let's start with Beta Pictoris.
The reason why it was unable to be confirmed is because the orbit was in this plane.
And the planet had moved in front of the star.
And if you subtract out the light of the star, you're going to subtract out the light of
the planet.
You can't confirm it.
In 2009, they were able to get this image and recover the planet on the other side of
the star.
Combining the two, you get the full orbit.
And now, the planet around Beta Pictoris has been confirmed.
So I've completed one story that I had to leave hanging at the end of the last episode.
For HR 8799, the update is actually, I'm going to use a sort of a digression, going back
to a planet not discovered with Hubble, but was discovered with this telescope.
This is Herschel's six-foot telescope, with which he discovered the planet Uranus.
Now it's a relatively small telescope, and you would have thought that, well, he was
an amateur astronomer at the time, couldn't the professionals see it?
Well, actually, Uranus is bright enough that the professional astronomers should have seen
it.
So what they did is they went back into the archives, they went back to previous observations
to see if anybody else had seen Uranus.
And they were able to recover it many years earlier, and able to get a much better orbit
prediction for how Uranus was moving.
They could refine Uranus's orbit by looking in previous observations.
As one example, this is John Flamsteed's Atlas Celeste in 1776.
And this is the constellation of Orion, and this is the constellation of Taurus the Bull.
And there's an extra star here.
That star right here is not a star.
That is actually Uranus.
Flamsteed had seen Uranus, and cataloged it as just a minor star in the constellation
of Taurus.
You can tell, because when he republished it in 1795, that star is now labeled Herschel.
Why not Uranus?
Well, we didn't settle on the name Uranus right away.
No one had ever discovered a planet before, so there was no real precedent for how you
name it.
Matter of fact, Herschel suggested that he name it after the current king of England.
And if he'd gotten his way, we would have had a planet named George.
But that's not really the point of this.
The point is that when you discover something new, you can go back to previous observations
and gain more information.
That's what we did with Hubble's observation.
This is Hubble's observation of the star HR 8799.
It's an infrared observation, so it's not a very clean observation.
But somewhere in there should be the three planets.
How are we going to pull them out?
To do that, you've got to understand the idea of a point spread function.
Here is Hubble's observation of the interior part of the globular cluster Omega Centauri.
And you can see there are stars of all sorts of brightnesses here.
The very small, faint ones, up to these very big and bright ones.
And you can also see that the faint ones are roughly small little circles, they're points
of light, whereas as they get brighter, they get bigger.
And they get really bright, they get the diffraction spikes, and they get all the little noise
around it.
This is what we call a point spread function.
Going from a point of light, as you get brighter, how does the image of that object change?
And in order to really understand our Hubble observations, we have to understand the point
spread function very, very carefully.
So, if we go back to this image of HR 8799, and if we understand that point spread function
carefully enough, could we possibly subtract out that point spread function and find the
planets?
Well, the answer is, of course, yes.
Taking the point spread function and subtracting it, we get this image.
And if you look carefully, you can recover the three planets around HR 8799, recovered
from a Hubble image, taken 10 years before these planets were discovered.
So again, like Uranus, we're able to find information from previous observations and
help refine the orbits for these planets.
It gives you an extra 10 years of baseline in the motions of these planets.
Finally, we'll go back to our biggest story, the discovery of the image of Fomalhaut that
we discussed in the previous episode.
Now, even when we announced this in 2008, we knew there was something slightly interesting
about it, because the truth is, Hubble shouldn't have seen it.
Hubble should not have seen the planet Fomalhaut b.
It's not bright enough.
So at the time, we had to assume, make the assumption, that there was a disk of material,
a dust disk of material around the forming planet, and that the light from Fomalhaut
would shine not just on the planet, but also on the dust disk, making it bright enough
for Hubble to see.
Now, of course, something like that is going to be challenged by other astronomers.
We're just not going to accept it.
We're going to go and double check it.
And one of the things is that it will depend a lot upon the orientation of the object.
They told me I needed a prop to show this off.
So we've got this ring here, okay?
And this ring could be tilted towards us, with the top portion towards us, or with the
bottom portion towards us, all right?
And how it's tilted is going to have how the light is going to shine on it, and it's going
to change the brightness of it.
And I can give you an example of that with the planet Saturn.
Here's an image of the planet Saturn, which is front-scattered light.
The light is coming in, hitting it, bouncing back to the camera, all right?
And when the light is front-scattered, it looks like Saturn.
Now, if we put our camera in behind the planet, into an eclipse, so the light was back-scattered
through the rings, you get not this image, but that.
Isn't that quite a difference?
All right?
When the light comes through Saturn's rings, the back-scattering, especially by the very
small particles in Saturn's rings, changes the appearance.
You can see a lot of more faint structure.
And so that was one of the arguments put against the Fomalhaut b planet.
Because the group went in and tried to measure the rotation of the star Fomalhaut, and determined
that it was rotating in one direction.
And they deduced from that, saying, well, if the ring is orbiting in the same direction,
then the orientation would be such that the planet would appear fainter, even if there
were a dust disk around it.
Well, it turns out that that argument is not quite solid.
It's not quite definitive.
A stronger argument actually comes from observations from this telescope, the Spitzer Space Telescope.
Now, Spitzer Space Telescope observes infrared.
And planets are actually brighter in the infrared.
And so the prediction by this group, who was based out of Princeton, was that the light
from the planet should be bright enough in the infrared that Spitzer should see it.
And what did Spitzer see?
Well, it didn't.
Spitzer did not see the planet.
And so the group based out of Princeton said that, well, I think that means that it's not
a planet, but instead a dust cloud.
Why would it be a dust cloud?
Well, here's an example from our own solar system.
There is a asteroid, a tiny, tiny asteroid that has collided with another one and created
a dust cloud.
And you can see how much brighter the dust cloud is than the tiny little asteroid down
here.
So if you have a dust cloud passing through the system, it occupies so much more volume,
can scatter so much more light, that it can be the brightness of what you might see in
the Hubble image.
So their argument was that it's not a planet, but it's really just a dust cloud moving through
the system.
Is that the end of the story?
Nope.
The folks who did the Hubble observations, who incidentally are based out of Berkeley,
went back to do more Hubble observations.
And in 2010, they were able to get this image.
Now, the same instrument was not available.
The instrument that they had used previously had failed and was not able to be repaired
in the service mission four.
But they're able to use another instrument and try and recover the planet.
And so again, we're blocking out the light of the star Fomalhaut.
And somewhere in here, they're trying to find the planet.
And again, it's that tiny little spot, the dot.
There it is.
Okay.
They did it not only once, but twice in 2010 and 2012.
And if we put all four images from Hubble together, you have here it is in 2004, and
in 2006, and in 2010, and 2012.
And from that, you get a much better prediction for its orbit.
Now, we're expecting the orbit to pass along the inside of this ring and help distort the
ring because the ring was off center.
And we're expecting that the interaction between the planet and the ring was causing changes
in the ring.
But that's not what we found.
Instead, the orbit seems to go right across the dust belt.
This orbit, instead of being circular like most of the orbits in our solar system, turns
out to be very elongated.
So this is a computer simulation.
And this whitish region here, this is supposed to be the dust belt looked at face on, okay,
and assuming that it's circular.
And this orange line here would be the orbit of the planet.
It goes way, way out from the star and way, way in close to the star.
Although, I got to say, even at its closest approach, it's still further from Fomalhaut
than Neptune is from our sun.
So although this looks small, it's actually quite large.
It's an enormous orbit.
And it's highly elliptical, highly irregular, and highly unexpected.
The other thing you might notice is that, well, it looks like it's going to cross the
dust belt.
Right?
Well, that's actually just a projection effect.
From our point of view, it could actually cross in front of it or it could cross behind
it.
And in the paper, they calculated that 95% of the possible trajectories actually do not
cross through the dust belt.
So it's unlikely it's going to hit the dust belt.
But if it does, it will reach there around the year 2030.
And at that time, we'll have some really serious evidence as to whether it is a planet or whether
it's a dust cloud.
So how are we going to decide this?
We've got the Berkeley group that says it's a planet with a dust disk around it.
We've got the Princeton group that's saying that it's a dust cloud moving through the
system.
And I've got to tell you, I have alliances with both because I did my PhD at Berkeley.
I did my postdoc at Princeton.
How are we going to tell what is the correct?
I mean, you've got the observations from Hubble versus the observations of Spitzer.
Who is going to be the referee in this battle of the telescopic titans?
Well, the referee, the answer is most likely to come from the James Webb Space Telescope.
JWST will be an infrared telescope, but it's a six and a half meter infrared telescope
that will give it the resolution of Hubble while observing in the wavelengths of Spitzer.
So JWST will have both the resolution and the wavelengths in order to be able to observe
Fomalhaut b in great detail.
And JWST should be able to settle whether Fomalhaut b is a planet or it's just a dust
cloud.
So although we've got a lot of exoplanet discoveries, even as we uncover more, sometimes
we uncover new questions.
And so for now, the Princeton group says strongly that they don't believe it's a planet.
The Berkeley group says strongly they do believe it's a planet.
The answer will have to come in a later episode, perhaps in JWST's universe unfiltered.
I look forward to presenting it to you.
Thanks for joining me.
We'll see you next time.