A History of the Mesosiderite Asteroid

Alan E Rubin. American Scientist. Volume 85, Issue 1. Jan/Feb 1997.

Mesosiderites comprise a strange group of about 30 known meteorites that have been recovered from locations on every continent, where they have inspired awe in local residents (except perhaps in Antarctica) and consternation among the scientists who study them. The puzzle of the mesosiderites resides in their composition, and their name alludes to this: The roots of the word are the Greek mesos, meaning "middle" (loosely, "half"), and sideros, meaning "iron." In terms of their mass, these meteorites consist of about 50 percent metallic iron-nickel and 50 percent silicates (minerals containing silicon and oxygen). This is a curious mixture for a meteorite that has undergone melting: Iron and nickel are dense metals that are usually found in the core of planetary bodies, whereas the comparatively buoyant silicates in a mesosiderite are usually found on the surface and in the mantle of such bodies. The oddness of finding this metal and silicate concoction in a meteorite that was once molten is analogous to discovering a blend of cashews and steel ball bearings on the side of the road. It's a combination that must certainly have an interesting story behind it.

Some investigators have abandoned the thought of explaining the existence of these meteorites, often referring to them as "messy-siderites." Certainly, there are many other interesting meteorites to study: Some come from the surface of the moon, whereas others originate from Mars. (One or more of the martian meteorites may even bear the imprint of primitive extraterrestrial organisms!) Other scientists such as myself, however, have developed a particular fondness for studying meteorites that were once part of an asteroid, one of the countless minor planetary bodies in our solar system. These meteorites hold their own special fascination because they are among the oldest rocks in our possession, dating back to the time when the planets were first forming from the solar nebula. Because celestial bodies such as asteroids have remained geologically inactive for billions of years, the composition and structure of mesosiderites hold vital information about some of the complex processes that planetary bodies must have experienced early in solar-system history.

Over the past three decades investigators have analyzed thousands of meteorites, including the mesosiderites, in an attempt to reconstruct this history. However, it was not until detailed chemical, petrologic and isotopic studies were made that the evolutionary history of the mesosiderite parent asteroid could be reconstructed. David Mittlefehldt, of the Lockheed Engineering and Sciences Company in Houston, and I have compiled information about these meteorites that has allowed us to outline the evolutionary history of their parent asteroid. Using geochemical and isotopic analyses of mesosiderite clasts (fragments) and whole rocks, we have discerned six major episodes in the evolutionary history of the mesosiderite parent asteroid.

The Case of the Missing Olivine During the early formation of our solar system, about 4.56 billion years ago, the planets were being shaped by the accretion of smaller precursors: Dust grains orbiting the nascent Sun agglomerated into planetesimals, which in turn would eventually accrete into the planets. It was during this era that the asteroids, or minor planets, were also formed by the accretion of primitive material in the solar nebula. The estimate of 4.56 billion years ago for the accretionary era is based on studies of radioactive isotopes found in the most primitive asteroidal material known, the chondritic meteorites. All known asteroids are believed to have formed by this accretionary process, and the parent asteroid of the mesosiderites undoubtedly formed this way as well. Accretion was the first major event in the history of the mesosiderite asteroid. Astrophysical models suggest that only a relatively short time span, perhaps one million years or so, was required for the formation of asteroids from dust grains.

After the accretionary period the evolutionary paths of the asteroids diverged. To a large extent this divergence was a function of the asteroid's distance from the center of the solar system because most heat sources increase in intensity closer to the sun. This is true for the generation of electrical currents by the solar wind as well as heating caused by impacts (because relative velocities are greater near the sun) and heating caused by the decay of short-lived radioactive nuclei (because bodies accreted faster near the sun and thus incorporated greater amounts of undecayed radionuclei). Most asteroids, especially those in the outer parts of the asteroid belt, remained unmelted. When the pieces of these undifferentiated asteroids fall to earth they are known as chondritic meteorites. The entire geological history of undifferentiated asteroids consists of occasional bombardment by smaller bodies (meteoroids) and, in some cases, mild thermal metamorphism and alteration by liquid water But some asteroids, particularly those concentrated in the inner part of the asteroid belt, melted and differentiated. One of these was the mesosiderite parent asteroid.

The undifferentiated parent asteroid of the mesosiderites probably contained a homogeneous distribution of silicates and metal, as found in most chondrites. In the process of melting, the metallic iron-nickel and silicates separated into immiscible liquids of different densities. Scientists can model what happens when an entire primitive asteroid is melted: The denser metallic liquid sinks to the gravitational center of the body and forms a core. As the silicate melt above the core cools, crystals of olivine-a dense, dark-green silicate mineral containing iron and magnesium, (Fe, Mg)2SiO4-form and settle outside the core. Over the course of tens of millions of years, an olivine-rich mantle eventually covers the core. Partial melting of this mantle produces a buoyant magma, which erupts to the surface and forms a basaltic crust above the olivine mantle.

Basalt is one of the most common rocks on the surface of planetary bodies. It is made up of the minerals pyroxene (an iron-magnesium-calcium silicate) and plagioclase (a calcium-aluminum silicate). Basaltic melts generally have low densities and rise through the mantle as a result of buoyant forces. Extrusion of a basaltic melt onto an asteroidal surface causes the melt to cool rapidly and solidify into a finegrained basaltic crust. A coarse-grained version of basalt-gabbro-forms if a deeply buried basaltic melt cools slowly. A melt of a somewhat different composition can crystallize into a rock, consisting largely of pyroxene, called pyroxenite.

Once completely solidified, the volume of a typical differentiated asteroid would consist of about 10 to 15 percent metallic core, 70 percent olivine-rich mantle and 15 to 20 percent basaltic crust. If the asteroid had a radius of 100 kilometers, the radius of the core would be about 50 kilometers, and the basaltic crust would be about 5 kilometers thick.

This brings us to the central riddle of the mesosiderites. These meteorites seem to represent an extremely efficient mixing of crustal material (basalts, gabbro and pyroxenite) and core material (metallic iron-nickel and sulfide). If the basalts and the metals formed in the same 100-kilometer body, they would have been initially separated by 45 kilometers of olivine-rich mantle. Remarkably, very little olivine is found in mesosiderites-only about 1 to 2 percent of a typical mesosiderite's volume.

Various ad hoc models have been proposed over the years to account for the missing olivine. These include the sinking of basaltic blocks of the crust through molten mantle to land atop the metal core, the melting of a primitive assemblage rich in pyroxene and metal but free of olivine, and the partial differentiation of a primitive asteroid. None of these models has gained wide acceptance.

Melting and Differentiation


It now appears that the melting of the mesosiderites' parent body occurred at almost the same time as the accretionary process. Isotopic studies of the silicates in mesosiderites suggest that melting took place a mere 30 million years after the formation of the oldest known inclusions in chondritic meteorites.

Basalts formed during the initial differentiation of the mesosiderite parent asteroid constitute about 11 percent of the large silicate clasts in mesosiderites. These basalt clasts are indistinguishable in texture, mineralogy and composition from the pure-basalt meteorites known as eucrites. An isotopic analysis by Brian Stewart and colleagues at the California Institute of Technology indicates that a basaltic clast from the Vaca Muerta mesosiderite found in the Atacama Desert of Chile was formed 4.48 0.09 billion years ago. This date is consistent with the rock having formed virtually simultaneously with accretion 4.56 billion years ago, but does not preclude the possibility that it was formed 170 million years later. Isotopic ages need to be measured on additional basaltic clasts in mesosiderites to resolve this uncertainty.

We can also deduce some of the igneous processes that formed the original basaltic crust of the mesosiderite asteroid. There are two basic mechanisms-fractional crystallization and partial melting-which are essentially opposite means of forming igneous rocks on planetary bodies. During fractional crystallization, crystals that formed from a cooling melt become isolated and are no longer able to react with the melt. A common way for crystals to become isolated is by settling to the floor of a magma chamber under the influence of gravity Isolation of the crystals depletes the melt of the elements that are concentrated in the crystals. For example, the magnesium-iron silicate olivine crystallizes early in many igneous systems; because early formed olivine has a high magnesium/iron ratio and has less silicon than the melt, isolation of olivine from the surrounding liquid depletes the melt in magnesium more than it does in iron and enriches the melt in silicon. Crystals formed after the isolation of the olivine will therefore have lower ratios of magnesium to iron. Partial melting (specifically, the type of partial melting called fractional fusion) is something like fractional crystallization in reverse: The first liquid to form during partial melting has the same composition as the last liquid to crystallize during fractional crystallization.

The different groups of eucrites can be resolved on diagrams where the abundance of an incompatible element (one that has a chemical preference to remain in the liquid state rather than to go into the lattice of a crystallizing solid) is plotted against the molar ratio of magnesium-oxide to magnesium-oxide plus iron-oxide, or MgO/(MgO + FeO).

This ratio, known as the mg-number, changes during crystallization: Early formed crystals such as olivine tend to be rich in magnesium (they have high mg-numbers); crystals formed after much of the melt has been crystallized tend to be rich in iron. (Yhey have low mg-numbers.)

Diagrams, such as Figure 6, that plot the abundance of incompatible elements in eucrites show two main trends. The Stannern trend, named after the eucrite found in Stannern, Czech Republic, is thought to represent basalts formed by partial melting of chondritic material and subsequent crystallization on the parent asteroid. The Nuevo Laredo trend, named after the eucrite found in Nuevo Laredo, Mexico, is thought to represent basalts formed by fractional crystallization of Stannern-like melts.

The basaltic clasts in mesosiderites plot together with the Nuevo-Laredotrend eucrites, indicating that they also formed by fractional crystallization processes. This was the second major event in the history of the mesosiderite asteroid, it occurred within 30 million years of accretion, and was the principal igneous process that formed the asteroid's original basaltic crust.

Crustal Remelting


The next major event in the evolutionary history of the mesosiderite parent asteroid is revealed by studies of coarse-grained igneous rocks called gabbros. These rocks constitute about 39 percent of the large silicate clasts in mesosiderites. The gabbroic clasts are similar in mineralogy to the basaltic clasts but have larger crystals. Many areas within the clasts contain threecrystal sets that meet to form 120-degree triple junctures. This texture is commonly found among rocks whose crystals settled through a melt and accumulated at the base of a magma chamber. The gabbros contain very little intercrystalline material, which represents the last liquid to crystallize (and is thus characterized by enrichments in incompatible elements). As a result, the gabbros are highly depleted in incompatible elements.

Most of the rare-earth elements, those with atomic numbers 57 (lanthanum) through 71 (lutetium), are among the incompatible elements. (Atomic numbers refer to the number of protons in the nucleus of an atom.) The rare-earth elements form large ions that do not readily fit into the crystal structures of most minerals; hence, they tend to concentrate in residual liquids during crystallization. They are particularly well suited for studying igneous processes because most of them have the same valence of +3 (they are missing three electrons in their outermost subshells), and the major differences between their geochemical behaviors result from small differences in their ionic radii (which decrease systematically with atomic number; this is the so-called 'lanthanide contraction"). The major exception is europium, atomic number 63. A significant fraction of europium is divalent (a +2 valence). Divalent europium is more compatible than the other rare-earth elements because it is similar in size to the common, mineral-forming divalent ion of calcium and tends to substitute for it in the silicate plagioclase.

The relative abundance of rare-earth elements in an igneous rock provides a clue to the history of its crystallization, and this can be examined pictorially in Figure 8. The basis of comparison in these diagrams is the relative abundance of rare-earth elements in chondrites; these meteorites have the same composition of rare-earth elements as the sun (which is effectively the average composition of the solar system). Rocks with a composition of rare-earth elements matching those of chondrites have flat rare-earth element patterns. Because europium tends to partition into plagioclase, even minor enrichment or depletion of plagioclase in a rock may cause europium to be significantly enriched or depleted relative to the other rare-earth elements. The commensurate peak or valley on the rareearth element diagram is known as an europium anomaly.

Geochemical analyses of gabbro clasts in mesosiderites indicate that these rocks have extremely low concentrations of most rare-earth elements but undepleted concentrations of europium. Indeed, these clasts have the most extreme positive europium anomalies known among rocks in the solar system. Calculations indicate that the gabbros could not have formed by a single episode of crystal settling from a basaltic melt. Instead, they probably formed by a two-stage process involving crystal settling to form a rock depleted in incompatible elements, followed by remelting of this rock and the accumulation of crystals from this second-generation melt.

The large abundance of gabbro clasts in mesosiderites and the twostage heating mechanism required to form these clasts indicate that a major remelting of the crust took place. This is the third major event in mesosiderite history. Brian Stewart and co-workers determined the samarium-neodymium isotopic age of one of the gabbros to be 4.47+/- 0.15 billion years. They also dated a remelted basaltic clast at 4.51+/0.04 billion years. I provisionally take 4.5 billion years as the date of this event, but its absolute age is not well resolved from either accretion or differentiation.

At about the same time as the crust of the asteroid was remelting, there was some mixing of metals and silicates. This remelting and mixing event appears to have affected the oxidation state of mesosiderite minerals. During fractional crystallization of a basaltic magma, the iron/manganese (Fe/Mn) ratio of pyroxene remains fairly constant while the iron/magnesium (Fe/Mg) ratio increases significantly. However, in many igneous clasts formed during crustal remelting the trend is very different: Both Fe/Mn and Fe/Mg decrease. This trend is indicative of reduction of iron oxide during crystallization of pyroxene in the clasts. In other words, while the pyroxene grains were crystallizing, some of the iron oxide in the melt was transformed into metallic iron and thereby prevented from entering the pyroxene crystal lattice. The decrease in iron oxide resulted in lower Fe/Mn and Fe/Mg ratios in these igneous clasts. It seems likely that the reducing agent (which acquired the oxygen from the iron oxide and produced metallic iron) was phosphorus, which initially was associated with the abundant metallic iron-nickel in the mesosiderites.

This model assumes that metal-silicate mixing occurred at (or before) the time these igneous silicate clasts were crystallizing. It seems plausible to equate the timing of the metal-silicate mixing event and the crustal remelting event. The concomitant oxidation of phosphorus may account for the abundance of phosphate in mesosiderites (about 1 to 4 percent by volume) relative to that in eucrites (about 0.1 percent).

The nature of the metal involved in the metal-silicate mixing event has also been investigated. Analyses of large metal nodules in 15 mesosiderites found that their composition is similar to that of mean IIIAB iron meteorites, the largest iron meteorite group. (Iron meteorites can be chemically classified into about a dozen groups based on the abundance of germanium and gallium relative to that of nickel.) This was not too surprising because the ratios of oxygen isotopes in mesosiderite silicates are similar to those of the rare oxide inclusions in the IIIAB iron meteorites. Because oxygen isotopic ratios vary considerably among different meteorite groups, the similarity between mesosiderites and IIIAB iron meteorites probably indicates that their parent asteroids formed in the same general region of the solar system.

The IIIAB iron meteorites formed by fractional crystallization of metal in asteroid cores. Each of the 200 or so known IIIAB meteorites represents the product of a different degree of fractionation and hence has its own distinct inter-element ratios. For example, during fractional crystallization of metals, nickel has a slight preference for remaining in the liquid, whereas iridium is strongly partitioned into the solid the result is that the iridium/nickel ratios in the earliest-formed IIIAB iron meteorites are approximately 3,000 times greater than those of the latest formed members of the group.

In contrast, mesosiderite metal is far more uniform in composition. The iridium/nickel ratios in the metal nodules we analyzed vary by less than a factor of two (excluding one structurally and compositionally anomalous mesosiderite from Reckling Peak, Antarctica). Therefore, if the metal in mesosiderites was derived from a IIIAB-like core, then that core must still have been largely molten at the time metal-silicate mixing occurred because fractional crystallization could not yet have proceeded very far. This implies that the metal-silicate mixing event took place very early in the asteroid's history because the core would have become fully crystallized about 100 million years after it was formed. This conclusion is consistent with the approximate date of 4.5 billion years assigned to this event based on isotopic dating.

The simplest model for mixing metal and silicate is a collision between two differentiated asteroids. The reason this model had not gained wide acceptance in the past is that most terrestrial and lunar impacts suggest that projectile material is widely dispersed over the surface of the target asteroid. The soils at several lunar landing sites suggest that the ratio may be as little as 1 part projectile to 100 parts of the target body. The absence of mesosiderite analogues on the earth and the moon (even from craters known to have been formed by impacting iron meteorites, such as Meteor Crater in Arizona) is the result of the high relative velocities of impacting projectiles. During such collisions the projectile material is largely vaporized, and much of the remainder is diluted by vastly larger amounts of target ejecta.

In the present-day asteroid belt, the relative velocities are about 5 kilometers per second, too high to allow a projectile/target mass ratio of unity. However, in the early solar system, less than 100 million years after accretion, there were many bodies in low-inclination, nearly circular orbits-consequently, the average relative velocities among asteroids may have been less than 1 kilometer per second. It is precisely during this early period that cores of differentiated asteroids were still largely molten. It therefore seems plausible that the basaltic surface of the mesosiderite parent asteroid was struck at a low relative velocity by an asteroid having a largely molten core.

Heat generated by this collision was responsible for remelting the crust of the asteroid and the production of gabbroic rocks with huge europium/samarium ratios. Phosphorus that was associated with the metal acted as a reducing agent during silicate crystallization and was itself oxidized into phosphate. The rare coarse olivine grains in mesosiderites may have been acquired from fragments of the mantle overlying the core of the impacting asteroid. No comparable event involving metal-silicate mixing, oxidation-reduction or crustal remelting is evident among the eucrites, suggesting that they were derived from a separate parent asteroid.

Local Impact Melting


The fourth major epoch in the history of the mesosiderite asteroid was characterized by localized impact melting occurring between 4.5 and 3.9 billion years ago. Brian Stewart and his colleagues found that a rapidly cooled impact-melt rock clast from the Vaca Muerta mesosiderite has a samariumneodymium isotopic age of 4.42+/- 0.02 billion years. Similarly, Marc Brouxel and Mitsunobu Tatsumoto, of the U.S. Geological Survey in Denver, and other workers have found lead-lead, uranium-lead and rubidium-strontium ages in the 4.5- to 3.9-billion-year range in the silicates of a mesosiderite from Estherville, Iowa. Most of this material probably represents fragments of impact-melt rock from numerous localized impact events during this era.

These impact events dramatically affected the textures of individual mesosiderites, and they have been classified into four different groups based on increasing degrees of recrystallization of their silicates. Type-1 mesosiderites have experienced little recrystallization, whereas type-2 and type-3 mesosiderites exhibit increasing degrees of recrystallization, and type-4 mesosiderites have been impact-melted. Different degrees of impact heating were probably responsible for transforming type-1 mesosiderites into higher types. In addition, the extensive fragmentation and chaotic mixing of the disparate materials in mesosiderites may also have been caused by localized impact events at this time. Some mesosiderites (such as one from Mount Padbury, Australia) contain metal-rich, silicate-bearing fragments that must have existed at the surface of the mesosiderite parent asteroid as independent mesosiderites prior to incorporation into their present hosts. The isolated fragments were probably assembled into whole rocks by impact-induced localized melting at grain boundaries; chilling of these tiny pockets of melt cemented the jumble of materials together.

Disruption and Reassembly


The fifth major event in mesosiderite history may have been the collisional disruption and gravitational reassembly of the mesosiderite asteroid about 3.9 billion years ago. Theoretical studies of the collisional evolution of asteroids indicate that disruption and reassembly may have been relatively common among large asteroids. Indeed, other studies have suggested that some of the moons of Saturn and Uranus were also disrupted and reassembled.

The conclusion that the mesosiderite asteroid may have been disrupted and reassembled was reached by Donald Bogard and his colleagues at NASA's Johnson Space Center in Houston. These authors used the argon age-dating technique to determine the ages of 14 mesosiderites. This method is based on the release of different isotopes of argon gas from retentive sites in mineral lattices in rocks that are experimentally heated. Rocks that have experienced severe shock metamorphism at some point in their history will have lost all of their argon and started the argon-accumulation clock anew. The ratio of the different argon isotopes in a sample can be converted into a meaningful age because the rates of decay of their parental potassium isotopes are well known. Bogard and coworkers found that all of the mesosiderites they studied were thoroughly degassed by a major thermal event about 3.9 billion years ago, although several mesosiderites (such as those from Veramin, Iran and Lowicz, Poland) may have experienced additional degassing more recently (less than 3 billion years ago).

One scenario that would account for such a thermal event 3.9 billion years ago is a major collision that shattered the asteroid and heated the mesosiderites to about 500 degrees Celsius. Most of the fragments failed to reach escape velocity, however, and quickly reassembled themselves into a disorganized jumble. Because this event would have occurred after metal-silicate mixing, the tough mesosiderites retained their integrity and did not get diluted with major amounts of olivine from the disrupted asteroid mantle.

Not all researchers believe that the mesosiderite parent body was disrupted and reassembled. Henning Haack, of Odense University in Denmark, and his colleagues have interpreted the 3.4 to 3.8 billion-year argon ages as indicating that the parent body was so large (400 to 800 kilometers in diameter) that it took 600 million years (from 4.5 billion to 3.9 billion years ago) to cool sufficiently to allow argon gas to be retained in the mineral lattice sites. The very slow mesosiderite cooling rates of about 0.5 degrees Celsius per million years (achieved while they cooled through 500 degrees) implies significant burial depths (probably several tens of kilometers) beneath the asteroid surface. These cooling-rate measurements are based on the concentrations of nickel at the centers of metallic ironnickel grains, the width of the metal grains and a diffusion model of how fast nickel can move from grain boundaries to grain centers. If the disruption/reassembly event occurred, it would have heated the mesosiderites, buried them deeply and caused them to cool slowly. Alternatively, slow cooling could be the natural consequence of a very large mesosiderite parent asteroid that was never disrupted and reassembled. The issue has not been resolved.

Excavation and Ejection


The sixth, and most recent, epoch of mesosiderite history involves a series of processes that happened long after the parent body cooled. The mesosiderites were excavated from their burial sites by meteoroid collisions and deposited on the surface of the asteroid. Noble gas measurements by Friedrich Begemann, of the Max Planck Institute for Chemistry (Mainz, Germany), and co-workers indicate that the Veramin mesosiderite acquired solar-wind gases at this time. Noble gases such as helium, neon, argon and krypton are carried through interplanetary space by the solar wind, encountering porous rocks on airless bodies such as the Moon and the asteroids where some of the gas molecules become attached to the surfaces of grains. Enrichment of a rock in solar-wind gases is prima facie evidence that it was once exposed at the surface of an airless body.

Some time after their excavation, the mesosiderites were ejected from the surface of their parent asteroid and launched into interplanetary space by a series of impact events. While in space, the mesosiderites were bombarded with galactic cosmic rays, highenergy protons originating outside the solar system. The mean energy of these protons is about 10 billion electron volts, allowing them (and their secondary particles) to penetrate interplanetary objects to depths of about one meter. These particles are energetic enough to cause nuclear reactions in the materials they penetrate, producing both radioactive and stable nuclides. Measurements of the ratios of the cosmic-ray-produced radionuclides and their daughter isotopes allow the determination of the cosmic-ray exposure age of a meteorite. This is actually a measure of the duration of the meteorite's residence in interplanetary space as a meter-size (or smaller) object. Begemann and his colleagues found that the cosmic-ray exposure ages of mesosiderites range from 10 to 160 million years. The mesosiderites's journey through interplanetary space was eventually cut short by yet another collision-this time with the earth.

Conclusion


This scenario describing six major epochs in the history of the mesosiderites's parent body should serve as a framework for testing hypotheses and prompting other workers to gather additional information. It would not be surprising if future work were to revise certain aspects of this chronology because some of the dates are based on isotopic analyses of only one or two samples. Our geochemical analysis of these mesosiderites is merely the latest in a series of events that has befallen these denizens of the early solar system.