Fasten your seatbelts, folks — I'm about to take you on a wild ride!

NASA's solar system

As portrayed on the home page for the NASA/JPL Photojournal, our solar system is an orderly arrangement of planets orbiting the Sun.


Compared with the systems of planets being found around other stars, our solar system is an orderly place, with each planet tracking around the Sun in a stable, roughly circular orbit. For centuries, the planets' long-term stability has been taken as evidence that they formed where they are now, sucking up gas, dust, and larger building blocks from the protoplanetary disk around them until reaching their final sizes.

But dig a little deeper, and you find serious problems with that simplistic view. For example, Uranus and Neptune should have ended up much smaller and less massive, because billions of miles from the infant Sun the protoplanetary pickings were slim and the assembly process too slow. Conversely, Mars formed in the fat of the disk and should have ended up at least 10 times more massive than it is today. And no one really understands the asteroid belt's existence — particularly why it's crudely divided into rocky bodies (called S types) nearer the Sun and dark, carbon-dominated hunks (C types) farther out.

Dynamicists solved the Uranus-Neptune dilemma several years ago by positing that the four giant planets were initially a much closer-knit family, coming together in a cozy zone 5 to 12 astronomical units from the Sun.

Evolution of outer planets' orbits

Computer models suggest that the outer planets formed within a narrow range of heliocentric distances, 5 to 12 a.u. from the Sun (vertical scale, lower left). After about 2 million years, however, the orbit of Saturn entered a 5:3 orbital resonance with Jupiter and became more eccentric. Neptune, which formed closer in than Uranus did, has repeated close encounters with all three of the other planets and is eventually ejected outward to its present location.

A. Morbidelli & others / Astronomical Journal

The Big Four coexisted peacefully at first, but after a couple of million years things got ugly. Jupiter's gravity jostled Saturn into an unstable, wide-swinging orbit, triggering a chain reaction of close encounters that ultimately threw Neptune and Uranus out to the distant depths of interplanetary space they now occupy.

Theorists now have computer models that get the outer solar system to come out right, more or less, but they're still vexed by the inner planets. The thorny problems of a too-small Mars and a compositionally stratified asteroid belt remain.

Worse, discoveries of other solar systems were revealing radically different inner-planet architectures: "hot Jupiters" whirling so close to their suns that a year for them is just days long, and massive planets in orbits so wildly out of round that any lesser worlds they encountered would have been tossed out. Given all the disorder so common among the exoplanets, it's remarkable that the Sun ended up with any small, close-in worlds at all.

But there's been a breakthrough in modeling our solar system's formation, details of which emerged at last week's meeting of the American Astronomical Society's Division for Planetary Sciences. It turns out that getting four right-size terrestrial planets and the right kind of asteroid belt is a snap — but it requires dramatic new thinking about the path Jupiter (and Saturn) took getting to their current locations.

Pulsar planet

Nobody really knows what the planet system orbiting pulsar B1257+12 would look like if viewed up close, but here's a plausible depiction. Three terrestrial-mass planets orbit the pulsar at close range; an asteroid-mass object (not pictured here) orbits much farther out. Click on the image for a larger view.

Solving for Mars

The stage for this revolution was actually set last year, when Brad Hansen (University of California, Los Angeles) tried assembling the inner solar system an entirely new way. He took a cue from the one other place known to have close-in, Earth-size planets: the system surrounding the millisecond pulsar B1257+12. Discovered in 1991, these pulsar planets are often overlooked because their host "star" is so extreme.

Prior computer simulations assumed that the inner planets accreted from a dense, massive belt of mile-wide planetesimals extending almost out to Jupiter. But invariably the outcome was a too-massive Mars and jumbled mess in the asteroid belt. However, Hansen realized that PSR B1257+12's planets must have assembled from a limited disk of hot material closely surrounding the pulsar.

Modeling the inner planets

By assembling the terrestrial planets from a narrowly confined disk (gray band), simulations yield a distribution of inner planets (open circles from 23 computer runs) that closely matches the actual arrangement (colored dots). Mercury's upper value assumes the planet formed with the same iron abundance as other terrestrial planets'.

Brad Hansen / Astrophysical Journal

When he tried that approach with our solar system, starting with a disk confined to just 0.7 to 1.0 astronomical unit from the Sun, voilà! — his computer runs routinely coughed up sets of planets with bigger ones (think "Earth" and "Venus") in the middle and smaller ones ("Mercury" and "Mars") near the inner and outer edges.

So why should Earth and its immediate neighbors have formed from such a limited disk? Hansen had no clue when he published his results last year. "In my paper I freely admit the choice was ad hoc," he allows. But it worked — far better, in fact, than any of the previous trials.

Meanwhile, the outer-planet crowd had wondered how Jupiter managed to avoid becoming a close-in captive of the Sun, as so many other beefy exoplanets had. On paper, tidal interactions between the King of Planets and the Sun's protoplanetary disk should have drawn Jupiter inward to its doom, or nearly so.

As early as 1999, however, theorists Frederic Masset and Mark Snellgrove (then at Queen Mary College) showed that Jupiter would have indeed migrated inward — but only until it linked up with Saturn in a 3:2 resonance, that is, with the two spaced such that Jupiter completed three orbits for every two of Saturn's. At that point the pair would have reversed direction and headed outward. (The mechanics of this coupled migration are a little involved; if interested, you can get the details here.)

Hansen's shot-in-the-dark simulations, combined with the realization that the gas giants could have migrated both inward and outward, gave solar-system modelers a "Eureka!" moment. What would have happened, they wondered, if young Jupiter had ventured much closer to the Sun than where it finally ended up?

Jupiter and Saturn on the move

This simplified sequence shows how the in-and-out migration of Jupiter and Saturn early in solar-system history could have created a truncated disk of material from which the inner planets formed. Their movement also created overlapping zones of rocky (S) and carbonaceous (C) bodies in the asteroid belt.

Kevin Walsh

The amazing answers came to light at last week's meeting. Kevin Walsh, who'd worked this problem with Alessandro Morbidelli while post-docing at Côte d'Azur Observatory in France, ran computer simulations that put Jupiter initially 3½ a.u. from the Sun and allowed it to creep inward to 1½ a.u. (about where Mars orbits now). The results were remarkable in their breadth and significance.

First, Jupiter's gravity would have forced the small stuff in its path inward too, creating a perturbation-driven snowplow that piled all the rocky planetesimals into a mini-disk with an outer edge 1 a.u. from the Sun. According to presenter David O'Brien (Planetary Science Institute), a member of Walsh's team, Jupiter took only 100,000 years to drive inward to 1½ a.u.and another 500,000 years to reach its current orbit, 5.2 a.u. from the Sun.

Second, the new computer runs confirmed what Hansen had already shown: a mini-disk of rocky material extending only to 1 a.u. provided just what's needed to assemble four terrestrial planets — and a Mars that's not too big.

At the meeting, David Minton and Hal Levison (Southwest Research Institute) described their own simulations using a truncated mini-disk, and they come to much the same conclusions. One key variation is that, in the Minton-Levison runs, Mars forms well within the disk and migrates to its outer edge and beyond.

This could be a good thing, because a moving Mars would provide the gravitational perturbations needed to kick iron-rich planetesimals out of the disk and into the inner asteroid belt, where they're commonly found today. "The original locations of Mars in the [disks] I calculated were quite variable," Hansen comments. "The outward migration was driven by scattering, so things shake up quite a bit."

Third, Jupiter probably would likely have come in even closer, perhaps sliding all the way into the Sun, had not Saturn (already in tow via the 3:2 resonance) grown massive enough to hit the tidal brakes and reverse both planets' movement. In this sense, the formation and survival of the terrestrial planets hinged not on Jupiter's existence but on Saturn's.

Fourth, Jupiter's inward trek would have completely swept clear the asteroidal region from 2 to 4 a.u. Most of the objects there were lost completely, but roughly 15% ended up scattered into a disk beyond Saturn. After reversing course and moving outward, the two planets scattered some of those previously displaced objects again, this time inward, returning them to what's now the inner asteroid belt.

Fifth, as Saturn and Jupiter continued outward to their final orbits, they encountered another group of asteroids. Unlike the rocky bodies that had boomeranged out and back, these were carbon- and water-rich objects that had formed 6 to 9 a.u. from the Sun. Tossed inward by perturbations from the dynamic duo, they formed most of what's now the outer asteroid belt.

Sun and forming planets

If recent computer simulations are correct, the Sun's innermost planets assembled from a narrow disk of rocky rubble that was only about 30 million miles wide. Such a tightly confined disk yields a Mars that's not too big — a failing of previous modeling.

NASA / JPL / T. Pyle (SSC)

A New Paradigm?

To recap: in one sweeping narrative, these theorists propose solutions for both a minimalist Mars and a stratified asteroid belt with a rock-rich inner region and a carbonaceous, water-harboring outer belt. As a bonus, the new mindset leads naturally to a set of four inner planets (correct sizes, correct orbits) that assembled on the right time scale (within about 30 million years of the Sun's formation). It even provides a source of water for Earth (C-type asteroids) and a near-Earth environment conducive to the presumed giant impact that formed the Moon.

This radical scenario represents "a paradigm shift in our understanding of the evolution of the inner solar system," says Walsh. That's an understatement! It all seems hauntingly Velikovskian to me, except that these folks have clearly done their homework.

Will "Jupiter's Grand Tack" (as Morbidelli dubs it) hold up to further scrutiny? Walsh and his team have submitted a fuller treatment to Nature for publication, but other dynamicists are already weighing in based on the presentations heard last week. "Many aspects of their model look good to me," observes SwRI officemate William Bottke, "but lots of first-order things have to be tested before they can declare victory on all fronts."

For example, it's now widely accepted that most of Earth's water was imported from the outer asteroid belt. Yet Bottke thinks the scenario envisioned by Walsh, Morbidelli, O'Brien, and others would require a vast reservoir of water-rich (C-type) bodies, totaling hundreds of times the mass of the current asteroid belt. "We need to vet these models with more physics and more cosmochemistry," he says. Also, the depth of Jupiter and Saturn's inward penetration would have depended critically on how fast Saturn grew to nearly full size and when. The broader the range of initial conditions that "work," the more confidence there'll be that this scenario is the right one.

Morbidelli remains confident that they're onto something profound. "We consider ourselves celestial geologists," he quips. "We're now able to 'read' the current solar-system arrangement well enough to figure out what the early planets did."


Image of MsColleen


October 16, 2010 at 10:52 pm

Once again, a question that seemed to be beyond possibility of any answer short of divine intervention ultimately falls into place with a very elegant and satisfying scientific explanation. I will never claim to "have faith" in science, but I am again validated in placing trust in the scientific method.

Which is a good thing, considering that I lost my "faith" in other explanations years ago.

You must be logged in to post a comment.

Image of chris.j.


October 17, 2010 at 8:26 am

i'm very curious to see how this model handles the LHB and correlates with the nice model.

You must be logged in to post a comment.

Image of Pete


October 17, 2010 at 11:26 am

To bad people don't have a sense of humor anymore. This along with a million other the pink elephant we see now. Twenty years from now....It will be some other alcohol (explanation) we all can get drunk on. This article IS the flavor of the month.

You must be logged in to post a comment.

Image of Dez


October 17, 2010 at 1:39 pm

Science is a work in progress, no one ever claims to know all the answers, we study the evidence. As our knowlege and instrumentation become more sophisticated we are sometimes forced to form new theories and explanations.We are constantly learning. That is why the scientific method works, unlike "You better believe this or you'll go to hell!"

You must be logged in to post a comment.

Image of KPW


October 18, 2010 at 8:07 am

The time scale for this process (as described above) is under a million years. The end point would have to be the starting point for the Nice model. This might put some additional constraints on that and the LHB. There's going to be a lot more modeling before we have the whole picture. (And it might turn out to look quite different from this.)

You must be logged in to post a comment.

Image of Eric V.

Eric V.

October 18, 2010 at 11:52 am

The Galileo spacecraft's drop probe contained a mass spectrometer that recorded only 2000 ppm of water in the atmosphere of Jupiter if hydrogen there was taken to define 1 million ppm. As I recall there was discussion that this seemed to be too dry for what we thought Jup. should have, given that its orbit is at the distance from the Sun where it should've sopped up lots of water. Perhaps this formation theory explains that by saying that Jup formed inside that "snowline" distance.

Then, could the Galilean moons have been captured from the "scattered planetesimals" region on Jup's outer trip?

You must be logged in to post a comment.

Image of Laurel Kornfeld

Laurel Kornfeld

October 18, 2010 at 5:16 pm

Given that pulsars are the remains of dead massive stars, is it assumed the planets orbiting B1257+12 formed after the star died and became a pulsar? Or could the planets have orbited the original star and survived its demise? If the planets formed only after the death of the star, is this solar system really comparable to one like ours?

You must be logged in to post a comment.

Image of Dieter Kreuer

Dieter Kreuer

October 19, 2010 at 5:51 am

Since pulsars form during supernovae explosions that happen only to stars above 8 solar masses, while the pulsar itself has an upper mass of only 3 solar masses (else, it would have become a black hole), this implies that the progenitor star of the pulsar must have lost more than half of its mass during the explosion. Suppose, you have panets circulating the progenitor star at circular orbit speed (circular orbit for simplicity), which is vo_progenitor = sqrt (G*M_progenitor/r), with G the gravitational constant and r the orbital distance. If you now replace the mass of the progenitor with the pulsar's mass, which is < 1/2 M_progenitor with the above, you will find that vo_progenitor exceeds the escape velocity of the final pulsar, which is ve_pulsar = sqrt (2*G*M_pulsar/r) < sqrt (2*G*1/2*M_progenitor) = vo_progenitor. That means, any planets that have survived the supernova explosion should have escaped the lower mass pulsar, so any planets found there now must have formed in some accretion disk that resulted from the supernova blastoff.

You must be logged in to post a comment.

Image of Ryan


October 20, 2010 at 6:52 am

lol Pluto

You must be logged in to post a comment.

Image of Richard Carroll

Richard Carroll

October 21, 2010 at 1:38 pm

It seems that these ideas, if true, say something about the likelihood of habitable, stable planets forming in other star systems. Has anyone worked through that, particularly considering the wide range of stellar luminosities?

You must be logged in to post a comment.

Image of Mike Haberer

Mike Haberer

October 22, 2010 at 10:13 pm

Interesting, but as current theory defines the formation of the earth's moon resulting from a Mars size planet hitting Earth, resulting in our current planet with a dense Nickel-Iron core and moon with a much less dense core, any theory of inner planet formation needs to provide for a fifth planet being formed, not just the current four. That's the inherent challenge with models, you can create one that results in the end your looking for, but to ignore a Mars sized fifth planet seems to me to be a big omission.

You must be logged in to post a comment.

Image of Rod


October 23, 2010 at 10:00 am

After reading the report, I will take this new model with serious caution. The present inventory of some 494 exoplanets challenges the earlier views of how the solar system formed here. Currently nearly 100 exoplanets previously published have been disputed or retracted. Does the new computer model start with dust grains and show how planets form from dust grains (without the dust disk dissipation problem)? My answer no, the computer simulations always begin at some *later stage of development* which determines the output. If people wish to believe that our solar system and habitable earth formed from random dust grain collisions they are free to do so. However, don't use preset conditions in a computer model simulation as science to justify such a faith position. The current crop of some 494 exoplanets and retracted data challenges the computer models and the preset conditions being used for origins science.

You must be logged in to post a comment.

Image of Edward Schaefer

Edward Schaefer

October 28, 2010 at 8:31 am

Let's see. Jupiter migrates inwards and somehow ends up in a 3:2 resonance with Saturn. That resonance somehow stops Jupiter's inwards slide towards the Sun and even works to draw both Jupiter and Saturn back outwards. Then having "towed" Jupiter outwards with the aid of the 3:2 resonance, Saturn is by some magic released from that resonance and later goes into a 5:3 resonance which triggers the migrations of Uranus and Neptune into their current orbits. This also moves Saturn through the 2:1 resonance which previously was thought to trigger the migrations of Uranus and Neptune.

Let's just say that I find this whole business to be very, very strange. The Nice Model for how the outer Solar System got to be as it is I found to be fairly elegant and reasonable. OTOH, this one raises more questions than answers. For one, I do not understand why the inner asteroid belt objects did not get mixed in with the outer asteroid belt objects as the former were scattered by Jupiter. Also, an early migration of the outer planets removes that migration as an explanation of the LHB.

On the other hand, it is nice to see progress being made in understanding how the inner solar system formed. Still, a model that only produces the four inner planets neglects the formation of the Mars-sized object whose collision with the Earth is believed to have resulted in the formation of the Moon.

So let's just say that I agree that modeling the formation and history of our Solar System remains a work in progress.

You must be logged in to post a comment.

Image of Sreedhar Rao Sonti

Sreedhar Rao Sonti

November 5, 2010 at 7:25 am

Almost all exoplanets, so far discovered, are Jupiter size living dangerously close to their parent star, which would go undetected if those were smaller in size.I think our solar system also had big size planets close to the sun with their outer layers losely packed, in the process of their evolution,which were, however, blown out by strong winds from the primordial sun along with the gasses outwards. Thus, the outer planets grew in size capturing the material blown out and those close remained rocky planets losing lot of material. If we consider this process as plausible we can read out the present solar system arrangement well enough.

You must be logged in to post a comment.

Image of Sreedhar Rao Sonti

Sreedhar Rao Sonti

November 5, 2010 at 7:42 am

Almost all exoplanets, so far discovered, are Jupiter size living dangerously close to their parent star, which would go undetected if those were smaller in size.I think our solar system also had big size planets close to the sun with their outer layers loosely packed, in the process of their evolution,which were, however, blown out by strong winds from the primordial sun along with the gasses outwards. Thus, the outer planets grew in size capturing the material blown out and those close remained rocky planets losing lot of material. If we consider this process as plausible we can read out the present solar system arrangement well enough.

You must be logged in to post a comment.

Image of J.Lawless


July 19, 2011 at 7:04 am

Does this mini capsule, as she wrote speaking about the micrometer zircon found in western Australia, be compared to this thought by Brandon Tingley in the article following Emily's, :"Even the best(scientist)will sometimes overlook,or too readily brush over, a small but crucial detail "? . Trying to put together all the "small details" brushed over by researchers' teams would, perhaps, give berth to the genuine theory ? In writing that, do I make Science Fiction ?

You must be logged in to post a comment.

You must be logged in to post a comment.