Building Planets: Cant Make Them, But Hurry
Constructing planets is a delicate business. Trying to get tiny bits of dust to join up into balls has never been found to work. It has to work fast, though, because unless the whole planet clears its dust lane, it will be dragged into the star in short order. It seems you can’t get there from the bottom up, and even if you could, you’d be in trouble. These and other problems with planet-building were discussed this month in two papers in the Annual Review of Earth and Planetary Sciences.
One thing we do know: our solar system is nicely decorated with planets, moons, asteroids, comets, meteors and other objects. And now we realize that planets are also common around other stars. To most planetary scientists, this lends plausibility to the belief that planets form naturally somehow – whether or not our models can figure them out. But the very same flood of discoveries about planets from robotic exploration of the solar system and detection of extrasolar planets has brought with it a host of new problems never anticipated by Laplace, Kant and other early theorists about the origin of planets. Life was so much simpler then.
Let’s start from the dust up. We know many stars possess spinning disks of dust and gas. Can we understand how small bodies might form? There are three problems right off the bat, discussed by Erik Asphaug in “Growth and Evolution of Asteroids” in the Annual Review journal.1 Turbulence, accretion and death spirals render theories of their origins problematic. More on that shortly. It must be remembered throughout the discussion that planetologists assume the Earth, and the solar system, is 4.5 billion years old. That means that any lifetimes numbered in 10s or 100s of millions of years represent tiny fractions of the overall age.
Asphaug [UC Santa Cruz] spent a lot of time discussing asteroid populations and characteristics. He even tossed in a little Tolstoy and Yeats to liven up the dry math and technical jargon. But when it came to the subject matter of building asteroids, he was less sanguine. Asteroids, he said, form from the top down, by destructive processes, in relatively short times:
Asteroid origin is ceaseless, as most asteroids are born in the process of catastrophic disruption. Any main-belt asteroid smaller than a few tens of km is unlikely to have survived intact throughout solar system history. Asteroid evolution is also ongoing: Most small asteroids have been drastically transformed in shape and structure over a few tens of millions of years by subcatastrophic collisions and cratering. Small asteroids are also prone to dynamical perturbations of various kinds, so their present orbits may be quite different from where they originated. Their surface textures and colors are also readily modified on short timescales.
Later, he added, “Asteroids represent the halfway point between the solar system’s turbulent beginnings and the quiescent 4 Ga [4 giga-annum or billion years] that have supported life on Earth.” Don’t look to them as planetary building blocks, in other words. Meteorites, by contrast, are primordial—they are remnants of the birth of our solar system, he said. So how did those smaller chunks form? Watch for any confidence in the following description of the environment of the planetary maternity ward. It’s like trying to get born in a shooting gallery with the heaters and air conditioners going berserk:
The first stage of planetary accretion is among the most complex studies in astrophysics. As they accumulate, planetesimals are entrained within a disk that undergoes violent shocks and propagates gravitational waves and eddies. Magneto-rotational instabilities might lead to high turbulent viscosities, which lead to radial transport and vertical mixing and viscous spreading. Solid particles settle to the mid-plane, increasing the density so that planetesimals might eventually coagulate. Electric discharge and impact heating take place sporadically. Outside this mid-plane the young sun blasts the gas and sweeps away small material. Drag against the gas disk forces meter-sized boulders to spiral in from the planet-forming region.
The chemical environment, too, is grossly out of equilibrium. The disk experiences a wide range of temperatures and pressures and oxidation states, with sharp gradients in time and space. Where the disk is optically thick, it can remain hot for thousands of years; where it is thin, it can cool in days or even minutes. Thermodynamic energy is available from solar heating, shock heating (by impacting globules striking the disk, by disk planetesimals colliding with one another, and by shocks in the gas), compression heating (adiabatic work), and radionuclide decay. This energy is transported radiatively and convectively.
It’s nice that Asphaug acknowledged these environmental hazards, but he so far has not given any reason for hope that bundles of joy will emerge from the wreckage. In fact, the next thing he talked about was how revolting the results of the Stardust mission were to theorists:
One of the most baffling results from any recent space mission is NASA’s Stardust sample return from periodic comet 81P/Wild 2 (Brownlee et al. 2006; see Figure 3), which includes an abundance of refractory silicates, metal sulfides, and refractory oxides. These small grains, captured in aerogel during a flyby through the coma, can have formed only in the terrestrial (inner) part of the disk; how did they make it to the ice-rich region where comets form? It is a revolutionary mission result, telling us that the solar system was exchanging matter across tens of AU during the early stages of primary accretion.
Take a moment from this Revolutionary Etude to focus on that phrase primary accretion: that’s the heart of the matter. In the primary stage of planet building, the tiny dust grains need to accrete, or stick to one another. Only when a planetesimal gets up to kilometer size or more in diameter will its self-gravity pull more material in. Little particles lack the gravitational potential to grow on their own. They’ve got to link up by other mechanisms while in the shooting gallery, with all that turbulence, wind, heating, cooling, colliding and electrical activity tending to disrupt them. Can this work? Let’s jump ahead to Asphaug’s section 3.2.2, “Primary Accretion.” Here’s where the pebble meets the spiral:
Direct knowledge of the earliest stage now exists from astronomical observations (e.g., Meyer et al. 2008). As the hot orbiting mixture of dust and gas cools, larger particles condense, leading to lowered opacity and further cooling and coagulation. The review by Podosek & Cassen (1994) remains highly relevant. Particles orbiting with any inclination or eccentricity must plow through the disk; the resultant damping settles them to the midplane on a timescale of 100 years at 1 AU (Cuzzi & Weidenschilling 2006). What happens next in this central sheet is a great unknown, and depends on the effect of turbulence.
Dust grains coagulate via Brownian motion and chemical or electrical sticking mechanisms (Dominik et al. 2007). This can lead to sand- to boulder-sized agglomerates. Brownian motion is an expression of modest turbulence, possibly set up by magneto-rotational instabilities in the plasma and the dusty gas. However, too great a turbulence disrupts agglomerates faster than they form. Benz (2000) and Leinhardt et al. (2000) studied collisions involving meter- to kilometer-scale aggregates at 1�10 m/s random velocities, and determined their disruption to be a bottleneck to further growth. A possible solution is that compactible aggregates damp the energy of collisions and resist subsequent disruption.
Not only must turbulence be low, but the gas must go away before the growing planetesimals spiral in…. Decoupled solids spiral towards the Sun at an estimated 1 AU [astronomical unit, 93 million miles] per 10�1000 years, so there is not much time!
It’s rare to find exclamation points in scientific journals, so let’s unpack this remarkable description. He doesn’t say how big these early-stage packages are, but they must be very small, because soon he will say, “The problem of accreting meter-scale planetesimals is far from solved.” Whatever they are, these hopeful bundles of dust have very little time to escape the giant sucking sound at the center, where the gravity of the star is pulling in every small pebble very rapidly (1000 years is about 20 millionths the assumed age of the solar system).
His evidence for the alleged “earliest stage” is observational: i.e., the presence of boulder-sized objects inferred (indirectly) from the properties of disks around other stars. We can only have confidence that material represents planet-assembling stages, however, if we can rule out the possibility the material is not, instead, leftover debris from the disruption of existing planets. It could be top-down material, in other words, not bottom-up material growing into planets.
Asphaug failed to describe how dust particles could stick together. The problem of getting meter-size boulders to accrete, let alone kilometer size, is “far from solved,” he said. He assumes it did happen on the basis of melt evidence in meteorites: “But we know it occurred rapidly because of the widespread melting of planetesimals caused by the decay of 26Al.” (Aluminum-26 has a half-life of 740,000 years.)2 Perhaps material puddled up in eddies. Perhaps it coagulated in local swarms. Or, maybe it was due to “random effects” (which, in science, amounts to hand-waving). Whatever happened, “Timing is everything,” he said. “So is location.” Nebular hypothesis or not, planets do not just emerge out of the disk without some very special conditions. From there, Asphaug left primary accretion to wrestle with problems regarding iron-silicate ratios and their relative breakdown rates. It appears now that meteorites, once thought to be pristine objects from the birth of the solar system, actually formed later. These and other problems had to be relegated to some maybes and perhapses. Shhh….don’t ask about chondrules:
The origin of chondrules adds another layer to the mystery, in that these are melt droplets of some kind, quite possibly themselves antecedent to global melting of planetesimals. Chondrules are melted silicate spherules a few mm in size that are abundant in chondritic meteorites (reviewed by Scott 2007). Over half the bulk mass of chondritic meteorites consists of chondrules, so they represent a widespread epoch in planet formation that has been attributed to processes as diverse as lightning, impacts, nebular shocks, compaction heating, and volcanic eruptions. Chondrules cooled through their solidus at rates ranging from 10 to 1000 K/h—much slower than radiative cooling of a mm-sized droplet, which is tens of seconds. Chondrule cooling appears to require a hot background that disappears over a timescale of hours.
What might have happened in the solar system, when planetesimals were presumably progressing from primary to secondary accretion, to form the mysterious chondrules? When in doubt, call Lucky Strike: “Although impact origin of chondrules is not currently in favor, we have a lot to learn, and large, late collisions deserve a closer look.” In logic this is known as special pleading.
One would hope that by late-stage planet formation, when the planetesimals are big enough to pull in their own material, the rest of the story would proceed smoothly. Sorry to pile on the difficulties:
In almost all simulations of late-stage planet formation, giant impacts are treated as “sticky ping pong balls” undergoing perfectly inelastic collisions when they hit, forming a larger equal-mass sphere that conserves linear momentum. This was proven untenable by Agnor et al. (1999), who tracked angular momentum during such a calculation and showed that perfectly inelastic collisions lead to planets with impossibly fast rotation.
Well, that was ten years ago. Surely they have solved it by now? Better simulations by Agnor and others have indeed been performed. But still, half the collisions end up as hit-and-run events that don’t grow planets. Only low-velocity, head-on collisions have any hope of causing more growth than damage. Most of the time, collisions break things down: “The prevalence of hit-and-run collisions makes it a late-stage pathway for the origin of exotic igneous asteroids, for volatile flux and iron-silicate intermingling, and for the bulk removal of planetary mantles and the stripping of iron cores.” The lucky leftovers might have been the planets as we know them. That, at least, is the hope. Maybe a little sweet odor will ease the pain:
The overall trend in a collisional accretionary environment is the loss of atmosphere, ocean, crust, and mantle, the preferential accretion of dense materials into growing planets, the shedding off of mantles, and the occasional disruption of single planets into multiples. This leads to a primary physical and chemical bias, a dichotomy among the accreted and the unaccreted. If finished planets are the loaves of bread, asteroids are the scraps on the floor of the bakery.
Catastrophic disruption continues to be the trend in the solar system today. “The prevalence of young dynamical families confirms that today’s NEOs [near-earth objects] are, by and large, discrete samplings of catastrophic disruption events in the main belt that happened thousands to millions of years ago,” he said, winding down his article. It is “incontrovertible” that the meteors hitting our atmosphere today are “punctuated by recent disruptions. The sampling of meteorites on Earth, and of small asteroids near Earth, reflects that bias.” Most of what we see falling from the sky, in other words, is recent material, not pristine remnants from the solar system’s origins.
We may not understand how planets formed, but we should be glad they did, he said in his conclusion. “Whether Earths are common depends in no small part upon the behavior of accreting planetesimals,” he said, waxing philosophical. In his view, we stand on lucky dust. “Nearly all of the original mass of our main belt was swept up in the chaos of planet formation, so we may be fortunate not to have lost everything from our habitable zone.” Add water (which had to be added later by another lucky strike) and the lucky dust became lucky mud, from which you and I sprang. That’s the tale.
Now, let’s turn to the other paper and see what happens after you have a planet. A whole new class of problems for planet formation theories has come to light in the last decade. John Chambers [Carnegie Institute of Science] discussed those in another article in Annual Review.3 In short, planets don’t stay where they were made, assuming they were made by processes Asphaug just described. Get ready for migration – the conveyor belt that sends planets toward the oven or the freezer:
Gravitational interactions between a planet and its protoplanetary disk change the planet’s orbit, causing the planet to migrate toward or away from its star. Migration rates are poorly constrained for low-mass bodies but reasonably well understood for giant planets. In both cases, significant migration will affect the details and efficiency of planet formation. If the disk is turbulent, density fluctuations will excite orbital eccentricities and cause orbits to undergo a random walk. Both processes are probably detrimental to planet formation. Planets that form early in the lifetime of a disk are likely to be lost, whereas late-forming planets will survive and may undergo little migration. Migration can explain the observed orbits and masses of extrasolar planets if giants form at different times and over a range of distances. Migration can also explain the existence of planets orbiting close to their star and of resonant pairs of planets.
These explanations that models “can” explain are clearly explanations after the fact, not predictions from previous theories. So here you have it: today’s models have to reflect the high likelihood that primary accretion, when the dust is plentiful and conditions are optimum, won’t help anyway, because the early world gets the doom. Migration adds a whole new class of problems that is “probably detrimental to planet formation.” Theorists were not expecting this grief. Chambers said that “for a long time migration received little attention from the planetary science community.” Scientists were “intent on understanding the formation of the Solar System, for which there seems little need to invoke migration…. The discovery of extrasolar planets has changed all this.” Hot Jupiter, there are giant planets spinning around their stars closer in than Mercury to the sun. Such things had never been dreamt of. With only one planetary system to look at (ours), planetologists had convinced themselves that rocky planets form close in, gas giants form far out, and they stay put. No model can account for Jupiter-size gas giants coagulating so close to the star. The only explanation available is that they formed far out, then migrated in. The disk material acts like a spiral conveyor belt – bringing even large planets toward the stellar furnace quickly. Migration is so rapid and efficient, in fact, it’s a wonder any planets escape the fate of falling into the star. Planets are now seen as lucky leftovers from this previously-ignored class of dangers. But then, woe be to any Earth-like planets in a habitable zone with a hot Jupiter nearby. The gravity of the gas giant would be sure to fling the rocky planet out of bounds – if not swallow it whole.
Those interested can obtain John Chambers’ paper for the full depressing story. “In this article, I describe how the standard model of planet formation is being modified to include planetary migration—a transformation that is by no means complete at present….” Later, he said, “the rapid pace of recent developments has left models of planet formation struggling to keep up.” Suffice it to say he was not able to rescue primary accretion from the death clutch of the previous article. Instead, he had to jettison the long-standing core accretion model and consider a radical alternative – disk instability – to get giant planets to form at all, and to form fast enough to get out of the death spiral. Advice to the revolutionaries: what thou doest, do quickly.
1. Erik Asphaug, “Growth and Evolution of Asteroids,” Annual Review of Earth and Planetary Sciences, Vol. 37: 413-448 (Volume publication date May 2009), doi:10.1146/annurev.earth.36.031207.124214.
2. The short lifetime of aluminum-26 and other short-lived radionuclides causes other problems for solar system theories, and leads to radical speculations: see, for instance, early speculations in Science Frontiers and the Smithsonian.
3. John E. Chambers, “Planetary Migration: What Does It Mean for Planet Formation?”, Annual Review of Earth and Planetary Sciences, Vol. 37: 321-344 (Volume publication date May 2009), doi:10.1146/annurev.earth.031208.100122.
Whether planet-building is depressing or fun might depend on your temperament, but it should be obvious that any theory that depends for its acceptance on a long string of lucky accidents falls short of the scientific ideal. Science is supposed to explain things with reference to natural law. It’s supposed to make predictions. It’s supposed to be falsifiable. It’s supposed to have observational support, not reinvent a theory every time new information throws a monkey wrench into the old theory. And when getting to the goal line requires a series of miracles or credibility gaps, it’s hard to judge whether this type of mythmaking improves on Aristotelianism (or Babylonianism, for that matter).
Much of this type of writing appears to be sophisticated storytelling masquerading as scientific explanation. Some of the constraints are well characterized – viscosity, half-lives, the physics of inelastic collisions and incidence angles. But these are like the boundaries and obstacles in a pinball game. Would you be impressed if we proposed a weirdly improbable path the ball took, circling one obstacle, jumping over others, and getting struck by lightning at some point? Would you like it if we assumed the boundaries were flexible and the obstacles movable? Would you be impressed if we said we don’t understand how the ball got from A to B, so we will just pick it up and move it to B for now and leave that problem to someone else in the future? Would you accept our explanation that the ball scored by taking a lucky random walk? How about if we said the game made itself and plays itself? Should such ideas be graced with the word science?
The only observational fact is that Earth hits the jackpot against an improbable odds that only increase with new discoveries. That lends credibility to the belief that a Master Designer not only built the game, but operated the controls.