Clues to the Origin of Jovian Outer Irregular Satellites from Reflectance Spectra

We first consider three satellites in the prograde cluster anchored by JVI Himalia: Himalia, JVII Elara, and JX Lysithea (Figure 1). Elara was observed on two nights. The spectrum of Himalia covers a larger wavelength range and has the best S/N compared to the spectra of the other two cluster members. Himalia's spectrum shows a well-defined feature centered near 0.69–0.70 μm covering a spectral range of ∼0.49 to 0.77–0.79 μm with a depth of ∼2%. This is consistent with most previous observations of Himalia (see Jarvis et al. 2000, and—as interpreted in Paper I—Degewij et al. 1980; Tholen & Zellner 1983; Grav et al. 2003). The feature's observed depth is comparable to the range observed in C-complex asteroid spectra of 1%–5% (Vilas et al. 1993b, 1994). The higher spectral resolution and greater wavelength coverage confirm the observable presence of this feature in broadband photometry in Paper I and negate the conclusion by Brown & Rhoden (2014)—based only on IR spectral observations—that there is no absorption feature centered near 0.7 μm present in Himalia's reflectance spectrum. This major absorption feature is shallower but spans a greater spectral range than that seen in 19 Fortuna (Figure 5). A small absorption feature from 0.40 to 0.44 μm is also present, at similar strength to the feature seen in the spectrum of 19 Fortuna. Jarvis et al. (2000) show three spectra of Himalia, each having a more limited wavelength range. Figure 6 compares their spectrum obtained on 1996 July 19 with our spectrum of Himalia, demonstrating the repeatability of the short-wavelength edge of the 0.7 μm feature. Future photometric observations determining rotational periods and orientation for all of these satellites would further address the question of large-scale surface compositional variations.

Zoom In Zoom Out Reset image size

The two spectra of JVII Elara also show absorption features centered near 0.7 μm, but they are slightly different. The 2006 spectrum shows a feature centered near 0.67–0.68 μm, covering a spectral range of 0.55–0.56 to ∼0.76 μm, having a depth of ∼2%. The 2009 spectrum shows a feature centered near 0.66–0.67 μm, with a shorter spectral edge beginning at 0.56 μm and extending longer than 0.8 μm, with an estimated similar depth of ∼2%. Both spectra show the shorter-wavelength edge of this feature shifting to slightly longer wavelengths than that observed in the spectrum acquired of Himalia in this study. The 2009 spectrum is slightly redder (reflectance increasing with increasing wavelength) than the 2006 spectrum. The UV/blue turnover is present in the 2009 spectrum, while it is much more subdued (if present) in the 2006 spectrum. These spectra potentially represent surface properties on different sides of Elara presented to the Earth. The differences in these two spectra are consistent with previous observations reported in Paper I, also suggesting surface variations (Degewij et al. 1980; Tholen & Zellner 1983). Bhatt et al. (2017) also observed a slight reddening for Elara compared to Himalia in the NIR.

The spectrum of the smallest of the three observed satellites, Lysithea (Figure 1), also shows an absorption feature centered near 0.64 μm with a shorter spectral edge beginning near 0.48 μm, a defined UV/blue absorption edge, an estimated similar depth of ∼2%, and a more reddened spectrum overall compared to Himalia and Elara. This increase in reddening was previously observed by Grav & Holman (2004) in their broadband photometry.

We next consider three satellites in the retrograde orbits with inclinations near 149° but varying jovicentric semimajor axes, including JVIII Pasiphae, JXII Ananke, and JXVII Callirrhoe (Figure 2). Ananke was observed on two nights. Pasiphae's spectrum has the highest S/N. It shows a shallow but observable feature centered near 0.69–0.70 μm covering a spectral range of 0.57–0.84 μm with an estimated depth of 2%. The spectrum is gray and also shows a UV/blue drop-off. An absorption feature spanning the wavelength range from 0.40 to 0.44 μm and centered near 0.43 μm is also present in the spectrum of Pasiphae. The 2009 Ananke spectrum shows a feature similar to the Pasiphae spectrum, with a lower wavelength edge near 0.57 μm but inadequate S/N to confirm the full spectral range. The shape and spectral placement of the absorption feature centered near 0.70 μm for Pasiphae and 2009 Ananke are very similar to those identified in the spectra of C-complex asteroids. These spectra are similar to that of 19 Fortuna, although the strength of the feature is reduced in the satellite spectra. The 2010 Ananke spectrum shows no sign of this feature, but an absorption feature centered near 0.75 μm and ranging spectrally from 0.66 to 0.84 μm is apparent. Callirrhoe has a very noisy and red spectrum, likely due to its apparent magnitude m_V = 20.09 when it was observed. The reddened spectrum was observed in broadband photometry of Callirrhoe (Grav et al. 2003, Graykowski & Jewitt 2018), but it was not so strongly reddened. We attribute this to an error with the narrowband spectrum here due to the difficulty of obtaining higher-resolution observations (dispersing the lower signal spread over a spectral range with finer granularity) of the object at this apparent magnitude, with no prejudice toward future interpretations. Figure 7 shows the scaled spectrum of Callirrhoe binned to a dispersion of 63 Å in order to reduce the scatter, with a linear background removed. The spectrum is presented to draw the reader's attention to the drop in reflectance from 0.54 to 0.60 μm, similar to that seen for Pasiphae and 2009 Ananke, suggesting that they have a similar surface composition component. Paper I examines single observations of Ananke and Callirrhoe, suggesting the presence of the 0.7 μm feature (from Grav et al. 2003).

Zoom In Zoom Out Reset image size

One satellite from the 165° inclination retrograde group, JXI Carme, shows a reddened reflectance spectrum (Figure 3). The broadband photometry (Paper I, using data from Tholen & Zellner 1983; Luu 1991; and Grav et al. 2003 ) is divided between the presence and absence of the absorption feature at 0.7 μm of a reddened object. Near 0.41 μm, the reflectance spectrum begins to flatten with decreasing wavelength (relative bluing).

The satellite JIX Sinope was observed twice (Figure 4). The 2006 spectrum shows an absorption feature at 0.40–0.44 μm but no obvious 0.7 μm absorption feature. The 2009 spectrum suggests an absorption feature beginning near 0.48 μm with a minimum near 0.64–0.66 μm (see Figure 4). This is consistent with the broadband photometry (Paper I, using data from Luu 1991 and Grav et al. 2003), which is divided between the presence and absence of this feature. No rotational period is known for Sinope. Debate exists as to whether Sinope is, in fact, a member of the retrograde cluster anchored by Pasiphae (Sheppard & Jewitt 2003). Sinope is roughly the same mean distance from Jupiter as the cluster (∼23.8 × 10⁶ km). Notably, if Sinope is a member of Pasiphae's cluster, it becomes the second-largest body in the cluster (see Table 3), estimated to be 1.2 times larger in diameter than JXII Ananke. Our compositional information addresses this question in Section 3.9.

As the most direct connection of solar system bodies that show these spectral features, we consider the spectra of these objects (our targets) in terms of the CM2 carbonaceous chondrites. To first order, these objects (CM2 chondrites) are made up of 57–85 vol.% matrix material, with the rest being chondrules, inclusions, and mineral fragments within the matrix (see Grossman & Olson 1974; McSween 1979; Zolensky & McSween 1988; Brearley & Jones 1998; Bland et al. 2004; Howard et al. 2009; Cloutis et al. 2011). The X-ray diffraction modal analysis of four CM2 meteorites shows that the overall composition of these chondrites includes Mg-serpentine (49–59 vol.%) and Fe-cronstedtite (43–50 vol.%) (see Howard et al. 2009, 2010), about equally split in amount and constituting a large percentage of the meteorites, with the exception of Cold Bokkeveld, where there is roughly twice as much Mg-serpentine as Fe-cronstedite. Other phases also show a lower abundance range across these CM2 chondrites: olivine (10.4–17.3 vol.%), pyroxene (3.3–8.4 vol.%), calcite (1.01–3 vol.%), gypsum (0–1.6 vol.%), magnetite (1.1–2.4 vol.%), pentlandite (0–2.1 vol.%), and pyrrhotite (0.7–3.1 vol.%) (see Howard et al. 2009; Cloutis et al. 2011). The mafic silicates are likely the only two materials of these other phases having a sufficient amount to be potentially spectrally detectable.

How does the division of CM2 meteorites by matrix and other components affect our study? Existing compositional/spectral analyses have been done on matrix material and the CM2 meteorites as a whole; reflectance spectra of individual chondrules do not yet exist.

The CM2 matrix composition is dominated by Fe-rich serpentine-group phyllosilicates and intermediate, more Mg-rich chrysotile-like members (Barber 1981; Zolensky et al. 1993). Serpentine is the dominant phyllosilicate in the CM2 chondrite matrix, present in a mixture of Mg–Fe serpentine, Fe²⁺ → Fe³⁺ cronstedtite, and a tochilinite–serpentine–cronstedtite intergrowth, described by Bunch & Chang (1980) as a fine-grained (<1 μm) mixture of phyllosilicates, carbonaceous matter, and sulfides. Based on the spectral similarities with C-complex asteroids and the presence of Fe-bearing saponite and serpentine-group phyllosilicates in CM2 meteorite laboratory analyses, we consider Fe-bearing phyllosilicates in our mixture of materials.

Based on modeling of the distribution and state of carbon in the solar system (Hendrix et al. 2016), the presence of carbonaceous matter in the CM2 meteorite laboratory analyses, the lower geometric albedos of the Jovian satellites and carbonaceous chondrites, and the ubiquitous presence of the presumably carbon-rich D-class asteroids at Jupiter's heliocentric distance, we include carbon products we expect from carbonization at different heliocentric distances.

Finally, assuming that the objects in an individual cluster are genetically related, we use the addition, subtraction, or change of materials to direct or adapt the fit to the highest-S/N cluster member (usually the largest satellite) to match the remaining cluster members. These changes are then considered in our interpretation of the history of the satellites.

While the process of aqueous alteration and metamorphism of carbonaceous chondrites shows many different components, the most volumetrically dominant phases in carbonaceous chondrites are not necessarily the phases that are spectrally dominant (e.g., Cloutis et al. 1990, 2011). We address the end-member compositions we expect to be spectrally apparent across the wavelength range we investigate here. We considered these end-member compositions in the modeling: Fe-bearing serpentine and saponite combined (0.7 and 0.43 μm absorption features) labelled here as "Himalia Phyllosilicate" (see discussion below), a lizardite labelled here as "Serpentine", ⁴ amorphous carbon (aC) labelled here as "Carbon" (Rouleau & Martin 1991), and a kerogen-like spectrum, which is darker and redder than the lab-measured kerogen (Arakawa et al. 1989) and is based in strength on the Sinope spectrum. The spectra of the starting-point end-member compositions are shown in Figure 8.

Zoom In Zoom Out Reset image size

The 28° inclination cluster is anchored by JVI Himalia and presents the first conundrum in the analysis. The feature centered on 0.7 μm in Himalia's spectrum is spectrally slightly wider compared to C-complex asteroids that show this feature (e.g., Vilas et al. 1993b). Figure 6 confirms the presence of this difference in Himalia's spectrum. The errors invoked by the separation of roughly 10–11 yr between the dates of observation of these older Himalia reflectance spectra and the newer spectrum presented here prevent us from stating unequivocally that these sample the same or opposite sides of the satellite.

We note, however, that—in its simplest form—the accepted sequence of aqueous alteration includes the conversion of Fe-bearing saponites to Mg-rich serpentines (see Cloutis et al. 2011). The saponites form before the serpentines. The reflectance spectra of terrestrial Fe-bearing saponites generally show the Fe²⁺ → Fe³⁺ charge transfer transition absorption feature centered at slightly shorter wavelengths than Fe-bearing serpentines (see Cloutis et al. 2011). Progress in this sequence could include both saponites and serpentines that could demonstrate the spectral shift in position of the Fe²⁺ → Fe³⁺ charge transfer transition. The repeatability of this configuration in all moderate-resolution reflectance spectra that we have of Himalia, as well as its appearance in the spectra of Elara (see below), suggest that the breadth of this absorption is due to the level of progression of aqueous alteration at the time it formed.

With no prejudice toward future work, we choose to define the spectral shape of Himalia as one of our end-point compositions for this analysis ("Himalia Phyllosilicate," which could be a blend of serpentine and saponite). We base this on evidence of the repeatability of this spectral absorption (Paper I, identified in six of seven observations considered; our three observations of this absorption feature in Himalia's spectrum) plus the presence of absorption near 3.0 μm suggesting bound water under aqueously altered conditions (Chamberlain & Brown 2004; Brown & Rhoden 2014) and the presence of a broad absorption centered near 1.2 μm suggesting ferric phyllosilicates and a darkening agent such as magnetite (Bhatt et al. 2017).

We also define the spectral shape of the kerogen-like end-member ("kerogen") based on the spectrum we obtained of JIX Sinope in 2006, representing an mixture of organics defined by the low albedo coupled with redder material presumed to be less processed (simpler hydrocarbons). Laboratory examples of like materials exist (e.g., Sill 1973); some of these are digitally unavailable to the authors.

We derived intimate mixture models to fit the observed spectra of each body using the Hapke-based formulations as described in Hendrix et al. (2010) and Hendrix & Hansen (2008). We opted to constrain the models to two-component intimate mixtures. In these models, we used lab-measured optical constants to derive the single-scatter albedos for the end-member aC and serpentine. For the Himalia Phyllosilicate and Sinope-based kerogen-like end-members, we used the method of Lucey (1998) to derive estimated optical constants based on a measured reflectance.

Table 4 shows the results of the modeling of the individual Jovian satellites, discussed in the sections below. The proportions could vary in future modeling efforts; they are likely not final percentages. They are, however, the relative proportions and significance of the materials (kerogen, graphitized carbon, serpentine) that we model here and that we use to address the histories of these objects.

The spectrum of JVI Himalia has two obvious absorption features: the slightly wider 0.7 μm absorption feature and the 0.43 μm absorption feature (Figure 9). The CM2 carbonaceous chondrites largely show the presence of the 0.7 μm feature in extraterrestrial materials (e.g., Vilas et al. 1994). This 0.7 μm feature is attributed to an Fe²⁺ → Fe³⁺ charge transfer transition in oxidized iron in phyllosilicates. Serpentines are the dominant phyllosilicate found in the CM2 chondrites (see above). The 0.43 μm absorption feature is consistent with, but not uniquely associated with, the presence of phyllosilicates containing tetrahedrally coordinated Fe³⁺ (Cloutis et al. 2011). It is also similar to the ⁶A₁ → ⁴A₁, ⁴E(G) Fe³⁺ spin-forbidden feature seen in the spectra of iron sulfate jarosite (e.g., Vilas et al. 1993a). In general, it can be attributed to a spin-forbidden oxyhydroxy-bridged pair of Fe³⁺ cations. It has been observed in multiple spectra of low-albedo C-complex asteroids (e.g., Vilas et al. 1993a), many showing the 0.7 μm absorption feature. Thus, defining Himalia's surface composition as containing some Fe-bearing serpentine is reasonable and consistent with observations and laboratory data.

Zoom In Zoom Out Reset image size

The 0.7 μm absorption feature is observed in spectra of about 60% of the main-belt C-complex asteroids that also contain the 3.0 μm water of hydration absorption feature (Howell et al. 2011; Rivkin et al. 2015). Brown & Rhoden (2014) conclude that there is no 0.7 μm absorption feature in Himalia's spectrum based on the association of the shape near 3.0 μm that they observe for Himalia with an adjusted spectrum of the C-complex asteroid 52 Europa; apparently, this rough shape is generally associated with C-complex asteroids such as 52 Europa that do not show a 0.7 μm feature (Takir & Emery 2012). Evidence for the presence of the 0.7 μm feature here and in Paper I suggests that the similarity of this particular shape of the 3.0 μm feature is not necessarily an indicator of the presence or absence of the 0.7 μm absorption feature in the spectrum of a small body.

Our modeling also finds significant quantities of aC that dominate the bulk material and likely darken the surface. The gray spectrum suggests that the carbon material is more graphitized, with increased removal of H.

The two spectra of JVII Elara differ slightly (Figure 9). The composition of the 2006 spectrum has the same phyllosilicate and aC components as Himalia, although it varies in the percentage amount of each component. The 2009 Elara spectrum differs in the kerogen (organics) present in lieu of the aC component (and apparent in a slight increase in reddening), and the amount of Himalia Phyllosilicate decreases. Elara has a slightly lower geometric albedo than Himalia (Table 3). The S/N does not allow us to discern a 0.43 μm absorption feature.

Based on their linear unmixing modeling of spectra obtained across 0.8–2.4 μm, Bhatt et al. (2017) determined that the compositions of Himalia and Elara contain the same components, resulting in a combination of iron oxides and ferric phyllosilicates. Our modeling suggests a common composition for one part of Elara with Himalia (Table 4), with Fe-bearing phyllosilicates and aC in varying quantities, consistent with the findings of Bhatt et al. (2017).

Overall, this suggests that Elara also contains material that has been subject to aqueous alteration, and carbonaceous material also exists, although the state of the aC (reflected in the differences in surface spectral reflectance) changes within the satellite. Referencing Hendrix et al. (2016), the shift from a reddish spectrum to a flatter spectrum can be linked to carbonization related to increased processing/weathering of the material. This result is consistent with Paper I, where the photometry of Elara is gray, and the presence or absence of the 0.7 μm feature varies. Since there is no accurately known rotational period for Elara, it is not certain whether the derived albedo correlates with specific material on Elara's surface, or whether there is a noticeable change in albedo with a change in surface location.

Lysithea has a spectrum that also indicates the 0.7 μm feature (Figure 9) and also has the lowest geometric albedo of these three objects (Table 3). Lysithea's spectrum shows increased reddening, a clue to the compositional change in Lysithea. The modeling shows an increase in kerogen content for the Lysithea spectrum over the 2009 Elara spectrum (Table 4). The presence of the 0.7 μm feature is consistent with that noted in the one observation in Paper I (from Grav et al. 2003). The S/N does not allow us to discern a 0.43 μm absorption feature. The increased spectral reddening and lower albedo than both Himalia and Elara (Table 3), interpreted as an increase in kerogen abundance, suggest that Lysithea is the satellite in this cluster with the lowest level of aqueous alteration and the greatest amount of relatively pristine organic material (where processing of organic material would ultimately lead to production of aC) and thus the most primitive material of the three satellites. The smaller amount of Himalia serpentine present is consistent with the aqueous alteration process having advanced the least in Lysithea.

We propose a scenario where the three observed satellites represent fragments probing the interior of the cluster's original parent body. We use here the diameters calculated by the NEOWISE project (Table 3; Grav et al. 2015; Mainzer et al. 2016) summed to estimate the diameter. Himalia is the largest satellite in this cluster. Its lack of known spectral variations supports a relatively uniform surface composition; Paper I examines six different data sets that suggest the 0.7 μm feature and one data set that does not (Degewij et al. 1980; Tholen & Zellner 1983; Luu 1991; Jarvis et al. 2000). Spectral indications of aqueous alteration products in its surface material support the idea that it was an internal part of a larger body that provided the environment for aqueous alteration to proceed, uniformly affected at the radius that was exposed. We suggest that Himalia is the core of the parent body.

We propose that Elara, as a smaller fragment ∼0.57× the diameter of Himalia, preserves the geochemical/mineralogical transition from the core to the outer layer(s) of the parent body. Lysithea samples a piece of a layer of the body closer to or at the surface of the satellite. Figure 10 shows a notional sketch of a parent body illustrating where Himalia is the proposed body core, Elara represents an adjacent portion of the body, and Lysithea probes farther from the center of the parent body. Our rough calculated diameter would be 303.8 km; although limited in accuracy, the general assumption would be that the parent body's diameter was at least on the scale of 300 km. We conclude that before the parent body fragmented, the interior was likely subjected to aqueous alteration.

Zoom In Zoom Out Reset image size

Drawing on laboratory studies of CM2 carbonaceous chondrites, the temperature of the aqueous fluid that altered the parent bodies of the CM2 meteorites has been estimated at ≤50 °C (323 K; e.g., Zolensky et al. 1993). Heating experiments on Murchison samples indicate that the 0.7 μm absorption feature is present in the laboratory reflectance spectra of samples at room temperature but disappear when the sample is heated to 400°C (673 K; Hiroi et al. 1993). Thus, the interior temperatures of the aqueously altered objects reached, but did not exceed, the range of 50 °C–400 °C. This does not preclude the fact that mafic igneous rocks requiring higher initial formation temperatures likely constitute some of the parent body's material that was subsequently aqueously altered.

Compared to the C-complex asteroid telescopic spectra and CM2 meteorite laboratory spectra, the 0.7 μm absorption features observed in the 28° inclination prograde cluster satellites are notably shallower. We observe this despite the NEOWISE mean albedo for Ch-class asteroids at 0.056 ± 0.003 (Mainzer et al. 2011), effectively the same as the NEOWISE value of 0.057 ± 0.008 for Himalia. The Ch asteroids were defined separately from the rest of the C-complex asteroids by the presence of the 0.7 μm absorption feature (Bus & Binzel 2002). Assuming the lower albedo is due to the presence of some combination of opaques such as carbon, organics, iron sulfides, and magnetite all potentially present or easily formed from "raw" material in the parent body, there are two potential causes for the shallow feature. First, a large amount of this darkening material is present in the parent body of these satellites, suppressing the depth of the absorption feature. Our modeling results support a large volume of carbon material (Table 4). Second, the aqueous alteration process did not progress beyond an intermediate state, where olivine altered to form magnesian serpentine (no 0.7 μm feature) and metal and sulfides reacted to form Fe-rich tochilinite (the beginning of the Fe-rich 0.7 μm feature). Given the location of the parent body when it broke apart (through whatever means), aqueous alteration likely stopped. Any form of space weathering that would occur at that location would have affected the surfaces exposed as part of the breakup; this space weathering would include effects invoked by both the postcapture fragmentation process (e.g., interactions with gases in a voluminous Jupiter atmospheric envelope) and the exposure to space following the fragmentation.

The overall change in albedo decreasing from Himalia → Elara → Lysithea follows the trend we see spectrally in the main-belt C-complex asteroids, where a decrease in geometric albedo is observed concurrently with a decrease in the presence and depth of the 0.7 μm absorption feature (e.g., Sawyer 1991; Vilas 1994; Fornasier et al. 2014). This is consistent with less aqueous alteration of the asteroids.

The 28° inclination prograde cluster serves as an example of a parent body that underwent limited aqueous alteration and was isolated at Jupiter's heliocentric distance following capture and fragmentation. This could be partly due to the increase in heliocentric distance affecting a change in micrometeoroid flux and solar insolation, as well as effects from the proximity to Jupiter, as all of the other known outer irregular Jovian satellites are at least twice as distant as the Himalia cluster (Himalia semimajor axis, a = 11.46 × 10⁶ km; e.g., Sheppard & Jewitt 2003).

The spectrum of JVIII Pasiphae shows the 0.7 μm feature similar to that observed in the Himalia Phyllosilicate spectrum (Figure 2) and the 0.43 μm feature described above, indicating that Pasiphae underwent similar aqueous alteration. The spectrum of Ch-class asteroid 19 Fortuna (Figure 5) is shown for comparison with Pasiphae. Paper I noted both the presence and absence of the 0.7 μm feature, suggesting rotational variation on the surface of Pasiphae; we cannot confirm or negate that possibility, as no rotational period is known for any of the objects in this cluster. The spectrum is well modeled as having 36% of the Himalia Phyllosilicate sample mixed with 64% of the flat aC (Figure 11).

Zoom In Zoom Out Reset image size

The Pasiphae spectrum has the best S/N among this cluster of objects, expected because its diameter is the largest, thus providing the strongest signal. Pasiphae's spectrum is gray, consistent with the broadband photometry (Paper I and Grav et al. 2003). The gray spectrum suggests that the carbon material is more graphitized, with increased removal of H. The visible geometric albedo of 0.044 ± 0.007 for Pasiphae is lower than the NEOWISE mean albedo for Ch-class asteroids at 0.056 ± 0.003 (Mainzer et al. 2011). This suggests two things. First, the darkening carbon and/or other darkening material is present in sufficient quantities to reduce the albedo. Alternatively, the aqueous alteration had not proceeded in extent before it was disrupted (see the above description of the aqueous alteration process above under 3.6). These two causes could both contribute to the lower albedo.

The two spectra of JXII Ananke differ slightly in spectral properties (Figure 11). The compositions of both the 2006 and 2010 spectra have the low-Fe phyllosilicate and aC components, and the percentage of Fe-bearing phyllosilicates in Ananke at 8% is significantly less than the amount modeled in Pasiphae at 36% (Table 4). There is a slight increase in reddening. Ananke has a geometric albedo of 0.038 ± 0.006, slightly lower than Pasiphae (Table 3). The S/N does not allow us to discern a 0.43 μm absorption feature. The variation in Ananke's background between gray or slightly red, coupled with the presence of the 0.7 μm feature, is also suggested in the broadband photometry (Paper I, using data from Grav et al. 2003).

The overall change in albedo, decreasing from Pasiphae → Ananke, again follows the trend in the main-belt C-complex asteroids, where a decrease in geometric albedo is observed concurrently with a decrease in the presence and depth of the 0.7 μm absorption feature (e.g., Sawyer 1991; Vilas 1994; Fornasier et al. 2014). In this case, we also note that the decrease in albedo and changed spectral attributes for Pasiphae → Ananke are similar to the changes seen for Elara → Lysithea.

Pasiphae is ∼57.8 km in diameter (Table 3). If Pasiphae is the core of a parent asteroid, the parent asteroid underwent some level of aqueous alteration before fragmentation. Again, the change in composition suggests that Ananke, as a smaller fragment roughly half the diameter of Pasiphae, preserves the geochemical/mineralogical transition from the core to the outer layer(s) of a parent body. Based on the albedo, the level of alteration inferred by the 0.7 μm feature, and the overall sizes of the two bodies, aqueous alteration did not progress beyond early stages in the larger parent body that contained both Pasiphae and Ananke. We propose that we are seeing different sides of Ananke preserved in the different spectra; this can be tested by an accurate rotational period determination. Using the diameters tabulated by the NEOWISE project (Table 3; Grav et al. 2015; Mainzer et al. 2016), we place a lower limit of 90 km on the size of the core of the (larger) Pasiphae parent body.

The spectrum of JXI Carme appears red (increasing reflectance with increasing wavelength) across the VNIR wavelength region (Figure 12). Incorporating the low albedo of 0.038, our modeling suggests that the surface composition consists of 4% low-Fe serpentine with 96% kerogen. The bluing or upturn near 0.4–0.5 μm in Carme's spectrum also suggests a graphite component to the surface composition of Carme. Laboratory reflectance spectra of graphite show a slight upturn at that location (Cloutis et al. 2011). The ONC-T data from Hayabusa2 and ground-based telescopic spectra of 162173 Ryugu (F. Vilas 2023, personal communication) show a similar bluing suggesting graphite.

Zoom In Zoom Out Reset image size

In Figure 13, we compare the relative reflectance spectra of Carme, JIX Sinope, and the mean, upper, and lower ranges of relative reflectance from ECAS photometry for the D-class asteroids that dominate the Trojan asteroids at Jupiter's heliocentric distance (Tholen 1984). Carme is a very close match to the mean D-class asteroid values. The ECAS broadband photometry also shows the bluing effect at the lower wavelengths of the least reddened D-class spectrum. This suggests that the graphite component could extend to a subset of the D-class asteroid surface composition. The two spectra for Sinope show a redder slope than Carme but fall within the spectral range of the ECAS values. A recent VNIR spectrum of Jovian irregular satellite JXVIII Themisto (Sharkey et al. 2023) shows similar reddening to the spectra of Carme and Sinope, suggesting a similar composition and possibly origin.

Zoom In Zoom Out Reset image size

Both spectra of JIX Sinope are in agreement with each other and are similar in the spectral region to the D-class asteroids (Figures 4 and 12). Both observations of Sinope suggest a very small amount of low Fe-bearing serpentine but a large kerogen component.

The significant compositional differences between Sinope and the two members of the 149° inclination cluster, Pasiphae and Ananke, would seem to argue against the common parent body origin suggested by Sheppard & Jewitt (2003). The mixed red and gray broadband photometry of this cluster (Grav et al. 2003), with mixed results from testing for the presence of the 0.7 μm feature in this photometry (Paper I), support, however, the potential coexistence of two types of material in the parent body of the cluster. Sharkey et al. (2023) show moderate-resolution NIR spectra of Pasiphae and Sinope, where the spectrum of Sinope is redder at lower wavelengths but alters at 1.6 μm to a less reddened slope in agreement with the spectrum of Pasiphae. This could agree with our spectral findings.

Statistical analyses show no great difference between the average shape of the irregular satellites of outer planets and the average shape of asteroids (as defined by variations in broadband magnitudes for an individual object; Graykowski & Jewitt 2018), lending support to theories that the origins of the irregular satellites are the collisional breakup of their parent bodies.