OSIRIS on Rosetta has imaged Comet 67P/Churyumov-Gerasimenko!

As a member of the OSIRIS Science Team I am happy to announce that our camera OSIRIS, that flies on ESA’s spacecraft Rosetta, now has imaged the target of its ten year long journey – Comet 67P/Churyumov-Gerasimenko!


Comet 67P/Churyumov-Gerasimenko in constellation Ophiuchus. This image taken with the Wide Angle Camera on March 20 shows a wide field 25 times larger than the diameter of the full moon. The color composite shows a background of hydrogen gas and dust clouds in the constellation Ophiuchus. The white box indicates the position of the close-up taken with the Narrow Angle Camera (below). The images were taken from about 0.03 AU distance to the comet. Rosetta was at a distance of approx. 4.4 AU from Earth. Image credit: ESA ©2014 MPS for OSIRIS Team MPS/UPD/LAM/IAA/SSO/INTA/UPM/DASP/IDA

Rosetta was launched in March 2004. The purpose of the spacecraft is to explore, in situ, what happens to a comet nucleus when it approaches the Sun from a very large distance, gradually is heated and therefore becomes active. Therefore, Rosetta first had to get very far out in the Solar System. The spacecraft swung by Earth three times, and Mars on one occasion, so that the gravitational perturbations from these planets gradually could make Rosetta’s orbit around the Sun wider. On its way, the spacecraft also passed near to two asteroids – (2867) Steins in September 2008 and (21) Lutetia in July 2010. I June the following year, Rosetta had come so far from the Sun that its solar panels no longer managed to generate the electric power necessary to keep the entire spacecraft up and running. Therefore, Rosetta was put in hibernation and all available power was used to heat the instruments to prevent them from break by freezing. The ground control had no contact with Rosetta at all.

In October 2012 Rosetta was farthest from the Sun, no less than 5.3 AU (one astronomical unit, 1 AU, is the mean distance between Sun and Earth, and corresponds to 150 million kilometers). It means that Rosetta was beyond the orbit of Jupiter, that is located 5.2 AU from the Sun. Two and a half years after Rosetta entered hibernation, on January 20, 2014 to be precise, it was time for the spacecraft to wake up. It was an enormous relief when the signals from Rosetta reached the ground control! After the wake-up, careful checks were made to make sure Rosetta was feeling well after its long sleep. We are now at a stage where the scientific instruments are switched on one by one, to see how they have coped with the hibernation. OSIRIS was switched on last week, and has now taken its first images of the comet – the camera works beautifully! We will therefore be ready when Rosetta reaches the comet in August this year, at a distance of about 4.5 AU from the Sun.


OSIRIS is the camera system on Rosetta. It actually consists of two different telescopes. One of them is called the Wide Angle Camera (WAC) and has a rather large field of view since it will be used to image the comet coma, the cloud of gas and dust that the comet nucleus surrounds itself with (see a previous post on comets). The camera has 14 different filters – glass plates with a special composition and surface coating that makes them transparent to light only at specific wavelengths. These filters are manufactured in Sweden and is the Swedish hardware contribution to OSIRIS. Seven of these filters are so-called narrowband filters – they are transparent only at very strict wavelength regions corresponding to the wavelength were seven different molecular fragments (radicals) emit light when they are illuminated by the Sun. These radicals are CS (a compound consisting of carbon and sulphur), the hydroxyle radical OH and the oxygen atom O (these are formed when the ultraviolet light of the Sun break down water molecules), NH and NH2 (compounds of nitrogen and hydrogen), CN (the cyano radical, consisting of carbon and nitrogen), and the sodium atom (Na), that can be outgassed by dust grains that are strongly heated by sunlight.

The dust grains in the comet coma will reflect sunlight, and some of this light will find its way through the narrowband filters. This is not good, since we will use the intensity of the light to calculate the abundances of radicals and atoms in the coma. Since the dust grains contribute with light, that does not originate from within the gas at all, the risk is that we overestimate the abundance of gas. Therefore, the WAC also has seven filters that is transparent to light just next to the wavelength regions of the narrowband filters. In this way, the contribution of the dust grains to the measured light can be estimated, and compensated for when determining the gas abundance. Four of these filters are transparent in the ultraviolet wavelength region (for example, a filter called UV375), while the others are located in the green, yellow and red wavelength regions.

The image above is really three different WAC images, taken through different filters. The red filter was used during an exposure that lasted one minute. The green filter was used during an equally long exposure. Finally, the UV375 filter was used three times with a total exposure time of nine minutes. By combining these images, the color photo above could be constructed.


The second camera is called the Narrow Angle Camera (NAC). It has a smaller field of view than the WAC, but is capable of resolving objects that are five times smaller than the ones the WAC manages to resolve. This camera will primarily be used to study the comet nucleus. This camera also has Swedish filters, but with quite different properties – a mixture of broadband filters in different parts of the visible wavelength region to make a rough characterization of the comet spectrum, and a number of filters that will be used to search for specific minerals, like pyroxene, hematite and hydrated silicates.

The figure below shows a picture taken with the NAC, and corresponds to the white square in the picture above. The strongly magnified picture shows a globular cluster called Messier 107 (or M107), as well as the comet nucleus within the small circle. It is still far too distant to be seen in detail, and is only a dot in the sky. But day by day Rosetta is closing in on the comet and soon we will be able to see how it looks like up close!



Comet 67P/Churyumov-Gerasimenko in constellation Ophiuchus. A zoom into an image taken with the Narrow Angle Camera on March 21. The comet is indicated by the small circle, next to the bright globular star cluster M107. The images were taken from about 0.03 AU distance to the comet. Rosetta was at a distance of approx. 4.4 AU from Earth. Image credit: ESA ©2014 MPS for OSIRIS Team MPS/UPD/LAM/IAA/SSO/INTA/UPM/DASP/IDA

For ESAs press release click here.


Why are comet nuclei dark and red?

A comet nucleus is a few kilometers in diameter and consists of a very porous mixture of ice, organics and a fine rocky powder. They often move on a highly elliptic orbit around the Sun, which means that they spend most of their time in extreme cold at large distances from the Sun. The ice is then so cold that it is stable, even when it is directly illuminated by the Sun and exposed to the vacuum of space. But periodically the comet passes its perihelion, its smallest possible distance to the Sun, and on its way there, the surface layer of the comet is heated sufficiently to vaporize the ice. The gases rush out from the comet nucleus, drags dust grains along and form the coma, a vast temporary atmosphere, along with two tails consisting of gas and dust, respectively. The basic properties of comets have been described in an earlier post.


Comet C/1995 O1 (Hale-Bopp) photographed in april 1997. The dust tail is yellow-white since it consists of small dust particles that reflect the light of the Sun. The plasma tail is blue since light with this color is emitted by ionized carbon monoxide in the tail.
Copyright: E. Kolmhofer, H. Raab; Johannes-Kepler-Observatory, Linz, Austria
Original image: http://en.wikipedia.org/wiki/File:Comet_Hale-Bopp_1995O1.jpg

But how close does a comet nucleus need to get to the Sun, in order to send gas and dust into space? How do the production rates of gas and dust vary with the distance to the Sun? Are all parts of the comet nucleus emitting the same amount of gas and dust, or are some regions more productive than others? In order to answer questions of this type astronomers use a tool called thermophysical modeling. The purpose of this modeling is to calculate the temperature of the comet nucleus at different depths below the surface, and how these temperatures are changing with time, as the Sun rises and sets in the sky as seen from a specific point on the comet surface, due to the rotation of the nucleus. These temperatures are depending on a number of different things, like the intensity of solar light, the capability of the surface material to absorb solar radiation, its heat conductivity, its porosity, the abundance of ice, and the physical properties of water as a chemical species. The final product of these calculations is an estimate of the number of water molecules that leaves a square meter of comet surface every second. By adding contributions from different parts of the comet nucleus, that are illuminated by the Sun in different ways at any given moment, it is possible to calculate the total production of gas of the nucleus.

My research has focused on thermophysical modeling for many years, and in a number of posts I will describe the thermophysical model I developed together with Dr. Yuri Skorov, a Russian scientist who presently works at Technische Universität Braunschweig in Germany. The model has several parts that are joined to form a whole – I here describe the first part, that deals with the way the surface material of comet nuclei absorb sunlight. Such research can provide clues as to why comets are so dark – they reflect merely a few percent of the sunlight – and why they are so red. As a fact, they reflect red light much more efficiently than blue light.

What happens to the solar radiation at the surface of the comet?

In order to describe the outgassing of a comet one needs to solve several sub-problems. One of these deals with the sunlight as it hits the surface of the comet. What fraction of this radiation is reflected back into space? What fraction of the radiation is absorbed, so that its energy can be used to vaporize ice? Where does this absorption take place – how deep are the solar rays penetrating?

When I started my PhD studies at Uppsala university in 1998, most thermophysical models used by scientists worldwide, assumed that solar radiation was absorbed at the very surface of the comet, within an infinitely thin layer. The reason for this assumption was very practical – there are so many other things to take into account, that the details of solar light absorption were overlooked in favor of other mechanisms considered more important. However, we envision the comet surface material as very porous – the material consists of loosely bound grains of ice and rock, surrounded by cavities and empty spaces since the gravitational force of the little nucleus is incapable of compacting the material. The solar radiation should be able to penetrate fairly deep into such a porous structure before being absorbed, especially since many of the grains consist of ice, which is rather transparent. Since ice at a depth of a few millimeters or centimeters below the surface therefore may be illuminated directly by sunlight, and if large amounts of vapor therefore are produced at depth, the outgassing of a comet would perhaps behave differently, compared to an object were all solar energy indeed is absorbed within the uppermost layer of tiny grains.

On a related note – the space probe Voyager 2 had flown by Jupiter in July 1979, and observed its ice-rich satellite Europa with an infrared spectrometer, that measures the heat radiated from the surface of Europa. Based on these measurements it is possible to calculate the temperature that prevails on different parts of the surface of Europa. When these temperatures, valid at different local times (night, early morning, noon), were compared to calculations, it was found that thermophysical models failed to reproduce the observations if they assumed that all radiation was absorbed at the very surface. Only when the model allowed radiation to be absorbed gradually within a thicker layer of ice near the surface was it possible to match the calculations with the measurements.

If all radiation is absorbed at the very surface, a sunlit spot on the dayside of Europa will have its highest temperature at the very surface, and the temperature decreases with depth below the surface. However, if the solar radiation is absorbed gradually within a thicker layer, the surface becomes rather cold. Now the temperature increases with depth and reaches a maximum a few millimeters or centimeters below the surface, and is then falling at even larger depths. This phenomena is called a solid-state greenhouse effect and is due to the fact that radiation easily can enter, but has difficulties getting back out. This is because sunlight primarily consists of visible light, and at such wavelengths ice is rather transparent. But the heated ice below the surface tries to get rid of this heat by emitting infrared radiation, at wavelengths where ice no longer is transparent. The energy cannot get out, but accumulates – this causes the temperature peak just below the surface – and it can only be removed by heat conduction or by consumption through ice sublimation.

Nobody had seriously considered solid-state greenhouse effects for comet nuclei, and since I needed a relatively poorly studied topic for my doctoral thesis, I thought I could work on this problem. But where to start?

Comet grains

First one needs to have an idea of how the grains look like in the surface region of the comet – what is the grain size, what minerals and substances are present, how are these materials arranged within the grains, how much void space surrounds the grains, and are the grains homogeneously distributed or lumped together into little clusters?

We know that comets contain grains consisting of silicates, sulfides and metals, typically having a size of a tenth of a micrometer, that is to say, ten thousand times smaller than a millimeter. This is because oxygen, silicon, magnesium, iron and sulphur are some of the most common chemical elements in the interstellar medium from which the Solar System formed (see the post about Voyager 1 and the interstellar medium), and since these elements team up to form silicates, sulfides, and metals – the mixture we call rock. We know that most of the grains have a size of a micrometer or smaller, because they are expelled in large numbers by active comets to form their huge dust tails. The sizes of these grains can be inferred from their motion – the dynamics is determined by the gravity and radiation pressure of the Sun, in a way that makes it possible to estimate the sizes of the grains.


A grain from Comet 81P/Wild 2, captured by the NASA spacecraft Stardust and brought back to Earth. The grain consists of the mineral forsterite, a silicate belonging to the olivine family, consisting of two magnesium atoms (Mg) and four oxygen atoms (O) for each silicon atom (Si); Mg2SiO4. The particle is about 2 micrometers across.
Image Credit: NASA/JPL-Caltech/University of Washington
Original image: http://photojournal.jpl.nasa.gov/catalog/PIA02190

It is likely that such small cores of rock are covered by a mantle of organic substances. This is because carbon is even more common in the interstellar medium than the previously mentioned elements, and it must be present in the comet in one form or another. This is because organic substances have a much lower condensation temperature than silicates, sulfides and metals, and require very cold environments to solidify from the gas phase. Therefore, organic species are rare in the inner Solar System (for Mercury, Venus, Earth, Mars and the inner parts of the asteroid belt), but much more common in the outer Solar System – for example in the comets. The inner Solar System was too warm for organic species to condense in large numbers, but in the outer Solar System this process could proceed undisturbed. In this way, mantles of organics likely formed on the pre-existing grains of rock. Such a grain can be seen in panel A in the figure below.


Drawings of the grain types discussed in this post.

Another species that was present in huge quantities was water – which has an even lower condensation temperature than the organic species. Therefore, essentially all water in the inner Solar System was in the form of vapor, while the substance gradually could condense and freeze in the outer parts of the Solar System. Thus thick crusts of ice were formed on the grains, on top of the layer of organics. This ice could either be rather pure (panel B) or possibly mixed with organic substances and rather dirty (panel C).

About 4.57 billion years ago such free-flying grains started to stick to each other as they collided at low speed, and with time this fluffy ball of rock, organics and ice grew to a size of several kilometers – this is the comet nucleus. We have reasons to believe that these grains are not strongly compressed within the nucleus, but form a loosely bound structure where grains barely touch, and where cavities of different sizes surrounds threads or membranes of loosely bound grains, which makes the material very porous and fragile. The force of gravity is too weak to squeeze the material into a more compact dirty snowball – it may have a structure resembling cotton sugar. We often see that comet nuclei fragment, split or are pulverized completely, which means that they are extremely fragile. We know that comets have a very low density and high porosity since we can measure their masses and volumes. Their masses are very low compared to what we could expect, given their volumes. It is necessary to take this very high porosity into account when trying to calculate the penetration depth of the solar radiation.


Comet 73P/Schwassmann-Wachmann 3 imaged by the Hubble Space Telescope in April 2006. The comet nucleus is fragmenting. At the upper right is one of the largest fragments, “fragment B”, which in turn has ejected three dozens of smaller pieces that drift outward in its tail. Many of these smaller fragments have tails of their own due to outgassing and expulsion of dust grains.
Image Credit: NASA, ESA, H. Weaver (JHU / APL), M. Mutchler and Z. Levay (STScI)
Original image: http://apod.nasa.gov/apod/ap060504.html

We do not know if the grains are homogeneously distributed, or if they form lumps here and there – something we call clusters. Panels D and E show examples of such clusters – in the former case the cluster consists of type A grains (they only contain rock and organics), while clusters of type E consist of type B grains (containing ice in addition to rock and organics). We can also imagine that ice evaporates but that the vapor freezes anew. If this happens over and over again, larger chunks of more compact ice could form, where particles of rock and organics (type A grains) are finely distributed within the chunk – such a dirty ice particle is seen in panel F.

The interaction of sunlight with single grains and clusters

Light, also called electromagnetic radiation, consists of an electrical field and a magnetic field that oscillate about each other. When light from the Sun penetrates the surface layer of the comet, these electric and magnetic fields start to interact with the electrically charged particles – atomic nuclei and electrons – that the grains consist of. As a result of such interaction some of the solar energy is absorbed – this gives the atoms higher vibration speed since they have absorbed the energy of the light. The parameter we call “temperature” is a direct measure of these vibration velocities, so the absorption of sunlight of the grains simply increase their temperature. But a part of the radiation will not be absorbed, but re-emitted by the grain in various directions – we say that the radiation is being scattered.

If we want to know how solar radiation behaves as it enters the surface layer of the comet we therefore first need to understand how the solar radiation interacts with a single grain, or a cluster of such grains. Exactly how much radiation is absorbed and how much is scattered? Exactly how much radiation is scattered into a specific direction, for example, perpendicularly to the original direction of propagation of the light? How do these things depend on the particle size, its chemical composition, and internal structure? How are these dependencies changing depending on the wavelength of the radiation?

There is a whole branch of science dealing with questions of this sort – light scattering theory. The first step was therefore to use techniques that had been developed in this research field. For the simplest grains (type A, B, and C), consisting of spheres or concentric shells with different composition, it is possible to use Mie theory. Mie theory is based on Maxwell’s equations – a group of equations that describe the properties of the electric and magnetic fields that light consists of. By using the Maxwell equations, Mie theory calculates exactly what happens to the electromagnetic wave as it encounters a spherical particle. The result is an analytical solution – a mathematical formula were one can enter the particle size, the thickness of the different layers, the wavelength of the radiation, and parameters that describes the optical properties of the materials (these differ strongly between various types of silicates, sulfides, organics and ice and are measured in the laboratory). The analytical solution then tells us how much radiation is being absorbed or scattered, and in what way.

However, it is not possible to apply Mie theory for the clusters seen in panels D and E. These grains have an extremely complex geometry – matter with very different optical properties are mixed in an elaborated manner, and there are a lot of cavities and spaces where radiation may bounce around and change direction. If a ray of light falls on such a complex particle, how much is absorbed and how much is scattered?

There are many techniques available to calculate the behavior of the electromagnetic wave when it interacts with such a strange grain. In our work, we used a technique called the Discrete Dipole Approximation or DDA. A dipole consists of a particle with positive electrical charge and a particle with negative electrical charge, separated by some distance. If the dipole is illuminated, the electrical and magnetic fields of the light will set the charged particles in motion. An oscillation takes place, so that the distance between the two parts of the dipole periodically is decreasing and increasing. Electrical charges that perform acceleration and deceleration will emit light themselves, into completely different directions compared to the light that initially set the dipole in motion. The energy theft of the dipole is the very reason for the light absorption of the material, while its emission is the source of the scattering phenomenon.

In DDA, the grains are represented by tens of thousands of individual dipoles, that each have their specific position. The dipoles located in a region consisting of silicates are given properties that are typical of silicate. The dipoles located in a region consisting of organics or ice are given properties consistent with those substances. Cavities between grains are not given any dipoles at all. A computer code then calculates what happens to every single dipole, when it is exposed to the combination of sunlight and radiation produced by all other dipoles within the grain. In this way, DDA can calculate very accurately how much sunlight is being absorbed or scattered by the cluster as a whole.

For grains of type F we used a third technique called geometrical optics. That is the type of physics used when calculating what happens to radiation when it passes through large objects like lenses (in telescopes, glasses, etc) or large drops of water. Geometrical optics assumes that the objects is much larger than the wavelength of the light. It is possible to calculate how much radiation that is reflected at the surface of the grain, and how much that enters the grain. For the transmitted radiation, the gradual absorption can be calculated, as well as the scattering, especially considering the presence of small particles of rock and organics that are baked into the grain. This calculation provides the same type of information as Mie theory and DDA – how much radiation is absorbed, how much is scattered, into what directions is the scattered radiation going, and how does these things depend on the size of the chunk and its content of type A grains?

Radiative transport

Thanks to these calculations we had come quite a way. We now had a fairly good idea of how individual grains, or clusters of such grains, interact with sunlight. The next step was to figure out how an entire medium of such grains behave – that is to say, large quantities of such grains or clusters that are spread out within a volume, with empty space between them in a porous yet coherent structure. Basically – what happens to sunlight as it shines on a dirty and very fluffy snowball consisting of millions of little grains?

The first step was to cut out small cubes of material and figure out how such cubes would absorb and scatter radiation, accounting for the fact that only a fraction of their interior consisted of grains or clusters, while most of their volumes were empty space. Through this calculation we could account for the porosity of the comet surface material.

The next problem was far more complicated. The surface region of the comet will consist of an entire network of such cubes. Each cube will receive some radiation from the Sun, keep some for itself, and pass on the rest into different directions. But since each cube is emitting radiation like this, it means that the Sun is not the only source of light. In reality, each cube will be illuminated both by the Sun and by all other cubes, and it is this combined radiation field each cube needs to deal with. Luckily it is possible to describe this complex network of interdependencies of cubes with the aid of a powerful mathematical tool – the so called equation of radiative transfer.

In fact, all of our work had aimed at realistically describing the terms for absorption, degree of scattering and type of scattering that need to be known before one even can attempt to solve this equation. However, the equation of radiative transfer is far too complex to be solved analytically, why we used the computers again. By doing so we obtained numerical solutions that finally allowed us to achieve our goal. We could tell, in a fairly realistic manner, what happens when the Sun shines on a comet nucleus.

Reflectivity and color

By observing comet nuclei we know that they are very dark – they reflect only a few percent of the sunlight that illuminates them. For example, Comet 1P/Halley only reflects 4% of the sunlight. We also know that they are rather red, in the sense that they reflect light at long wavelengths (red) more efficiently than light at short wavelengths (blue). We also know that most comets produce very little water vapor compared to the available nucleus surface area, which means that large parts of this area is ice-free and inactive. The inactive areas, that dominates the surface of comets, could perhaps consist of grains similar to type A or D, that lack ice.

Therefore it is interesting to compare our theoretical calculations to the actual observations. As it turns out, a porous medium consisting of 0.1 micrometer type A grains with 70% porosity, will reflect 9% of the sunlight if they are not forming clusters. This is a bit too much, but we also found that the reflectivity is strongly dependent on the size of the grains – if the grain size is cut in half, the reflectivity drops to 2%. This is very close to the reflectance of real comets. However, there is a problem – our modeled medium is strongly blue – it reflects the short-wavelength light more effectively than the long-wavelength light, which is not the case for real comets.

If the grains are allowed to form small clusters (like in panel D), something interesting happens – the reflected radiation becomes very red, and is similar to that of real comets. A medium consisting of small clusters of type D do reflect a bit too much radiation, about 13%. However, this is possibly due to the porosity of the medium. We had assumed a very high porosity in our calculations, about 76%, but laboratory experiments show that this type of media may be even more porous, reaching values in excess of 90%. Such a high porosity makes it even easier for radiation to enter, and if a reflection takes place, the radiation most likely does not manage to exit, but illuminates some other part of the medium, where it may get absorbed. Therefore, if the porosity is increased further, the reflectivity could be substantially lowered.

It is therefore possible that the inactive parts of cometary surfaces are dominated by extremely small grains of rock and organics, that often form little clusters, and that this medium is extremely porous. In fact, such a medium would be as dark and as red as real comet nuclei.

How would a spot of dirty ice look like? According to our calculations, a porous medium consisting of type E clusters would not look much different than a porous medium of ice-free clusters. It is still very red, but reflects a bit more – about 20% of the sunlight if the ice is clean, but only 13% if the ice is dirty. These clusters contain three times as much ice as rock. An icy spot therefore does not necessarily look white and shiny like newly fallen snow, but can be very dark. The reflectivity depends on the way ice, organics and rocks are mixed with each other, and the size of pieces of “pure” rock and organics compared with the wavelength of the light that illuminates them.

If the grains look like those in panel F, the calculations show that the material may become very dark – it only reflects 3% of the light and the color is fairly neutral. There is not a big difference in reflectivity at long and short wavelengths.

How deep does the radiation penetrate?

The single most important property of the medium that decides where and how sunlight is absorbed is the size of the cavities between the grains. If a medium consists of 0.1-1 micrometer grains or clusters that are homogeneously distributed, the cavities will also have sizes of a few micrometers. On its way through the medium, the radiation is “forced” to pass through grains, where it rapidly becomes absorbed.

If the grains look like those in panel B, about 90% of the radiation will be absorbed at a depth of 10 micrometers, while 99% of the radiation is absorbed in the upper 45 micrometers. This is very deep in a relative sense – the grains are assumed to have a radius of merely 0.16 micrometers – but is very short in an absolute sense. For such a medium, an assumption of immediate absorption at the surface is probably accurate.

If the grains look like those in panel F, and if they have a size of 100 micrometers (i.e., a tenth of a millimeter) due to repeated cycles of sublimation and condensation, about 1% of the sunlight is still present at a depth of two millimeters. If the grains instead have a size of a centimeter, there may be unabsorbed sunlight even at a depth of 20 centimeters. This is because the cavities have sizes of 0.01-1 centimeters in these media, which allows radiation to reach rather large depths.

It is therefore clear that the properties of the surface material – not only its composition and porosity, but also its purely geometrical arrangement – has a potentially important effect on the absorption of sunlight, and hence the heating and outgassing of comets. Many other factors play a role as well – but that is for another post!


Brown, R. H., Matson, D. L. (1987). Thermal effects of insolation propagation into the regoliths of airless bodies. Icarus 72, 84-94.

Davidsson, B. J. R., Skorov, Y. V. (2002). On the light-absorbing surface layer of cometary nuclei. I. Radiative transfer. Icarus 156, 223-248.

Urquhart, M. L., and Jakosky, B. M. (1996). Constraints on the solid-state greenhouse effect on the icy Galilean satellites. Journal of Geophysical Research 101, 21,169-21,176.

Mysteries of the asteroid belt

Most of the asteroids in the Solar System are located between 2.1 and 3.3 AU from the Sun and constitute a population called the main belt (1 AU = one astronomical unit, corresponding to the mean distance between Earth and the Sun, or roughly 150 million kilometers). The main belt is outside Mars that is located at 1.5 AU, and interior to Jupiter that is 5.2 AU from the Sun. This post is about two of the largest mysteries of the main belt – why does it contain such a small mass and why are the orbits of asteroids around the Sun so extreme?


The Asteroid (21) Lutetia with a diameter of about 100 kilometers, imaged with the camera OSIRIS onboard the European spacecraft Rosetta.
Original image: http://www.esa.int/spaceinimages/Images/2010/07

The total mass of main belt asteroids is only about 0.0005 Earth masses, which means that one would need about 2000 asteroid belts to build a planet of Earth’s size. If we also consider that Mars only has 0.1 Earth masses, it means that the region between about 1.3-5.0 AU from the Sun only contains roughly a tenth of an Earth mass, although there should have been several Earth masses there. Where has all of this mass gone? Furthermore, the orbits of main belt asteroids are characterized by rather high eccentricities (their orbits are clearly elliptic) and inclinations (they often tilt significantly with respect to Earth’s orbital plane). We say that the population is dynamically hot, which is different from the large bodies in the region – Venus, Earth, Mars and Jupiter – that all have more or less circular orbits in almost the same plane, which makes them dynamically cold. How have the asteroids obtained these strange orbits?

There are currently two different scenarios that seek to explain both why the area between Earth and Jupiter is almost empty, and why most bodies there are dynamically hot. Both scenarios are very dramatic. If we could decide which of them that is correct, an important but yet unwritten part of the history of our Solar System could be clarified. By describing these scenarios in detail we also touch upon some other important questions – how is natural science conducted, what does it take to prove a scientific theory, and what is the difference between unfounded opinions and scientific hypotheses? However, before addressing these questions it is necessary to describe how the asteroids formed, and how the asteroid belt looks like today.

Planetesimals in the Solar Nebula

About 4.57 billion years ago, something happened that frequently takes place in our galaxy, the Milky Way – a part of a molecular cloud (see an earlier post about the interstellar medium) contracted due to its self-gravity and formed a starless core – a cold lump of gas and dust that measured 10,000 AU across. It may have maintained an equilibrium configuration for a long time, up to one million years, before it collapsed further and formed the protosun – a structure from which the Sun formed – and the Solar Nebula, a flat and warm cloud of gas and dust that revolved around the protosun. This final collapse could have happened spontaneously, but it is more likely that it was initiated by the explosion of a nearby type II supernova – such supernovae produce a cocktail of short-lived radioactive isotopes such as aluminum-26, iron-60, chlorine-36, manganese-53, and calcium-41, whose decay products are found in the meteorites that impact Earth.

We can see such young, newly collapsed systems around us and they are called class 0 or I protostars, depending on how far they have come in their evolution. These phases last a couple of hundred thousand years in total. The hot environment in the Solar Nebula close to the protosun was suitable for creating a kind of grain cluster called calcium-aluminum-rich inclusions or CAI. They consist of different minerals like melilite (a mix of åkermanite and gehlenite) and fassaite that are rich in calcium and aluminum. These can be found in meteorites and can be dated through radiometric methods – it is the age of these grains of 4.57 billion years that we define as the age of the Solar System. The measurement of time in the Solar System use CAI as a reference – a given moment in time is clocked as a certain numbers of years “after CAI”.

Also other types of particles formed at lower temperature, that were rich in oxygen, silicon, magnesium, iron, and sulphur, such as amoeboid olivine aggregates (AOA) and agglomeratic olivines (AO). All these particles have typical sizes of 0.01-1 centimeters, are often very porous, and there are reasons to believe that the vast majority of grains stopped growing at this size due to a phenomenon called the bouncing barrier – if two particles of these kinds collide with each other, the probability that they will stick to each other and build something bigger, is very low.

A fraction of these CAI, AOA and AO managed, in spite of the difficulties, to merge early on into boulders measuring decimeters or meters across. These could, in turn, merge to form even larger bodies – planetesimals – with sizes measuring hundreds of kilometers. The planetesimals that formed to within one million years after CAI contained sufficient amounts of radioactive aluminum-26 to get heated to the point that they melted. Thereby the grains were destroyed, and the minerals broken down into their atomic constituents. The gravity of the planetesimal forced heavy elements like iron, nickel and sulphur to sink towards the center of the body, where they formed a nucleus rich in metal and sulfides. On top a mantle formed that contained lighter elements – mostly oxygen, silicon, and magnesium – recombining to form minerals like olivine and pyroxene. Possibly, an outer crust formed that consisted of the lightest minerals, like pyroxene and feldspar – the latter rich in the aluminum and calcium that originally had been located in the CAI. Such a layered planetesimal is said to be differentiated.

However, we have reason to believe that the vast majority of the CAI, AOA and AO did not participate in this process. They continued to orbit the protosun, along with the differentiated planetesimals. The heat generated within such small particles due to the decay of the short-lived radioactive substances could easily escape to space, and the particles remained rather cold. After 2-3 million years, the gas began to leave the inner parts of the Solar System, and the disk around the protosun became much dustier. This appear to have given rise to some form of electromagnetic phenomenon – perhaps current sheets or electric discharges – capable of flash-melting large amounts of CAI, AOA and AO, which transformed the porous clusters of dust to small, hard, compact balls of rock. These new types of particles are known as Type C CAI and chondrules, respectively. Large amounts of chondrules have been found in meteorites, and radiometric dating shows that the majority have been formed 2-4 million years after CAI. The process also seem to have given rise to a fine-grained mixture of grains consisting of olivine, pyroxene, sulfide, metal, and organic substances called matrix material. Although chondrules and matrix material individually have a chemical composition that differs from that of the Sun in terms of elements heavier than hydrogen and helium, the sum of chondrules and matrix material is very solar-like.

The chondrule formation appears to have given rise to a second, and perhaps dominating, wave of planetesimal formation. It seems like the bouncing barrier suddenly could be crossed as soon as the porous collections of grains (AOA and AO) were transformed to smaller compact spheres of rock (chondrules). The large bodies that formed at this stage did not contain very high abundances of aluminum-26 since most of the substance already had decayed. Therefore, these relatively late planetesimals were never molten, and they did not differentiate. Their interiors still contain surviving and largely unmodified grains – calcium-aluminum-rich inclusions (CAI), amoeboid olivine aggregates (AOA), agglomeratic olivines (AO) and matrix material.

Asteroids and meteorites

The asteroids we see today are a few surviving examples of these different types of planetesimals – the rest have been used to build the planets in the Solar System. Some of these asteroids are very old, and constitute the oldest differentiated type. For example, Asteroid (4) Vesta is a body that is known to contain a core of iron, nickel and sulphur, a mantle of olivine and pyroxene, and a basaltic crust – volcanic rock rich in olivine, pyroxene, silica and feldspar. We also know that many of these extremely old differentiated asteroids have been smashed to pieces in violent collisions. Such pieces of differentiated asteroids often impact Earth as meteorites – iron meteorites from the metallic core of the parent body, stony irons from the transition region between the core and mantle, and achondritic meteorites from their crusts. Vesta itself is the parent body of a large group of achondritic stony meteorites called howardites, eucrites, and diogenites, or HED with a common name. A few asteroids, apart from Vesta, also seem old enough to have differentiated. For example, so called M-asteroids may be parts of the iron core of larger smashed-up planetesimals, while A-asteroids possibly are parts of the mantle from such bodies.


The Asteroid (4) Vesta photographed by the NASA spacecraft Dawn.
Image credit: NASA/JPL-Caltech/UCAL/MPS/DLR/IDA

However, the majority of main belt asteroids seem to belong to the younger undifferentiated variant of planetesimals, and their collision fragments. Pieces from these undifferentiated bodies also frequently impact Earth, and are called chondritic meteorites, since they are so rich in chondrules. The innermost parts of the main belt is rich in E-asteroids, that are believed to be related to a certain type of stony meteorite called enstatite chondrites. Besides that, the inner half of the main belt is dominated by S-asteroids, that are known to be related to another type of stony meteorite called ordinary chondrites. We know this since the Japanese spacecraft Hayabusa went to an S-type asteroid named Itokawa, and brought back small parts of its surface material to Earth. When investigated in the laboratory they turned out to be identical to ordinary chondrite meteorites. The Chelyabinsk meteorite that I have written about previously, was also an ordinary chondrite. Finally, the outer parts of the main belt is dominated by C-asteroids, that are believed to be related to yet another type of meteorite – the carbonaceous chondrites. All chondrites have one thing in common – they contain a mixture of chondrules and matrix material, in different proportions. The differences between enstatite chondrites, ordinary chondrites and carbonaceous chondrites include; the mixing ratio of chondrules and matrix material; whether iron is located in separate metallic grains (is reduced) or finely distributed within the minerals (is oxidized); whether olivine is present among the dominating pyroxene; the abundances of rare oxygen isotopes compared with the most common form, oxygen-16. These differences reflect systematic changes in temperature and pressure within the Solar Nebula, as function of time and distance from the protosun.

The asteroid belt today

There are in total 220 asteroids in the main belt with diameters D of 100 kilometers or larger. The four largest ones are called (1) Ceres (D=930 km), (2) Pallas (D=580 km), (4) Vesta (D=525 km) and (10) Hygiea (D=410 km). There are in total 680 asteroids with diameters of 50 kilometers or more, and the number of asteroids with diameters in the 10-50 kilometer interval is about 7,000. The known population is considered complete down to sizes of 10-15 kilometers, which means that we know all individual objects that are at least that big. The number of asteroids larger than a kilometer is estimated to be 1.3-1.4 million. Currently, we know about 630,000 asteroids, which means that about half of all asteroids larger than a kilometer already have been discovered.

The largest asteroids – the 220 objects larger than 100 kilometers – are most likely surviving planetesimals, i.e., they formed 4.57 billion years ago in their current form. Almost all other asteroids are considered collision fragments, i.e., they are pieces of even larger objects that have been broken up in collisions. Therefore, they have not had their current appearance since Solar System childhood, but have formed throughout the long history of the Solar System during violent collisions. There are three properties of asteroids that change systematically around a size of 100 kilometers. First, there is a clear change in the size distribution of asteroids, i.e., a list of the number of asteroids having a given size. If one consider the total number of asteroids larger than a certain size (the cumulative size distribution), it increases fast when D is reduced from 930 kilometers to 120 kilometers. But if D is reduced further, the cumulative size distribution does not change very fast at all, until a size of about 30 kilometers is reached, at which the increase is fast anew. Second, asteroids larger than 100 kilometers are almost spherical, while smaller asteroids systematically become more irregular the smaller they are. Third, there are systematic changes in the rotational periods of asteroids – in the D=100-930 kilometer interval the rotational period increases with decreasing size, which means that the smaller asteroids tend to rotate slower than the larger ones. But around D=100 kilometers this trend is reversed, so that even smaller asteroids tend to have shorter rotational periods – small asteroids spin faster the smaller they are. These three properties show that we are dealing with a population of objects with diameters larger than 100 kilometers that are primordial surviving planetesimals – their sizes, shapes and spin properties are consequences of the process or processes that formed planetesimals early in Solar System history. Instead, objects with diameters smaller than 100 kilometers are collision fragments – their sizes, shapes and spin properties are consequences of what happens when to large asteroids collide with each other at a high velocity.

Most of the mass in the main asteroid belt is locked up in the largest objects. The fact is that it is sufficient that 10-20 asteroids with sizes in the range 100-1000 kilometers collide, to explain the number of all asteroids that are smaller than this. Since we know all asteroids larger than 10-15 kilometers, and have a fairly good idea of how the size distribution looks like at even smaller sizes, we can say with certainty that the total mass in the main belt is not higher than about 0.0005 Earth masses.

However, the original amount of mass (in the form of rock, metal, sulfides and organic substances) in the 1-4 AU region must have been significantly higher, and may have been 5-8 Earth masses. This estimate is based on our observations of circumstellar disks around foreign protostars in the Milky Way, that currently are in the same stage of evolution as our Solar System was 4.57 billion years ago. In such disks, the amount of mass in the disk changes in a characteristic way with increasing distance to the parent star. If this is compared to the amount of matter that has been locked up by the planets in our own Solar System internal and external to the 1-4 AU region, we can conclude that the primordial asteroid belt must have been several thousand times more massive than today. A very dramatic event, or chain of events, must have led to this drastic mass loss in the asteroid belt.

We also know that the main belt asteroids originally must have moved on almost circular orbits, that all were located more or less in the same plane. The consequence of having such orbits is that the objects meet at low velocity when they collide – perhaps a few tens of meters per second. Such gentle collisions are needed to allow gravity to keep colliding bodies together, so that a larger body can form as a result of the collision. In order to build bodies with sizes measuring several hundreds of kilometers, it is necessary that even smaller bodies collide with each other at very low speed.

However, today the asteroid orbits have high eccentricities and inclinations. The orbits are no longer circular but are shaped as ellipses. The ellipse has a center, but the Sun is not located there but in a focus point that is displaced towards the point in the orbit where the asteroids is as closest to the Sun as possible, called the perihelion. The eccentricity is defined as the distance between the center and the focus point, divided by the distance between the center and the perihelion. The average value for the eccentricity of asteroids is 0.15, which means that the displacement of the Sun from the center of the ellipse is 15% of the distance between the center and perihelion. The average value of the orbital inclinations is 8 degrees, which is the average tilt of the asteroid orbital planes compared to the orbital plane of Earth. These high eccentricities and inclinations mean that asteroids have very high velocities when they collide – the average velocity is 5 kilometers per second. It is not possible to build anything during such powerful crashes – all collisions become destructive and only makes the population slowly grind itself to dust. Something must have happened in the asteroid belt that changed growth in dynamical cold to grinding in dynamical heat.

The first scenario – embryos in the asteroid belt

According to the first scenario the main belt originally was a few thousand times more massive than today, and all bodies moved more or less on circular orbits in a common plane. For this reason, small planetesimals could merge into larger planetesimals rather fast. Computer simulations show that it takes about one million years to form bodies as large as the Moon. The Moon has a diameter of D=3470 kilometers and a mass of about 0.01 Earth masses. It is even possible that bodies formed with masses around 0.1 Earth masses, that would be as large as the planet Mars. Bodies with masses of 0.01 Earth masses or larger are normally not referred to as planetesimals, but are called embryos.

The presence of embryos in the main belt had a large effect on the orbits of smaller planetesimals. They could no longer maintain their circular orbits in a common plane, but gradually obtained ever higher eccentricities and inclinations due to gravitational perturbations. Paradoxically, the formation of embryos – a consequence of efficient growth due to low collision velocities – therefore leads to the end of growth, and increasingly efficient fragmentation since the collision velocities are too high.

However, we know that Jupiter formed around the same time – the gas giants must have formed at most five million years after CAI since observations of gas disks around foreign protostars show that they do not live longer than that. Jupiter and Saturn must have had time to grow by consumption of such gas while it was still available. Uranus and Neptune, that mostly consist of ice, only managed to bind smaller amounts of gas since they formed at a stage when the gas disk already was dispersing. The formation of Jupiter had an extremely important effect on the asteroid belt – mean motion resonances were formed.


The asteroid mean distance to the Sun measured in Astronomical Units (AE in the figure) on the horizontal axis, and the orbit inclination on the vertical axis. Every dot in the diagram corresponds to a known asteroid. The vertical lines show the location of strong mean motion resonances. The name of the resonance as well as its distance to the Sun is marked at the top of the figure. Note that several resonances virtually lack objects (so-called Kirkwood gaps). The 3:2-resonance contains a population of objects called Hilda asteroids.

The orbital periods of the asteroids increase systematically with increasing distance to the Sun. At certain specific distances from the Sun, the orbital period will constitute a small-integer fraction of Jupiter’s orbital period. For example, in the inner parts of the asteroid belt we find the 3:1 resonance at 2.50 AU – here the asteroids revolve exactly three times around the Sun in the same time as Jupiter performs one revolution. Further out, at 2.82 AU we find the 5:2 resonance, where the asteroids complete exactly five orbits around the Sun in the same time as Jupiter makes two revolutions. There are several such resonances and they all have one thing in common – objects that end up in these resonances are subjected to very strong gravitational perturbations by Jupiter. The most common effect is a strong increase of the orbital eccentricity, which means that the perihelion point is moved closer to the Sun. It is exactly this mechanism that creates Near Earth Asteroids. The asteroid is normally destroyed by colliding with the Sun, or more rarely, by colliding with one of the terrestrial planets.

Embryos and mean motion resonances is a deadly combination for asteroids – it took about one million years for the embryos to shuffle about 99% of the asteroids into the resonances, so that these objects left the main belt. The embryos also perturbed each others orbits, and disappeared one by one via the resonances. When the last embryo disappeared from the main belt, only traces remained of the once so massive population, and the remaining objects had received a high degree of dynamical heating – a memory of the ravaging of the embryos. If this scenario is correct, the current asteroid belt is therefore the remains of a region where the planet formation process had time to progress rather far, before Jupiter cleaned away all large bodies and only left a few smaller asteroids behind. The E-, S- and C-asteroids, the parent bodies of enstatite meteorites, ordinary chondrites and carbonaceous chondrites, thus formed very close to each other which means that the physical and chemical properties of the Solar Nebula changed very fast with distance to the protosun. The embryos that redecorated the main belt did not manage to fully erase this strong chemical gradient within the asteroid belt.

The second scenario – Jupiter visits the asteroid belt personally

However, it is possible that the clean-up and dynamical heating of the asteroid belt happened in a completely different manner. In another scenario, the entire 0.5-2.0 AU region initially consists of E-asteroids while the 2.0-5.0 AU region is completely dominated by S-asteroids. The C-asteroids are formed among and beyond the giant planets, which are jostled in a region around 5-15 AU from the Sun. If this is correct, the difference between the various chondritic meteorites is caused by a change in the chemical and physical properties of the Solar Nebula that happened very gradually across a region that covered a vast range of distances from the Sun.

In this scenario Jupiter is formed somewhere beyond 5 AU, but does not stay there since it starts to drift towards the protosun. This is called migration. As is evident from my previous post about the Kepler exoplanets, such migration is rather common for gas giants around other stars in the Milky Way. It has been investigated what would happen if Jupiter was allowed to drift all the way to 1.5 AU – were we find the planet Mars today – before Saturn catches up with Jupiter, and both gas giants drift back outwards in formation. As it turns out, a smaller external gas giant can reverse the migration of an internal larger gas giant.

If Jupiter plowes through the main belt twice – once on its way in and a second time on its way out – the asteroid belt will be heavily depleted, and subjected to dynamical heating. On the way in, Jupiter will force many S-asteroids to relocate to larger distances where they mix with C-asteroids. When Jupiter travels back, such S- and C-asteroids are replanted into the asteroid belt, which could explain why we observe such a variety of bodies within a fairly narrow region from the Sun. This scenario is called The Grand Tack.

The difference between scientific hypotheses and unfounded opinions

The two scenarios above have one thing in common – they are both solutions to the so-called N-body problem. This problem can be formulated in the following way: if there are N bodies with known masses, that are located in specific positions and are having specific velocities at a given moment, where will the bodies be at a later moment of time, and what will the velocities be, if every body feels the gravitational pull of all the other bodies? In brief, the N-body problem is about tracking the motion of a swarm of bodies as function of time, where the movements of any given body is the result of the gravitational pull from the other bodies in the swarm.

This problem was formulated very early, e.g. by Robert Hooke in 1674 and by Isaac Newton in 1687. The latter found an analytical solution to the problem for N=2, which is called the two-body problem. It is this solution that states that a single planet that orbits a star will move in an elliptic, parabolic or hyperbolic orbit, depending on its distance to the star and its velocity at the beginning of the calculation. The equation that describes how the acceleration of the body depends on the forces that acts on it – the so-called equation of motion – can be formulated rather easily when N bodies are involved. However, it is not easy to solve the equation, which means that one get access to the positions and velocities of the bodies at a given moment of time. Massive efforts were made in the 19th century to find such solutions, but it was only possible to find exact analytical solutions for N=3 under very special conditions – this is the so-called restricted three-body problem. However, methods were developed that yielded approximate solutions, that turned out to be extremely useful.

For example, it was possible to show that Uranus did not move as expected, when taking into account the gravitational force from the Sun, Jupiter and Saturn (a four-body problem). Independently of each other, the Englishman John Couch Adams and the Frenchman Urbain Jean Joseph LeVerrier used these discrepancies between the theoretical and observed orbit of Uranus to calculate the position of an hitherto unknown planet, assumed to be responsible for these discrepancies. By observing the sky in the vicinity of the calculated position of the unknown planet, the German astronomer Johann Gottfried Galle could located the object in 1846 – the planet that is now known as Neptune. While Uranus was found by accident in 1781, Neptune was found since the mid-19th century scientists had perfected the art of measuring and calculating the positions of planets in the sky to a level where they almost had complete control over the mechanical properties of the Solar System.

In a similar manner, LeVerrier could later show that the innermost planet, Mercury, did not move exactly as is was supposed to, even if the gravity of the other planets were taken into consideration. This discrepancy was later to become one of the most important proofs that the general theory of relativity by Albert Einstein was correct, since it led to an adjustment of Newtonian mechanics that exactly matched the “error” in Mercury’s movements.

Astronomers and physicists have therefore studied the problem of calculating the motion of a body affected by the gravitational forces from several other moving bodes for a long time, and fantastic progress in this field was made already 150 years ago. Nowadays, when we have access to computers with enormous capacity, it is no longer a problem to solve the N-body problem with extreme accuracy, even when N is a rather large number. For example, a spacecraft that travels through interplanetary space will be subjected to significant gravitational forces from the Sun and all planets. Yet there are no problems for ESA, NASA or any other organization to navigate in space – this sort of calculations is routine.

It is the same type of calculations that are made in order to understand how the asteroids in the main belt will react when they are exposed to the gravitational force from embryos and Jupiter. De scenarios described above are therefore based on very accurate and detailed calculations, and are extremely realistic. The starting point is some basic assumption (i.e., whether Jupiter is migrating or not), and then well-known physics and established methods are used to calculate the consequences of this assumption. Details on how these calculations are made, and the results of the calculations, are published in scientific journals, that only can be read by experts since they are filled with mathematics, physics and technical terminology. Such documentation describes a scientific hypothesis. It is a hypothesis in the sense of being a proposal, but this proposal is scientific since it is substantiated and supported by arguments and calculations that are based on centuries of research, and methods that have a demonstrated correctness.

Popularization of science aims at describing the essence of a scientific hypothesis, using few and simple words, which e.g. is made in this blog. This is an important task, because it allows large numbers of people to take part of scientific discoveries, which enriches the mind of every human, increases his or her understanding of the world and the capability to interpret what happens around us. However, it is impossible for a popularized text to give justice to the richness, complexity and depth of the investigations on which the hypothesis is based – if one wishes to understand the hypothesis at depth one must consult the original texts. The simplicity and brevity that characterizes a popularization of science can therefore be treacherous – it is difficult or impossible to imagine the complexity of the underlying machinery. It is easy to get the impression that a hypothesis is nothing but an opinion – that some scientist “believes something”. It is easy to confuse this with opinions and beliefs that are not based on any form of deeper knowledge, or investigation of the actual properties of Nature. This is what we may called unfounded opinions. In an increasingly media-based world, where the internet and social media have given people the possibility to express their opinions at an unprecedented extent, and were the flow of information is accelerating, it can be difficult to tell the difference between scientific hypotheses and unfounded opinions – at a first glance they may look similar. This becomes even clearer when unfounded opinions are presented by using a language and a “technical terminology” that gives the impression of being scientific – that is to say, pseudo science.

To claim that the Earth is a few thousand years old is an unfounded opinion, and if one tries to sell this opinion by using various technical arguments it is pseudo science. When a scientist claims that Earth was formed about 50-150 million years after CAI, that are 4.57 billion years old themselves, it is not an opinion. It is not even a scientific hypothesis, since this specific example is not a proposal – it is a fact.

Astronomers and physicists therefore do not sit and do “philosophy” in some general manner. The idea that embryos or a migrating Jupiter are responsible for the current properties of the asteroid belt are not just fantasy or speculation, it is not a matter of opinion or believing – they are scientific hypotheses since they have been formulated by using scientific methods. But if so, how come there can be two or more hypotheses claiming to explain the same thing – is not natural science about “knowing” and telling right from wrong? Why do we speak of facts and “to know” when the age of Earth is concerned, while the properties of the asteroid belt is the subject of hypotheses? The answer to that question gives further insight into the workings of the scientific process.

When a scientific hypothesis becomes fact

A scientific hypothesis is built on a number of basic circumstances, established and well-known physics, and mathematical calculations performed with the goal to explore the consequences of the basic circumstances. In order for a hypothesis to dominate over other hypotheses, and finally be referred to as a fact, comprehensive comparisons with Nature itself are needed. Such empirical data must lend support to the basic circumstances, as well as to the consequences of the hypothesis. If measurements or observations of Nature do not agree with the hypothesis, the hypothesis must be rejected or reformulated so that it no longer contradicts reality. The problem is that such empirical data often are not known, why two or more hypotheses may co-exist for a long time, until observations, measurements or laboratory experiments can shed light over the unknown circumstances needed to prove one hypothesis, and disprove all others.

Scientists are working hard to figure out how grains, not larger than a thousandth of a millimeter, merge to build boulders that are a few decimeters or meters across. There are several hypotheses on how such boulders grow to planetesimals with sizes in the 100-1000 kilometer range. The following process – the formation of embryos – is better known. In spite of this, it is difficult to know exactly how long it took for embryos to form, what size distribution planetesimals and embryos had at that time, and exactly how large the eccentricities and inclinations had grown. Another source of uncertainty is exactly when Jupiter had grown big enough that its mean motion resonances in the main belt were strong enough to effect the evolution there – did this happen only once the embryos had existed for a few hundred thousand years, or were they present even before the embryos formed? Yet another source of uncertainty is whether Jupiter migrated substantially or not. Scientists send spacecraft to asteroids and comets, make advanced computer calculations, perform experiments in laboratories and measure the properties of meteorites to answer such questions as far as possible, or at least limit the range of possible answers.

In such an uncertain situation, there is room to formulate various realistic but hypothetic basic assumption. For example, one can assume that embryos have formed, that these and planetesimals have a certain size distribution, that they have dynamically cold orbits, that Jupiter forms at a time when embryos already exist, and that Jupiter does not move substantially – one then ask, what happens then? The answer can be found in the detailed computer calculations that tracks the motion of thousands of planetesimals and embryos. Once the calculations are finished, one can explore what has happened – how many asteroids are left, what distances to the Sun do they have, what are their eccentricities and inclinations? If E-, S-, and C-asteroids initially are placed within well-defined regions at different distances from the Sun, to what extent have these populations mixed due to the ravaging of the embryos? It is exactly this type of detailed information from the computer models that are compared with observations of the Solar System today – accounting, as far as possible, for processes that may have changed various properties during the long history of the Solar System.


The Asteroid (253) Mathilde photographed by the NASA spacecraft NEAR. Mathilde is a C-asteroid, that are parent bodies of the meteorites called carbonaceous chondrites.
Image Credit: NASA
Credit: NSSDC Photo Gallery
Original image: http://solarsystem.jpl.nasa.gov/multimedia/gallery

As it turns out, the eccentricities and inclinations of the surviving asteroids in computer models that include embryos are very similar to those seen in Nature, which is encouraging. However, there are some difficulties – the innermost parts of the main belt become too sparsely populated, since embryos in those regions turn out to have a long lifetime before they are removed, and in the meantime they manage to clear the region completely. Furthermore, the region around the orbit of Mars is not cleared enough – models of this kind lead to a mass of Mars that is many times larger than that of the real planet.

The Grand Tack scenario seems to solve these problems. If Jupiter is given the opportunity to drift into the asteroid belt, and then head back out again, the amount of matter around 1.5 AU from the Sun is reduced to a level that explains why Mars only carries a tenth of an Earth mass. The distribution of matter within the modeled asteroid belt corresponds fairly well to the observed one. The computer model makes a very good reproduction of the distribution of S- and C-asteroids at different distances from the Sun, if they originally formed within, and far outside, today’s main belt, respectively. The asteroids in the computer model have inclinations very similar to those in the asteroid belt today. But also this hypothesis has problems – the eccentricities of the modeled asteroids do not resemble the ones found in the main belt today. This may be explained by the Late Heavy Bombardment, a powerful disturbance to the main belt caused by the giant planets, that is believed to have happened when the Solar System reached an age of about 700-800 million years.

Both these scientific hypotheses constitute advanced working models, that will co-exist until the amount of data gathered about the Solar System becomes large enough to prove one of the hypotheses and disregard the other – unless it turns out that a hybrid of the hypotheses is more suitable, or a third yet unexplored scenario turns out to be the right one.

This shows that Solar System science is alive and is constantly evolving. It shows that the asteroid belt contains traces of large-scale dramatic events that took place early in the history of the Solar System. The exploration of the main belt is therefore extremely important, since it helps us to understand what happened near the region where our own planet formed. Assume, for example, that Jupiter actually penetrated deep into the inner parts of the Solar System. An interesting consequence of the Grand Tack scenario is that the inner Solar System receives large amounts of mass from farther regions. If this was the case, could it be that our planet would not have formed, or that it would have looked very different, had Jupiter stayed away? Does our planet, and thereby humanity itself, exist as a result of Jupiter’s migration? It is questions like this that starts to erase the difference between Solar System history and human history – we cannot understand our own past without describing that of the Solar System.


Petit, J.-M., Morbidelli, A., Chambers, J. (2001). The primordial excitation and clearing of the asteroid belt. Icarus 153, 338-347.

Walsh, K. J., Morbidelli, A., Raymond, S. N., O’Brien, D. P., Mandell, A. M. (2011). A low mass for Mars from Jupiter’s early gas-driven migration. Nature 475, 206-209.

Walsh, K. J., Morbidelli, A., Raymond, S. N., O’Brien, D. P., Mandell, A. M. (2012). Populating the asteroid belt from two parent source regions due to the migration of giant planets – ”The Grand Tack”. Meteoritics & Planetary Science 47(12), 1941-1947.