Was a Fifth Giant Planet Expelled from Our Solar System?

by Ray Sanders on November 10, 2011

Want to stay on top of all the space news? Follow @universetoday on Twitter

Artist’s impression of a fifth giant planet being ejected from the solar system.
Image credit: Southwest Research Institute

Earth’s place in the “Goldilocks” zone of our solar system may be the result of the expulsion of a fifth giant planet from our solar system during its first 600 million years, according to a recent journal publication.

“We have all sorts of clues about the early evolution of the solar system,” said author Dr. David Nesvorny of the Southwest Research Institute. “They come from the analysis of the trans-Neptunian population of small bodies known as the Kuiper Belt, and from the lunar cratering record.”

Nesvorny and his team used the clues they had to build computer simulations of the early solar system and test their theories. What resulted was an early solar system model that has quite a different configuration than today, and a jumbling of planets that may have given Earth the “preferred” spot for life to evolve.


Researchers interpret the clues as evidence that the orbits of Jupiter, Saturn, Uranus and Neptune were affected by a dynamical instability when our solar system was only about half a billion years old. This instability is believed to have helped increase the distance between the giant planets, along with scattering smaller bodies. The scattering of small bodies pushed objects both inward, and outward with some objects ending up in the Kuiper Belt and others impacting the terrestrial planets and the Moon. Jupiter is believed to have scattered objects outward as it moved in towards the sun.

One problem with this interpretation is that slow changes to Jupiter’s orbit would most likely add too much momentum to the orbits of the terrestrial planets. The additional momentum would have possibly caused a collision of Earth with Venus or Mars.

“Colleagues suggested a clever way around this problem,” said Nesvorny. “They proposed that Jupiter’s orbit quickly changed when Jupiter scattered off of Uranus or Neptune during the dynamical instability in the outer solar system.”

Basically if Jupiter’s early migration “jumps,” the orbital coupling between the terrestrial planets and Jupiter is weaker, and less harmful to the inner solar system.

Animation showing the evolution of the planetary system from 20 million years before the ejection to 30 million years after. Five initial planets are shown by red circles, small bodies are in green.
After the fifth planet is ejected, the remaining four planets stabilize after a while, and looks like the outer solar system in the end, with giant planets at 5, 10, 20 and 30 astronomical units.
Click image to view animation. Image Credit: Southwest Research Institute

Nesvorny and his team performed thousands of computer simulations that attempted to model the early solar system in an effort to test the “jumping-Jupiter” theory. Nesvorny found that Jupiter did in fact jump due to gravitational interactions from Uranus or Neptune, but when Jupiter jumped, either Uranus or Neptune were expelled from the solar system. “Something was clearly wrong,” he said.

Based on his early results, Nesvorny added a fifth giant planet, similar to Uranus or Neptune to his simulations. Once he ran the reconfigured simulations, everything fell into place. The simulation showed the fifth planet ejected from the solar system by Jupiter, with four giant planets remaining, and the inner, terrestrial planets untouched.

Nesvorny concluded with, “The possibility that the solar system had more than four giant planets initially, and ejected some, appears to be conceivable in view of the recent discovery of a large number of free-floating planets in interstellar space, indicating the planet ejection process could be a common occurrence.”

If you’d like to read Nesvorny’s full paper, you can access it at: http://arxiv.org/pdf/1109.2949v1

Source: Southwest Research Institute Press Release

About 

In addition to being a published astronomer specializing in variable stars, Ray Sanders has blogged for Universe Today, and The Planetary Society blog, among others. He runs his own blog, Dear Astronomer, teaches classes for CosmoQuest, and is a NASA/JPL Solar System Ambassador.

Anonymous November 10, 2011 at 8:50 PM

I wish the solar system could eject a few people from earth…

Baris Bicer November 10, 2011 at 9:00 PM

I love how that’s your first response to this. You’re too hateful, lighten up. =]

Ray Fowler November 10, 2011 at 9:04 PM

This sounds interesting, but simply could be an example of fixing a flawed model by changing the input data to give a correct result. I was about to click on the pdf to read, but UT has just recommended an article from 2008 for me to read so that’s probably more important!

Torbjörn Larsson November 12, 2011 at 12:11 AM

Nice models are very predictive, and they keep getting better, so how are they “flawed”? You can’t really tell a flaw with a series of related viable theories until you hit a “no go” part, and that is not as easy to establish as we may think. It is, I would argue, much easier to identify a flaw when they can’t be improved.

Also, in this case the initial “input” data is given from other models of planetary disk formation, so there is not much constraint. And many giants can happen, for example HD 10180 has 5 giants.

Tim McDaniel November 10, 2011 at 9:10 PM

I Have Nothing To Add, I Just Wanted To Repeat “Jumpiter”, the felicitous typo of the day.

Ray Sanders November 11, 2011 at 6:35 PM

I was trying to be cheeky.

Ray Sanders November 11, 2011 at 6:35 PM

I was trying to be cheeky.

.. .__. November 10, 2011 at 9:20 PM

Ejected? Or just put into an extremely long orbit?

David Wees November 11, 2011 at 7:12 AM

I was thinking the same thing. Maybe this could be the introduction to a science fiction movie… what would happen when it returns from it’s long orbit?

I suspect though that we’d have other evidence of a very long orbit planet wouldn’t we? Surely some of the other planets would have to have some weirdness to their orbits as a result of interactions when this potential ancient sibling returned?

David Wees November 11, 2011 at 7:12 AM

I was thinking the same thing. Maybe this could be the introduction to a science fiction movie… what would happen when it returns from it’s long orbit?

I suspect though that we’d have other evidence of a very long orbit planet wouldn’t we? Surely some of the other planets would have to have some weirdness to their orbits as a result of interactions when this potential ancient sibling returned?

squidgeny November 11, 2011 at 1:15 PM

A science fiction movie… or a conspiracy theory. There are plenty of nutters waiting for evidence that there really is a big planet out there waiting to swing by and cause all sorts of havoc. Oh no here they come, everyone look busy

squidgeny November 11, 2011 at 1:15 PM

A science fiction movie… or a conspiracy theory. There are plenty of nutters waiting for evidence that there really is a big planet out there waiting to swing by and cause all sorts of havoc. Oh no here they come, everyone look busy

Jerry Jackson November 13, 2011 at 12:20 PM

A Neptune size planet with Sedna’s orbit? Haha that would make a good scifi show.

Kevin Heider November 11, 2011 at 11:51 AM

It would have to have been ejected or else the gas giant planets would NOT have stable orbits. We can see that the gas giants have stable orbits.

Kevin Heider November 11, 2011 at 11:51 AM

It would have to have been ejected or else the gas giant planets would NOT have stable orbits. We can see that the gas giants have stable orbits.

Julian Maytum November 10, 2011 at 9:49 PM

Anyway to get rid of that horrible slide down the bottom right? I know you can “opt out” but can’t we get rid of it completely?

Julian Maytum November 10, 2011 at 9:50 PM

adblock worked :P

Anonymous November 11, 2011 at 9:44 PM

The problem being that everyone using Adblock would mean no ad revenue which would mean no site. Yes, internet advertising does generate money even without clicks.

So the ACTUAL solution is for UT to find ads that aren’t this obnoxious. A good start would be for the opt-out to be anything else than a LIE.

Anonymous November 11, 2011 at 9:44 PM

The problem being that everyone using Adblock would mean no ad revenue which would mean no site. Yes, internet advertising does generate money even without clicks.

So the ACTUAL solution is for UT to find ads that aren’t this obnoxious. A good start would be for the opt-out to be anything else than a LIE.

Anonymous November 11, 2011 at 1:11 AM

This website is much more readable using that (chrome extension) :
http://stylebot.me/styles/842

Joshua Zelinsky November 10, 2011 at 10:27 PM

This doesn’t say much about how common event this is. This might be one detail indicating that getting small planets in the habitable zone can be really tough.

Kevin Heider November 11, 2011 at 11:56 AM

Hot Jupiters migrating inwards has already shown how difficult it can be to get terrestrial planets in a habitable zone.

Torbjörn Larsson November 12, 2011 at 12:01 AM

From a model perspective it can perhaps be difficult. But nature doesn’t seem to agree with the models, as Kepler and Corot finds more small planets than large, and more planets further out than closer in. Hot Jupiters was an effect of observational bias, and they are less common.

So AFAIU we should expect many habitable terrestrial analogs in larger stars (F,G) habitable zone. And that has set modelers scrambling, or at least the Kepler group seem to think they should do so. (I was listening to Marcy’s seminar here in Uppsala this year.)

magnus.nyborg November 14, 2011 at 7:07 PM

Kepler and Corot might provide an observational bias that is on the other end of the scale, favoring planets that are closer to the star wich might overall be planets of less mass. Not saying they surely are, just that they might.

Planets at greater distances from the star is still beyond the confirmation simply because they may not have transited yet, atleast not the 3 times needed for confirmation, and planets further out are less likely to actually transfer (statistically).

Kevin Heider November 11, 2011 at 11:56 AM

Hot Jupiters migrating inwards has already shown how difficult it can be to get terrestrial planets in a habitable zone.

Matt Hickman November 11, 2011 at 1:42 AM

Obviously it was not a planet, giant or not, which was ejected. A planet would clear the neighborhood of its orbit. And this object did not do the clearing.

Ciaran Ruane November 11, 2011 at 11:47 AM

Would that make it a giant dwarf-planet?

Ciaran Ruane November 11, 2011 at 11:47 AM

Would that make it a giant dwarf-planet?

Torbjörn Larsson November 11, 2011 at 11:50 PM

Someone got their planets in a bunch. =D

If you study the model, you will see that the ejected planet had time to clear the disk to become a defined planet. (It takes a few Ma to do this.)

Since the planet definition only belong to our system so far, there is no definitional problem for ejected objects.*

——————-
* Except that they have no official definition at all as of yet what I know of. Meanwhile “free-floating planet” is as good a description as any.

Anonymous November 11, 2011 at 4:15 AM

For this fifth gas giant to have been ejected it means that since its total energy increases the energy of the remaining planets decreases. This means the orbital radii of the major planets, Jupiter and Saturn in particular, should have decreased.

LC

Prof. Michael O. Zimmermann Ph November 11, 2011 at 3:31 PM

Hi, LC, while not being in your class I do not buy the story about a gas giant being ‘expelled’;-). There is only a theory, not any scientific evidence…
In my book the “Rock-Belt” between Mars and Jupiter most likely is what’s left of another planet, not a gas giant. I wonder if there’s any evidence about when the event took place… I like the more ‘down-to-earth’ theories.

Anonymous November 11, 2011 at 5:54 PM

Is the asteroid belt due to a planet that was torn apart or due the failure of these materials to form a planet because of Jupiter’s gravitational or tidal influence?

The appearance of gas giants at large radial orbit is probably due to the gravitational interaction with material that was sent in towards the sun. Maybe a Neptunian mass planet was ejected in this process. The problem is that it is tough to back out the initial conditions of a many body system due to nonlinear processes or chaos which occurs.

LC

Anonymous November 11, 2011 at 5:54 PM

Is the asteroid belt due to a planet that was torn apart or due the failure of these materials to form a planet because of Jupiter’s gravitational or tidal influence?

The appearance of gas giants at large radial orbit is probably due to the gravitational interaction with material that was sent in towards the sun. Maybe a Neptunian mass planet was ejected in this process. The problem is that it is tough to back out the initial conditions of a many body system due to nonlinear processes or chaos which occurs.

LC

Torbjörn Larsson November 11, 2011 at 11:43 PM

I have to ask the same as lcrowell, which variant do you mean?

The theory is validated in spades, see my reply to Roger Overcash. A validated theory is worth a thousand facts, as it predicts many interconnected such correctly.

Most times, as for gravitation and evolution, the process is an observed fact. These type of planet scatter models seems likely to become the same.

Prof. Michael O. Zimmermann Ph November 11, 2011 at 3:31 PM

Hi, LC, while not being in your class I do not buy the story about a gas giant being ‘expelled’;-). There is only a theory, not any scientific evidence…
In my book the “Rock-Belt” between Mars and Jupiter most likely is what’s left of another planet, not a gas giant. I wonder if there’s any evidence about when the event took place… I like the more ‘down-to-earth’ theories.

Roger Overcash November 11, 2011 at 9:39 AM

“Researchers interpret the clues as evidence…”

“Nesvorny and his team used the clues they had to build computer simulations…”

The orbit of mini asteroid 2005 YU 55 was precisely determined as it passed earth recently. The orbit has been calculated up to about one hundred years with decent accuracy, after that, it’s anybody’s guess. The reason for this is when it passes Venus latter on, a very small distance closer or farther away from that planet will make a very large difference one hundred plus years down the road. Now imagine calculating the position for it four billions years in the past.

Though computer simulations are useful to determine “possible outcomes” for events out of hand for direct measurements i.e., an event four billion years in the past, it is still just possible outcomes, literally thousands of them. The “butterfly effect” cannot be ignored.
The programmed input is known data and also assumptions. Runs of animations are tweaked (more of this, less of that) while in progress to get a desired conclusion. The outcome is in no way evidence but a possibility out of many many. I will consider it.

Kevin Heider November 11, 2011 at 11:54 AM

Actually we can reliably predict the positions of the planet for about 60 million years into the future and past. Since NEO are so dynamically weak it is difficult to predict their positions for more than 200 years.

Roger Overcash November 11, 2011 at 8:34 PM

Thank you for these constraints, Kevin.
For the sake of highlighting my intended point, I used 2005 YU 55 as an extreme example to easily show where I was going with this. I did not give an upper limit on accurate predictions of planetary orbits but instead used the four billion year simulated runs to show implausibility. As you have said, the upper limit of accurate predictions of planetary orbits is about 60 million years. This supports my point. My point is that a four billion year simulated run is so full of unknown dynamics that many outcomes are available to choose from to support a hypothesis especially if the runs are tweaked. In some cases (not here) these outcomes are established as proof. That is ludicrous.
Astronomers not so long ago discovered about a dozen “free floating” Jupiter sized planets about 20,000 lt yrs towards the center of our galaxy using a microlensing technique. Recent direct observations have shown some “free floaters” in nursery clouds.
Might these be a lower limit on failed star attempts? Ejected planets also seem likely for the crowded dynamics of a stellar nursery or the crowded neighborhood towards the center of our galaxy. Perhaps both scenarios are true.
If there are any giants out there, it is the bloggers of Universe Today whose blogs are far more intuitive, imaginative, and scientific than the articles blogged about. I feel honored just to read some of them.

Torbjörn Larsson November 11, 2011 at 11:39 PM

Sure there are problems, but no worse than simulating a dice throw to convince that it is possible to get one particular face up.

The generic Nice model is very predictive, it is a long list of observations it gets correct: terrestrials, outer vs inner asteroids, giants, Trojans, captured moons, Kuiper objects, scattered disk, Oort cloud and impactors (LHB). Recently predicting the rapid formation of a small Mars have been added, and here the formation of Neptune and Uranus is predicted.

Kevin Heider November 11, 2011 at 11:46 PM

The reason we can only predict the motion of the inner planets for ~60 million years is largely because of the perturbations of Mars by the asteroids in the asteroid belt. The perturbations do add up after millions of years. Once we accurately know the masses of the largest ~30 asteroids(?) the simulations should be able to be expanded much deeper.

Roger Overcash November 11, 2011 at 8:34 PM

Thank you for these constraints, Kevin.
For the sake of highlighting my intended point, I used 2005 YU 55 as an extreme example to easily show where I was going with this. I did not give an upper limit on accurate predictions of planetary orbits but instead used the four billion year simulated runs to show implausibility. As you have said, the upper limit of accurate predictions of planetary orbits is about 60 million years. This supports my point. My point is that a four billion year simulated run is so full of unknown dynamics that many outcomes are available to choose from to support a hypothesis especially if the runs are tweaked. In some cases (not here) these outcomes are established as proof. That is ludicrous.
Astronomers not so long ago discovered about a dozen “free floating” Jupiter sized planets about 20,000 lt yrs towards the center of our galaxy using a microlensing technique. Recent direct observations have shown some “free floaters” in nursery clouds.
Might these be a lower limit on failed star attempts? Ejected planets also seem likely for the crowded dynamics of a stellar nursery or the crowded neighborhood towards the center of our galaxy. Perhaps both scenarios are true.
If there are any giants out there, it is the bloggers of Universe Today whose blogs are far more intuitive, imaginative, and scientific than the articles blogged about. I feel honored just to read some of them.

Kevin Heider November 11, 2011 at 11:54 AM

Actually we can reliably predict the positions of the planet for about 60 million years into the future and past. Since NEO are so dynamically weak it is difficult to predict their positions for more than 200 years.

Roger Overcash November 11, 2011 at 9:39 AM

“Researchers interpret the clues as evidence…”

“Nesvorny and his team used the clues they had to build computer simulations…”

The orbit of mini asteroid 2005 YU 55 was precisely determined as it passed earth recently. The orbit has been calculated up to about one hundred years with decent accuracy, after that, it’s anybody’s guess. The reason for this is when it passes Venus latter on, a very small distance closer or farther away from that planet will make a very large difference one hundred plus years down the road. Now imagine calculating the position for it four billions years in the past.

Though computer simulations are useful to determine “possible outcomes” for events out of hand for direct measurements i.e., an event four billion years in the past, it is still just possible outcomes, literally thousands of them. The “butterfly effect” cannot be ignored.
The programmed input is known data and also assumptions. Runs of animations are tweaked (more of this, less of that) while in progress to get a desired conclusion. The outcome is in no way evidence but a possibility out of many many. I will consider it.

ikma November 11, 2011 at 2:36 PM

in the second to last paragraph, Nesvorney attempts to support his theory by saying it “…appears to be conceivable in view of the recent discovery of a large number of free-floating planets in interstellar space”.

what is the story with these ‘recently discovered interstellar planets’? is this a theory based on orbital mechanics that predicts planet ejection is common, or did we actually spot some interstellar planets?

Torbjörn Larsson November 11, 2011 at 11:30 PM

Both, see my reply to scott chambers.

ikma November 11, 2011 at 2:36 PM

in the second to last paragraph, Nesvorney attempts to support his theory by saying it “…appears to be conceivable in view of the recent discovery of a large number of free-floating planets in interstellar space”.

what is the story with these ‘recently discovered interstellar planets’? is this a theory based on orbital mechanics that predicts planet ejection is common, or did we actually spot some interstellar planets?

Prof. Michael O. Zimmermann Ph November 11, 2011 at 2:47 PM

Absolutely! First politicians, the cause of greed and wars… and then get rid of the copy-cats…. haha

Prof. Michael O. Zimmermann Ph November 11, 2011 at 2:47 PM

Absolutely! First politicians, the cause of greed and wars… and then get rid of the copy-cats…. haha

scott chambers November 11, 2011 at 5:01 PM

I wasn’t aware that we had discovered ” a large number of free-floating planets in interstellar space”. when did I miss that?

HeadAroundU November 11, 2011 at 8:22 PM
HeadAroundU November 11, 2011 at 8:22 PM
Torbjörn Larsson November 11, 2011 at 11:29 PM

First there is the simulations that we see an example of here. Many of those ejects planets, so it was a longstanding prediction, I believe.

Then came direct observations that HeadAroundU links to. Personally I don’t think the statistical sample all that convincing as of yet, but it accords with the above prediction.

scott chambers November 11, 2011 at 5:01 PM

I wasn’t aware that we had discovered ” a large number of free-floating planets in interstellar space”. when did I miss that?

Torbjörn Larsson November 11, 2011 at 11:26 PM

The strength of Nice type scattering models is evident I think. First their overall predictivity, and now the elaborations that makes first Mars small and early enough (a result earlier this year), and here Uranus and Neptune originating sufficiently close to the Sun.

Especially nice to see our system eject one giant as most systems should do to meet the current tentative constraints on free giants. Every system is an individual, but our system looks more and more like sharing characteristics with the large majority. Which is encouraging from the aspect of astrobiology.

Marcin November 14, 2011 at 1:56 PM

How fast was this potential planet ejected? How far would it be today? Is it possible to detect it using todays telescopes? As far as I know Jupiter emits some heat so maybe it’s possible to use infrared telescopes to find wandering large gasous planet.

Torbjörn Larsson November 14, 2011 at 6:23 PM

Good question!

For a quick estimate we can assume it was ejected at a speed comparable with orbital speed. At Jupiter radius of today that is ~ 10 km/s.

If it was ejected at the time of Late Heavy Bombardment it happened ~ 0.7 Ga after our system formed or ~ 4.2 Ga bp.

The ejected planet would have traveled ~ 4*10^9*60*60*24*365*10*10^3 ~ 10^21 m. A light-year is v*t ~ 3*10^8*60*60*24*365 ~ 10^16 m. Hence it would have traveled ~ 100 000 ly.

As a comparison our sun is ~ 26 000 ly out from the Milky Way (MW) center. Or about halfway, since the MW is ~ 100 000 ly in diameter. The “5th giant” would hence have traveled ~ 10^5/(2*π*26*10^3) ~ 0.5 orbits around MW relative to us. (Roughly, assuming it was unlikely to be ejected on a direct radius out from the MW center.)

Meanwhile, at our systems clip at ~ 200 km/s relative to the MW center, our system would have traveled ~ 4*10^9*60*60*24*365*200*10^3 ~ 3*10^22 m ~ 3*10^6 ly or ~ 16 orbits around the MW. [We have clocked 20; only 30 more to go!]

Unfortunately the giant is dispersed among all other “ejects of the systems” and likely rather obscured by being roughly halfway around the MW, on the other side of the galactic center.

We could wait another 4 Ga and have a better position for observation, of course. :-/

Comments on this entry are closed.

Previous post:

Next post: