Dawn; the Vesta–HED connection; and the geologic context for eucrites, diogenites, and howardites



The Dawn mission has provided new evidence strengthening the identification of asteroid Vesta as the parent body of the howardite, eucrite, and diogenite (HED) meteorites. The evidence includes Vesta's petrologic complexity, detailed spectroscopic characteristics, unique space weathering, diagnostic geochemical abundances and neutron absorption characteristics, chronology of surface units and impact history, occurrence of exogenous carbonaceous chondritic materials in the regolith, and dimensions of the core, all of which are consistent with HED observations and constraints. Global mapping of the distributions of HED lithologies by Dawn cameras and spectrometers provides the missing geologic context for these meteorites, thereby allowing tests of petrogenetic models and increasing their scientific value.


Asteroid 4 Vesta was an especially enticing exploration target for the Dawn spacecraft mission, because it is generally thought to be the parent body for the most voluminous collection of achondrites. These are the howardite, eucrite, and diogenite (HED) meteorites, comprising >1000 named samples. However, the Vesta–HED linkage has not always been embraced (e.g., Schiller et al. 2011; Wasson 2012). The occurrence of many iron meteorite groups indicates that differentiated asteroids, perhaps similar to Vesta, were common in the early solar system, even though only their metallic cores have survived. Although a few other small asteroids with basaltic surfaces have been identified from their spectra, Vesta appears to be the only intact, differentiated, rocky protoplanet (Consolmagno and Drake 1977; Keil 2002; Russell et al. 2012). Data from Dawn instruments provide new tests for the proposed Vesta–HED connection.

Comprehensive reviews of the petrology, geochemistry, and geochronology of the HEDs abound, most recently by Keil (2002) and McSween et al. (2011). Here, we focus on those characteristics of HEDs that link them to a large, differentiated parent body or to Vesta specifically, and those characteristics of Vesta that link it to HEDs. Some of these connections derive from telescopic observations and from the meteorites themselves, and these will be briefly reviewed, but of particular interest here are new linkages based on observations and measurements by instruments on the Dawn spacecraft.

However useful the petrologic insights from these meteorites might be, they are still samples from unknown geologic settings. Dawn also provides an opportunity to fill in some of the missing geologic context, and thereby markedly increase the scientific value of these meteorites.

Vesta and HEDs—the astronomical perspective

McCord et al. (1970) first noted that the visible/near-infrared (VISNIR) reflectance spectrum of Vesta resembled the laboratory spectra of eucrites, and proposed that Vesta was the HED parent body. As spectral surveys of the asteroid belt advanced (e.g., Wisdom 1985), it became increasingly apparent that Vesta's spectrum was unique among the 500 or so largest asteroids. Moreover, the rarity of olivine-rich mantle samples among achondrites argued that the basaltic crust sampled as HEDs must still be intact somewhere in the Main Belt. Based on this argument (Consolmagno and Drake 1977), Vesta became further accepted as the HED parent body. Over many years, this hypothesis has been strengthened by subsequent spectral measurements of Vesta, including rotational observations using the Hubble Space Telescope that revealed regions where HEDs were thought to be concentrated (Binzel et al. 1997; Gaffey 1997; Pieters et al. 2006; Reddy et al. 2010). For decades, Vesta was thought to be spectrally unique—the only known candidate for the HED parent body. More recent discoveries of a limited number of other Main Belt asteroids with spectral characteristics similar to those of eucrites (Lazarro et al. 2000; Hardersen et al. 2004; Sunshine et al. 2004; Roig et al. 2008) now allow the possibility of other sources for some HEDs.

Although the spectral link between Vesta and HEDs was initially persuasive, Vesta was recognized to be in an orbit that made transferring its impact debris into Jovian or secular resonances, and thence into Earth-crossing orbits, difficult (Wetherill et al. 1987). This problem was surmounted by the discovery of numerous Vesta family asteroids (“Vestoids”), generally <10 km in size, having HED-like spectra (Binzel and Xu 1993; Vilas et al. 2000; Burbine et al. 2001; De Sanctis et al. 2011; Mayne et al. 2011; Reddy et al. 2011). The orbits of the Vestoids extend between that of Vesta and the Jovian 3:1 and ν6 resonances, demonstrating a migration path from Vesta to nearby escape hatches (Wisdom 1985). The Vestoids' link to Vesta is seemingly irrefutable as they reside tightly in the same orbital inclination and eccentricity planes as Vesta itself. (These planes are largely insensitive to drifting forces such as Yarkovsky, which primarily migrates objects in semimajor axis orbital distance.) A number of “fugitives” that have orbits outside the boundaries of the Vestoid family appear to have evolved dynamically away, enlarging the original number of these bodies (Nesvorný et al. 2008). Overall, more than 13,000 asteroids are assigned as being dynamically linked to Vesta, amounting to more than 10% of all asteroids with orbital semimajor axes between 2.1 and 2.5 AU (Nesvorný 2010). Because of the likelihood of Vestan samples reaching Earth through migration of this abundance of Vestoids, or dislodged chips of them, into nearby resonances, it seems probable that Vesta really is the parent body of most HEDs. A number of small HED-like bodies have also been found in Earth-approaching orbits (Cruikshank et al. 1991; Migliorini et al. 1997).

Vesta and HEDS—The meteorite perspective

Basaltic eucrites are basalts; cumulate eucrites are gabbros; and diogenites are orthopyroxenites, harzburgites, or dunites (cumulates of orthopyroxene and/or olivine). Most HEDs are monomict, dimict, or polymict breccias. Polymict eucrites contain up to 10% noneucritic material (Delaney et al. 1983), dimict diogenites are mixtures of orthopyroxenite and harzburgite (Beck and McSween 2010), and howardites are breccias composed of eucrite and diogenite (Beck et al. 2012). Thermal metamorphism has also affected most HEDs (Yamaguchi et al. 1997). The petrologic and geochemical properties of the meteorites themselves argue for a large parent body, although not specifically Vesta.

The HED parent body experienced complex igneous processes that are not yet fully understood. In fact, the petrogenetic relationship between eucrites and diogenites is still contested (see McSween et al. 2011). It is agreed, however, that the various HED lithologies formed on the same body; the incorporation of both eucrite and diogenite clasts into howardites, in concert with their identical oxygen isotopic fingerprints (discussed below), confirms their common parentage.

The early differentiation of the HED parent body is indicated by analyses of the decay products of short-lived radionuclides. Measurements of 26Mg*, the daughter isotope of 26Al, in diogenites (Schiller et al. 2011) argue for rapid formation of this body within approximately 1 Ma after CAIs, and the former existence of 26Al in eucrites has also been documented (Srinivasan et al. 1999; Nyquist et al. 2003). A 53Mn-53Cr isochron for eucrites has been interpreted as indicating a crust-mantle differentiation age of 2.2 ± 1.1 Ma after CAIs (Trinquier et al. 2008), and 53Mn systematics of diogenites indicate formation within 2–3 Ma after CAIs (Day et al. 2012). The 182Hf-182W isochron for HEDs gives a slightly later but less constrained time for core formation as 3 ± 6 Ma (Kleine et al. 2009).

Schiller et al. (2011) suggested that the very rapid solidification and cooling of the HED parent body (within 4 Ma after CAIs), which they inferred from 26Al systematics of diogenites and eucrites, are inconsistent with their origin on an asteroid as large as Vesta. Most thermal evolution models for Vesta suggest timescales >10 Ma for igneous crystallization and cooling through the blocking temperatures for some radiometric chronometers (Ghosh and McSween 1998; Gupta and Sahijpal 2010), and 182Hf-182W ages of zircons in eucrites define a range of crystallization ages spanning 45 Ma (Roszjar et al. 2012). The younger, concordant 147Sm-143Nd, 207Pb-206Pb, and sometimes 87Rb-87Sr ages for the plutonic HEDs (cumulate eucrites and diogenites) also support a protracted magmatic or cooling history for the body (Takahashi and Masuda 1990; Tera et al. 1997), although rapid global differentiation is allowed.

Nevertheless, rapid differentiation of the HED parent body certainly occurred, and is consistent with the idea of pervasive melting. Although a Vestan magma ocean has often been advocated (Righter and Drake 1997, 2001; Ruzicka et al. 1997; Takeda 1997; Warren 1997; Greenwood et al. 2005; Schiller et al. 2011), all the members of the HED suite cannot be unambiguously explained by magma crystallization on a global scale. The trace element abundances of diogenites, in particular, are difficult to explain as products of a magma ocean, and appear to require multiple parent magmas having different compositions (Mittlefehldt 1994; Shearer et al. 1997; Barrat et al. 2008; Beck and McSween 2010; Mittlefehldt et al. 2012a). Petrogenetic models for eucrites are not simple either, and must account for two distinct groups—the so-called Nuevo Laredo (or main group) and Stannern trends (Stolper 1977). The Neuvo Laredo group eucrites have been variously interpreted as primary partial melts or as fractionated magmas (summarized by McSween et al. 2011), and Barrat et al. (2007) modeled the Stannern group magmas as hybrids formed by comingling of Nuevo Laredo magmas with anatectic crustal melts. A possible compromise solution is suggested by models that posit 60–70% equilibrium crystallization of an early magma ocean, followed by extraction of melt into magma chambers that undergo fractional crystallization and periodic eruption (e.g., Mandler and Elkins-Tanton 2013). Alternatively, Wilson and Keil (2012) developed asteroid models that indicate very efficient removal of melts from the mantle, so that only a small amount of melt is present at any one time, probably precluding a magma ocean. Moreover, comparison of magma flow rates and stresses needed to keep fractures open support the formation of subsurface magma chambers that then erupt episodically. Both models thus converge on the necessity of plutons as the immediate sources of diogenites and eucrites. Regardless of the correct details, it is clear that the compositions of eucrites and diogenites reflect complex magmatic processes.

Bodies that have experienced extensive or near-total melting and differentiation might be expected to have homogenized oxygen isotopes (although this would require overcoming lid effects and impact mixing of residual solids). The Δ17O measurements for HEDs (Wiechert et al. 2004; Greenwood et al. 2005; Day et al. 2012), summarized in Fig. 1, define slightly different average values for the various laboratories, presumably due to different methods of correcting data to the internationally accepted standard. Regardless of which data set is considered, however, almost all eucrites and diogenites lie along a common oxygen isotope mass-fractionation line. Howardites (Wiechert et al. 2004) show some small deviations, attributable to foreign chondritic inclusions. The oxygen isotope data have been interpreted differently, depending on how much emphasis is placed on the few anomalous samples. Wiechert et al. (2004) thought that eucrites deviating from the mass-fractionation line reflected an isotopically heterogeneous, and hence incompletely melted parent body. Greenwood et al. (2005) argued that the isotopic uniformity of most HEDs was evidence for global melting, with aberrant eucrites resulting from projectile contamination. Scott et al. (2009) and Day et al. (2012) explained the few eucrites and diogenites, respectively, with distinct Δ17O as samples of parent bodies other than Vesta. Eucrites with a non-Vestan parentage might also be expected to have different oxidation states, as reflected in pyroxenes with distinct Fe/Mn ratios (Papike et al. 2003). However, most eucrites with distinctive oxygen isotopic compositions have Fe/Mn ratios like the bulk of HEDs, whereas at least one eucrite with a distinctive pyroxene Fe/Mn ratio has normal Δ17O (Mayne et al. 2009). Although basaltic eucrites probably sample more than one parent body, it is not clear which geochemical criteria are definitive.

Figure 1.

Oxygen isotopic compositions of HED meteorites. Black symbols and left-hand vertical line (average) from Greenwood et al. (2005), open symbols and right-hand line from Wiechert et al. (2004), gray diogenite symbols from Day et al. (2012). Isotopically anomalous eucrites are identified by name. Differences in average values are attributed to problems with interlaboratory calibration.

Wasson (2012) noted that the oxygen isotopic compositions of silicate inclusions in the largest group of magmatic iron meteorites (IIIAB irons) and HEDs are consistent with derivation from the same asteroid. If this were correct, Vesta could not be the parent body as it still has an intact and presumably unsampled central core (see below). However, there are a number of examples of multiple bodies having the same oxygen isotopic composition, e.g., Earth, Moon, enstatite chondrite, and aubrite parent bodies. Moreover, we might expect to see iron asteroids as part of the Vestoid dynamical family, which has not been observed. Even more damaging is the “olivine mantle problem.” Volumetrically, the olivine-rich mantle of a disrupted asteroid should dominate the basalt both among observed asteroid fragments and meteorites. Yet, no surviving olivine-rich fragments are seen within the Vesta family and remain exceedingly rare among HEDs (Beck et al. 2011a). The argument that the HED parent body is intact (Consolmagno and Drake 1977) remains firm.

Variations in the isotopic composition of chromium (ε53Cr), the decay product of short-lived 53Mn, were established very early in solar system history and reflect initial heterogeneities in the 53Cr/52Cr ratio and/or the Mn/Cr ratio. Lugmair and Shukolyukov (1998) noted that measured ε53Cr values for the Earth, Moon, Mars (SNC meteorites), and Vesta (HEDs) showed a linear increase with heliocentric distance (Fig. 2). If this parameter is a true measure of body location, it provides further evidence linking the HEDs to Vesta. It would be a remarkable coincidence if the HEDs were derived from another differentiated asteroid (Wasson 2012), now destroyed, that happened to orbit at the same heliocentric distance as Vesta.

Figure 2.

Variation in chromium isotopic composition, expressed as deviation of 53Cr/52Cr from a standard (ε53Cr), with heliocentric distance, assuming SNC = Mars and HED = Vesta. After Lugmair and Shukolyukov (1998).

In summary, the eucrites and diogenites formed through magmatic processes that have not yet been fully understood but are clearly complex. These igneous rocks were later pulverized and mixed by impacts, producing breccias that sampled the asteroid's different materials. Radiogenic isotopes reveal early magmatism that may have extended for a few tens of millions of years. Such magmatic, petrologic, and chronologic diversity might be expected for a large asteroid like Vesta, but would be difficult to reconcile with smaller bodies. The proposed relationship between ε53Cr and heliocentric distance directly links the HEDs to Vesta's orbital position.

Vesta and HEDs—Dawn's perspective

The Dawn mission has provided new observations and measurements that link HEDs specifically to Vesta. Three instruments provided compositional data during mapping from orbit (Russell et al. 2012). The Visible and Infrared Spectrometer (VIR) measured reflectance spectra between 0.25 and 5.1 μm, allowing surface mineral identification using diagnostic absorption bands (De Sanctis et al. 2012b). The Framing Camera (FC) provided geologic context for VIR spectroscopy, limited (seven color filters between 0.44 and 0.98 μm) spectral measurements for geologic mapping, and images used to construct a topographic map (Sierks et al. 2011). The Gamma Ray and Neutron Detector (GRaND) determined the abundances or detection limits for specific elements and element ratios, including H, Fe, Fe/O, Fe/Si, and K, with a large spatial footprint and to a depth of several decimeters (Prettyman et al. 2011). In addition, radiotracking of the spacecraft's orbit provided data that were used to model Vesta's density, gravity field, and interior structure.

Petrologic Complexity

A false-color map of color ratios on Vesta (Fig. 3), compiled from FC multispectral images (Reddy et al. 2012a), reveals considerable surface compositional diversity. The various units differ in albedo, band depth, spectral slope, and inferred abundances of eucrite versus diogenite. The latter are based on the 0.98/0.92 μm band depth ratio, which qualitatively identifies eucrite- and diogenite-rich terrains (Reddy et al. 2012a). Pyroxenes in eucrites have more ferroan compositions than those in diogenites, causing this ratio to shift toward longer wavelength.

Figure 3.

Color ratio map of Vesta showing the ratio of the 0.98/0.92 μm filters of the Dawn Framing Camera obtained during the High Altitude Mapping Orbit (HAMO) at a resolution of approximately 60 m per pixel. The areas in red are more diogenite-rich and those in blue are eucrite-rich. The map uses the Claudia coordinate system that is different from the IAU coordinate system; the zero longitude in the IAU system is offset by approximately 212° from that of Claudia.

Some minor spectral variations may be attributed to differences in space weathering, rather than composition, as discussed below. Nevertheless, the remarkable compositional diversity on Vesta's surface is unlike the spectrally bland surfaces of all smaller asteroids previously visited by spacecraft, and is consistent with the petrologic diversity among HEDs.


The diversity of Vesta's surface in terms of mineral proportions and compositions has been demonstrated by VIR spectra (De Sanctis et al. 2012a, 2013). The center positions of the 1 μm (BI) and 2 μm (BII) bands are characteristic of pyroxenes in eucrites and diogenites. The different HED lithologies can be distinguished using these band center positions (Beck et al. 2011b), although there are some overlaps between howardites and the other lithologies, reflecting different mixing ratios (Fig. 4). A cloud plot of Vesta's global spectra (Fig. 4) shows that the highest proportion of pixels corresponds to howardite, with a more limited number of pixels falling within the eucrite or diogenite fields (De Sanctis et al. 2012a, 2013). VIR spectra indicate that diogenite is concentrated within and around the Rheasilvia basin, and diogenite-rich howardite extends farther northward as a result of its impact excavation (McSween et al. 2013).

Figure 4.

Cloud representing VIR band center positions for Vesta superimposed on a spectral classification diagram for HEDs (De Sanctis et al. 2012a, 2013). The mineralogic similarities inferred from this diagram indicate that the Vestan surface compositions are consistent with HEDs. The boxes are defined by spectrally well-characterized HEDs, each having at least 30 spectra measurements.

There is no reason, a priori, why the relative proportions of HEDs should mimic the proportions of Vesta's surface covered by the different lithologies, but a comparison might reveal sampling biases. The relative proportions of HEDs recovered from Antarctica by US expeditions during the period 1977–2010, distinguished by mass and by number of samples, are shown in Fig. 5. Direct comparison of these meteorite abundances with the Vestan surface (De Sanctis et al. 2013) is problematic because, as previously noted, Vesta's regolith at the VIR pixel scale reveals the preponderance of howardites, i.e., mixtures of lithologies. (VIR resolution was 70 m per pixel at closest approach, but was 800 m per pixel at the altitude where mapping occurred.) Nevertheless, if the Vestan howarditic surface is subdivided into regions rich in eucrite or diogenite (Fig. 4), eucrites dominate, as they do among the HEDs (Fig. 5).

Figure 5.

Relative proportions of HED lithologies in the US Antarctic collection, broken down by number of samples and by mass.

Although a subset of diogenites contains significant amounts of olivine (Beck and McSween 2010; Beck et al. 2011a), no spectral detections of olivine have yet been made within the huge Rheasilvia basin (McSween et al. 2013). This apparent lack of olivine probably results from the difficulty in spectrally distinguishing olivine in concentrations <30% in the presence of abundant orthopyroxene (Beck et al. 2013).


A comparison of global Fe/Si versus Fe/O ratios (Fig. 6), determined from GRaND γ-ray spectra of Vesta, indicates consistency with the compositions of HEDs, but rules out all chondrites, stony-irons, and most non–HED classes of achondrite (Prettyman et al. 2012). Although some angrites and ureilites overlap HEDs on this plot, their VIR spectra would make them readily distinguishable.

Figure 6.

Global Fe/Si and Fe/O weight ratios for Vesta, determined from GRaND measurements (Pieters et al. 2012), compared with those ratios analyzed in HEDs. The 1σ and 2σ ovals indicate analytical uncertainties in the GRaND element ratios.

Variations in neutron counting rates due to absorption by certain elements measured by GRaND reflect differences in the abundance of Fe, Ca, Al, Ti, and Mg on Vesta (Prettyman et al. 2011, 2013). The Rheasilvia basin and its ejecta blanket show lower neutron absorption than the Vesta average, supporting the VIR and FC conclusions that it contains significantly more diogenite (Prettyman et al. 2012; McSween et al. 2013). Areas on Vesta farther from Rheasilvia have neutron counting rates appropriate for howardites and eucrites.

Geochemical analyses of HEDs indicate chondritic refractory element abundances, but with strong depletions in volatile elements (Warren 1983). Consequently, a high Vestan abundance of K would argue against the Vesta-HED connection (Wasson 2012). The global surface K abundance (<1000 μg g−1, below the 3σ detection limit) determined by GRaND (Prettyman et al. 2012, supporting information) is consistent with HED compositions (e.g., the average K content of howardites is 230 μg g−1).

Distinctive Space Weathering

Processes in the space environment that weaken or mask diagnostic spectral absorption features are collectively referred to as space weathering. Solar-wind irradiation and micrometeorite bombardment on the Moon (Keller and McKay 1997; Pieters et al. 2000) and on the chondritic asteroid Itokawa (Noguchi et al. 2011) have produced nanophase particles of iron metal or iron sulfide that accumulated on the rims of mineral grains. The spectral consequence is a correlation between band strength and albedo, but without modification in the continuum slope across the near-infrared (0.7–1.5 μm) region.

Pieters et al. (2012) reported that Vesta exhibits a form of space weathering that is distinct from other airless bodies. No VIR spectral indication attributable to the accumulation of nanophase iron (npFe0) is observed, although the spectral properties of fresh craters on Vesta do change over time. The inhibition of npFe0 production on Vesta, relative to the Moon and Itokawa, presumably results from some combination of the following (Gaffey 2010; Pieters et al. 2012): (1) Vesta receives a lower solar flux; (2) impact velocities in the asteroid belt are lower; (3) native iron and FeS that could be mobilized are almost absent on Vesta; and (4) pyroxene, which dominates Vesta's surface, is more resistant than olivine to the formation of npFe0 (Vernazza et al. 2009). A crustal remanent magnetic field, inferred from measurement of one eucrite (Fu et al. 2012), might also have shielded the solar wind ion flux, limiting space weathering.

The lack of space weathering on Vesta has been used as an argument against its connection to HEDs (Wasson 2012). However, this unusual spectral characteristic of Vesta's surface is fully consistent with the rarity of nanophase iron particles in howardites (Noble et al. 2010) and so is actually evidence for the Vesta–HED connection.

Protracted Impact History and Petrologic Mixing

Although monomict breccias are common among HEDs, many eucrites, some diogenites, and all howardites are mixtures of multiple lithologies. The preponderance of breccias and the degree of petrologic mixing among the HEDs indicate significant gardening of surface materials. The megaregolith of Vesta is estimated to be perhaps 1–2 km thick (Jaumann et al. 2012), requiring a protracted impact history, as documented by Vesta's crater density (Marchi et al. 2012). This is, of course, a qualitative argument, as all asteroids are heavily cratered. However, the chronology of substantial impacts on the HED parent body has been determined from 40Ar-39Ar ages of eucrite breccias (diogenites contain so little K that they produce negligible radiogenic Ar). These ages (Fig. 7, Bogard 2011) record a series of events interpreted to indicate an increase in impactor flux starting near 4.1 Ga and defining the Late Heavy Bombardment (Marchi et al. 2013). The 40Ar-39Ar ages are set during cooling through the blocking temperature within ejecta blankets. A disrupted asteroid, like that advocated for the HED parent body (Wasson 2012), could not have recorded impacts during the Late Heavy Bombardment period, unless its disruption occurred after approximately 3.4 Ga. The 40Ar-39Ar ages of eucrites (Fig. 7), however, are consistent with a Vestan origin, as these rocks were presumably not ejected until the Rheasilvia impact event at approximately 1 Ga (Russell et al. 2012). The ejection event would not have reset these ages, as the ejected meteorites were not buried within a hot ejecta blanket (Marchi et al. 2013).

Figure 7.

Probability plot of 40Ar/39Ar ages of unbrecciated and brecciated eucrites, modified from Bogard (2011).

Distinctive Chondritic Materials in the Regolith

Clasts of dark carbonaceous chondrite (mostly CM and some CR) occur in many howardites and some polymict eucrites (Zolensky et al. 1996; Gounelle et al. 2003; Lorenz et al. 2007). The volume of chondritic debris in these breccias, estimated from their augmented bulk Ni contents, is generally 0–4% (Warren et al. 2009). However, one unusual howardite (PRA 04401) contains up to 60% carbonaceous chondrite clasts (Herrin et al. 2011), illustrating that very high proportions of chondrite debris can occur locally.

Broad regions of low albedo on Vesta's surface (McCord et al. 2012; Reddy et al. 2012b) are larger manifestations of this contamination by exogeneous carbonaceous materials. These regions have high H abundances (>400 μg g−1), as analyzed by GRaND (Prettyman et al. 2012). Hydrogen at this high concentration cannot be explained by solar-wind implantation, but is consistent with the incorporation of hydrous phyllosilicates in carbonaceous chondrites. Assuming that the accreted material contains equivalent H to that in CM and CR chondrites, the proportion of carbonaceous material in the Vestan dark regions would be comparable to a few percent (Prettyman et al. 2012), as observed in howardites. Laboratory spectra of mixtures of powdered eucrite containing 1–6% CM chondrite provide an excellent match to the spectra of the dark regions on Vesta (Reddy et al. 2012b). This interpretation is further bolstered by the detection of a 2.8 μm mineral OH absorption feature on Vesta by VIR (De Sanctis et al. 2012b), and by the discovery on Vesta of pitted terrains thought to represent volatilization of OH-bearing materials during impact (Denevi et al. 2012). The dark materials in the Vestan regolith, now confidently interpreted to be exogenous carbonaceous chondrites, provide a strong link with howardites.

Absolute Ages of Surface Units

The absolute ages of Vestan surfaces, as determined from counting crater densities, are somewhat controversial and models differ markedly for Vesta's ancient terrains. This difference stems not so much from varying crater counts as from contrasting methods. One group (Schmedemann et al. 2012) extrapolates the crater production rates for the Moon to the asteroid belt. This method, however, violates constraints such as the number of observed large craters on Vesta, the early dynamical evolution of the Main Belt, and the pattern of eucrite 40Ar-39Ar ages (O'Brien et al. 2012). If application of the lunar chronology function is questionable, then another chronology function must be devised. The function chosen is based on collisional evolution models of asteroids in the Main Belt. This method gives a crustal age of approximately 4.5 Ga based on the abundance of large craters (>100 km diameter) and younger ages between 3 and 4 Ga based on smaller craters (O'Brien et al. 2012). Many Vestan terrains are saturated with craters, so the 3–4 Ga ages must be considered minimum ages.

The crystallization ages of HEDs, described earlier, not only indicate rapid melting and differentiation, but also suggest continued magmatism and a protracted cooling history for perhaps 50 Ma. This is consistent with some thermal evolution models for Vesta (Ghosh and McSween 1998; Gupta and Sahijpal 2010), but not with others (Schiller et al. 2011). The radiometric ages are older than the Vestan surface areas, as expected for crater-saturated terrains.

Ages of Vestan Basins and Vestoids

The most probable age for the Rheasilvia basin is 1.0 ± 0.2 Ga (Marchi et al. 2012). We previously noted the likelihood that most Vestoids were ejected in the Rheasilvia impact event, and that the Vestoids were the immediate parent objects of the HEDs. Marzari et al. (1996, 1999) showed that if the Vestoid family was older than approximately 1 Ga, its observed steep size distribution would have been ground down to match the background asteroid population. The tight clustering in orbital elements of the Vestoids supports this conclusion, as they would have been dispersed in a much earlier asteroid belt subject to the influence of planetary embryos and giant planet migrations (Noguchi et al. 2011). The fugitive V-type objects that orbit outside the Vesta family (Nesvorný et al. 2008) required at least 1 Ga to migrate away from the family, so it is possible that they were ejected in an older (≥2 Ga) cratering event that produced the Veneneia basin that underlies Rheasilvia (Russell et al. 2012). Thus, there is a chronologic link between impact events on Vesta and the Vestoids that are thought to have eventually supplied the HEDs to nearby escape hatches from the Main Belt.

Dimensions of the Core

Depletion of siderophile trace elements in eucrites (Newsom and Drake 1982; Righter and Drake 1997) and the former occurrence of a dynamo magnetic field, as revealed by remanent magnetism in an eucrite (Fu et al. 2012), substantiate that the parent body has an iron core. Models of the HED parent body constructed from meteorite compositions (Righter and Drake 1997; Ruzicka et al. 1997) suggested cores with mass fractions of 15–20%. These estimates compare favorably with constraints on the mass fraction (approximately 18%) and radius (110 ± 3 km) of Vesta's core that fit Dawn's measurement of the gravitational moment J2, using appropriate constraints on core and bulk silicate densities (Russell et al. 2012). One caution is that Vesta's calculated core dimensions assume an iron-nickel composition with no density correction for sulfur, although the core size is more dependent on the density of the silicate than of the core.

The geologic context for HEDs

Vesta's giant impact basin Rheasilvia (Schenk et al. 2012) was discovered by Hubble Space Telescope images and immediately implicated as the presumed source for the Vestoids (Thomas et al. 1997), fragments of which probably were injected into nearby resonances to become the HEDs (Migliorini et al. 1997). As a consequence, most HEDs were probably excavated from a terrain that only exists as the basin out of which they were carved. Nevertheless, eucrites, diogenites, and howardites still comprise other parts of the surface of Vesta, and images and spectral maps of Rheasilvia and of the rest of Vesta reveal their geologic settings and provide insights into their formation.

Eucrites and the Crust

Models of volcanism on Vesta predict the dimensions of surface flows and feeder dikes and suggest the possibility of pyroclastic materials (Wilson and Keil 2012). Although some unbrecciated basaltic eucrites show fine-grained textures indicative of solidification in lava flows (Mayne et al. 2009), no flow units or pyroclastic deposits have been recognized in Vestan images (Jaumann et al. 2012). This is not surprising, given the preponderance of eucrite breccias and the ubiquitous regolith cover on Vesta. Nevertheless, some regions on Vesta, especially near the equator and in parts of the northern hemisphere, exhibit the FC spectra properties of eucrite (Fig. 3). This distribution is confirmed by maps of the global distribution of HED lithologies based on VIR spectra and inferred from GRaND neutron counting rates (Prettyman et al. 2013). The regolith in these areas, all located away from Rheasiliva, has not undergone substantial admixture of diogenite debris.

Only a portion of the eucrite “box” in Fig. 4 is actually populated with global VIR pixels (De Sanctis et al. 2013). Part of the Vestan portion of the eucrite box corresponds to the band center positions for cumulate or polymict eucrites, but their band centers are indistinguishable from those of howardites (Fig. 4). Most basaltic eucrites plot in the portion of the box unpopulated by Vestan data (Fig. 4). An abundance of cumulate eucrites is understandable if much of the asteroid's surface has been covered by Rheasilvia ejecta, as suggested by the model of Jutzi and Asphaug (2011). With a crustal thickness of approximately 20 km (see constraints in McSween et al. 2013) and assuming that the upper 2 km of the crust was basaltic flows, the volume of excavated plutonic rocks (cumulate eucrite) should be 10 times greater than basaltic eucrite. If basaltic eucrite debris was mixed with cumulate eucrite to form polymict eucrite or with diogenite to form howardite, it would become spectrally unrecognizable. This explanation does not seem consistent, however, with the high abundance (63%) of basaltic eucrites relative to cumulate (3%) and polymict (34%) eucrites among HEDs in the US Antarctic collection. Chemical mixing relationships also suggest that basaltic eucrite is a more abundant component than cumulate eucrite in howardites (McSween et al. 2011; Mittlefehldt et al. 2012b; Usui and Iwamori 2013). However, the compositions of pyroxene fragments in howardites do show that cumulate eucrites are present (Labotka and Papike 1980; Fuhrman and Papike 1981). Taken together, these observations may suggest that the portion of Vesta's surface falling within the howardite box is actually dominated by howardite and that cumulate eucrite was not so abundant, at least in the Rheasilvia region.

Diogenites and the Mantle

The Rheasilvia basin exposes materials interpreted as diogenite at the base of its central uplift (corresponding to the lowest points on the crater floor) and within some portions of the crater wall (Reddy et al. 2012a; McSween et al. 2013), and diogenite-rich howardite comprises the widespread ejecta blanket around the basin. Diogenite is identified both by its VIR spectra and by GRaND neutron counting rates (McSween et al. 2013). Estimates of the excavation depth of Rheasilivia (30–45 km, Jutzi and Asphaug 2011) are twice the nominal thickness of the Vestan crust (15–20 km, Righter and Drake 1997; Toplis et al. 2013). Although a model of Ruzicka et al. (1997) suggests a thicker crust, it is based on an improbable (enstatite chondrite) bulk composition for Vesta. The Rheasilvia impact, when coupled with earlier excavation of the same area by the underlying 400 km-diameter Veneneia basin (Schenk et al. 2012), must have exposed materials from the upper mantle. The coarse-grained textures of diogenites, coupled with estimates of their slow cooling rates (Zema et al. 1997), indicate their formation as deep-seated rocks, although no quantitative indicators of depth are available.

Similar to the eucrites, the diogenite “box” in Fig. 4 is only partially populated by VIR pixels (De Sanctis et al. 2013). The explanation for this observation is unclear, but it cannot reflect different proportions of orthopyroxene and olivine. As significant amounts of olivine are added to orthopyroxene to form harzburgitic diogenite, the BI center would shift to longer wavelength, but the BII center would remain unchanged (olivine has a slightly longer wavelength 1 μm band than pyroxene, but no 2 μm band). The diogenites in Fig. 4 show covariation in their BI and BII center positions, presumably reflecting differences in mineral compositions rather than in mineral proportions.

The occurrences of diogenite deduced from Dawn's maps are fully consistent with petrologic constraints, but do not provide unambiguous support for any of the proposed modes for diogenite formation. Magma ocean models (Righter and Drake 1997; Ruzicka et al. 1997; Warren 1997; Mandler and Elkins-Tanton 2013) predict a thick, cumulate diogenite layer below a basaltic crust, whereas serial magmatism models (Mittlefehldt 1994; Shearer et al. 1997; Barrat et al. 2008; Beck and McSween 2010; Wilson and Keil 2012) predict emplacement of diogenite-bearing plutons near the crust-mantle boundary or within the lower crust. The depth to diogenite (approximately 20 km), as exposed on the floor of Rheasilvia, is consistent with magma ocean models based on some bulk asteroid compositions, but its lateral and vertical extents are not well constrained (McSween et al. 2013).

Howardites and the Megaregolith

Howardite comprises the regolith on Vesta. However, howardite is not the most abundant lithology among HEDs (Fig. 5). We might then infer that (1) the event(s) that excavated and launched HED meteorites primarily sampled subregolith materials, or (2) eucrites and diogenites occur within the regolith as blocks that are smaller than the 800 m per pixel resolution for VIR maps. Both of these explanations are likely to be correct. As noted earlier, the huge volume of materials excavated by Rheasilvia would have been dominated by subsurface igneous rocks, if we accept that the thickness of the regolith is generally <1–2 km (Jaumann et al. 2012). And incorporation of the various lithologies into the regolith is supported by the fact that HEDs comprise a continuum from unbrecciated eucrites, through monomict and polymict eucrite breccias, to howardites and dimict diogenite breccias, and finally to unbrecciated diogenites (McSween et al. 2011).

The regional diversity in the proportions of eucrite and diogenite in the Vestan regolith, so apparent in VIR spectral maps (De Sanctis et al. 2012a, 2013), is also reflected in the diverse mixing ratios in howardites (Warren et al. 2009). Considerable variations in the proportions of eucrite and diogenite occur even at the hand-sample scale, as illustrated by the paired PCA 02 meteorites (nine howardites and two diogenites) that collectively constitute the largest single sample of Vesta's regolith (Beck et al. 2012). The Vesta data clearly demonstrate that its regolith is not compositionally homogenized, although they may not rule out the possibility that the ancient regolith prior to the Rheasilvia impact may have been homogenized (Warren et al. 2009).

Impact melt clasts are common constituents of howardites (e.g., Olsen et al. 1990). Glass fragments and beads having unusual compositions (Barrat et al. 2009a; Beck et al. 2012) found in howardites appear to represent impact melts of distinctive materials not otherwise found as meteorites. However, agglutinates are very rare in howardites (Noble et al. 2010), indicating that impact melting is not as common a phenomenon on Vesta as on the Moon. The low apparent volumes of impact melt in fresh craters seen in FC images (Schenk et al. 2012) are consistent with models that imply limited melt production resulting from low impact velocities (<5 km s−1) in the asteroid belt (Bottke et al. 1994; Keil et al. 1997).

Unusual K-rich glasses in howardites have been suggested to be impact melts of a granitic or KREEP-like lithology not represented as meteorites (Barrat et al. 2009a). Small clasts of unusual low-Mg rock also have high K and other incompatible element abundances and might represent a lithology with KREEP affinities (Barrat et al. 2012). Singerling et al. (2013) suggested that the small K-rich glasses, as well as glasses rich in Ca and Al occurring in the same howardites, could be small-scale micrometeorite impact melts that preferentially sampled mesostasis or plagioclase in eucrites. The paucity of K in the global regolith (<1000 μg g−1, Prettyman et al. 2012, supporting information) and the lack of any detectable K concentrations do not support the idea that such highly fractionated rocks exist in any significant abundance.


Although there were numerous differentiated rocky protoplanets in the early solar system, as evidenced by the surviving remnants of their iron cores, most have been destroyed through impacts or swept up during the accretion of the terrestrial planets. Only Vesta remains as an intact representative of this population. Vesta's huge Rheasilvia basin and the resulting Vestoids demonstrate ejection of its crust and mantle materials, and the orbital positions of Vestoids near resonance escape hatches have allowed dislodged Vesta samples to reach the inner solar system and intersect the Earth.

The Dawn mission has provided new evidence to strengthen considerably the Vesta–HED connection:

  1. High-resolution VISNIR spectra obtained from orbit indicate pyroxene compositions and abundances like those of HEDs, confirming prior telescopic comparisons (Binzel 2012).
  2. The petrologic complexity mapped on Vesta is unlike the homogeneity of smaller asteroids, but consistent with the range of rock compositions among HEDs.
  3. Geochemical measurements of Vesta's regolith are similar to the compositions of HEDs and unlike those of other achondrite types.
  4. The unique space weathering characteristics of the Vestan surface, reflecting an absence of npFe0, are in agreement with the scarcity of npFe0 in howardites.
  5. The occurrence of a thick regolith of howardite on Vesta suggests protracted mixing by impacts, as demonstrated by pervasive breccias containing all the HED lithologies and evidence for numerous impact events from 40Ar-39Ar ages of eucrites.
  6. The ancient ages of Vestan surface units obtained from crater densities are 3 to 4 Ga. Because the surfaces are saturated with craters, these are minimum ages, and are thus consistent with older radiometric ages for crystallization of eucrite and diogenite magmas.
  7. Low-albedo regions on Vesta contain high abundances of H and exhibit an OH spectral feature, consistent with the proportions of dark, hydrated carbonaceous chondrite xenoliths in howardites.
  8. Geophysical constraints on the size of Vesta's core, based on shape and gravity models from orbital tracking, agree with size estimates from parent body models based on HED compositions.

Having more firmly established that Vesta is the HED parent body, we can make some conclusions about the geologic context for these meteorites. Although no volcanic flows are observed, broad regions dominated by eucrite occur. These show spectra more like cumulate or polymict eucrites than basaltic eucrites. Perhaps upper crustal (basaltic) rocks were ejected from Vesta more efficiently, accounting for their abundance among HEDs. The estimated excavation depth of Rheasilvia implies that the mantle was exposed, and diogenites occur at the lowest stratigraphic levels on the basin floor. Thus, mantle rocks appear to have been excavated and incorporated into an ejecta blanket covering half the body. This produced a globally heterogeneous regolith with varying proportions of eucrite and diogenite. Impact melts are not commonly observed in Vesta's craters, consistent with their relative scarcity in howardites and reflecting only modest collisional velocities in the asteroid belt. There is no discernble evidence on Vesta of lithologies not already represented as whole HED meteorites. Although observations of Vesta are consistent with HEDs, they do not yet distinguish between competing petrogenetic hypotheses, nor do they provide compelling evidence for or against an early magma ocean.


This work was funded by NASA's Discovery Program through a contract to the University of California, Los Angeles, by NASA's Dawn at Vesta Participating Scientists Program, by the Italian Space Agency, by the Max Planck Society and German Space Agency (DLR), and by the Planetary Science Institute under contract with the Jet Propulsion Laboratory, California Institute of Technology. We appreciate reviews by K. Keil, G. Consolmagno, and an anonymous reviewer.

Editorial Handling

Dr. David Mittlefehldt