JWST near-infrared spectroscopy of the Lucy Jupiter Trojan flyby targets: Evidence for OH absorption, aliphatic organics, and CO₂

The most prominent feature in the Trojan spectra is the broad absorption band that is centered around 3 $\mu$m. This band is deepest on Eurybates and clearly discernible on the other less-red objects Patroclus and Polymele. Meanwhile, on the red objects Orus and Leucus, the presence of this band is considerably more marginal. A characteristic property of the absorption is its distinct short-wavelength edge at roughly 2.7 $\mu$m. Previously, Brown (2016) detected a portion of this 3 $\mu$m band in Keck/NIRSPEC spectra of 16 Trojans, with less-red objects displaying deeper band depths than red objects. The significantly higher signal-to-noise ratio JWST/NIRSpec spectra confirm the earlier detections and robustly corroborate the trend between surface color and 3 $\mu$m band depth.

We determined the center of this broad absorption by dividing each spectrum by a linear continuum fit across the feature, fitting a third-order polynomial to the bottom of the band, and taking the location of the minimum of the best-fit polynomial as the band center. For Orus and Leucus, we excluded the region around the 3.4 $\mu$m organic band from the polynomial fit (see Section 3.2). The band depth was calculated by averaging the continuum-divided spectrum within a 0.05 $\mu$m wide window centered around the band center. This window width was selected so that $\sim$25–30 channels are averaged. This process was repeated 10,000 times, each time resampling the data points using normal distributions with widths equal to the respective uncertainties. The reported values and uncertainties are the mean and standard deviation of the resulting parameter distributions. We also computed the band depth at 2.9 $\mu$m to enable more direct comparisons with ground-based spectra of near-Earth and main belt asteroids, whose band depths are often reported at this wavelength in the literature. Table 3 shows the results of our band analysis. The continuum-subtracted Trojan spectra in this region are presented in Figure 3.

The continuous wavelength coverage of the JWST spectra enables us to perform a more detailed characterization of the 3 $\mu$m band than was previously possible. High-quality spectra of small bodies covering the 3 $\mu$m region are scant in the literature. Ground-based observations in this wavelength range are challenging due to significant telluric absorption, which almost entirely suppresses the incoming light at 2.5–2.9 $\mu$m. While a significant body of IRTF/SpeX spectra is available for main belt asteroids (Takir & Emery, 2012), including many that show superficially similar 3 $\mu$m absorption features to the Trojans, the gap in wavelength coverage precludes definitive compositional inferences. The only available continuous reflectance spectra in this region were obtained from space telescopes or spacecraft flyby missions. In Figure 3, we plot a collection of representative spectra that show various types of absorption features that have been observed on main belt asteroids.

The spectrum of Pallas, derived from measurements made by the AKARI spacecraft (Usui et al., 2019), shows a distinctive sharp 2.7 $\mu$m absorption that is diagnostic of hydrated silicates — specifically, the O–H stretch feature in phyllosilicates (e.g., Lebofsky, 1980; Jones et al., 1990; Rivkin et al., 2002). The spectra of Ceres (from globally averaged observations by the Dawn spacecraft; De Sanctis et al. 2018) and Themis (compiled from AKARI and ground-based IRTF spectra; Rivkin & Emery 2010; Usui et al. 2019) display the same 2.7 $\mu$m hydrated silicate band. For Trojans generally, we can exclude hydrated silicates as a major component in the surface regolith due to the absence of a sharp absorption band centered at $\sim$2.7–2.8 $\mu$m as seen in Ceres, Palas, and Themis.

Themis and Ceres show a second spectral feature around 3.0–3.2 $\mu$m; at even longer wavelengths, there are other features from organics and/or carbonates, which will be discussed in the following subsection. Several interpretations have been forwarded for the 3.0–3.2 $\mu$m band. In the case of Themis models incorporating very fine water frost ($<0.1$ $\mu$m) coating underlying refractory grains provide a good match (Campins et al. 2010; Rivkin & Emery 2010; but see McKay et al. 2017; O’Rourke et al. 2020 for arguments against this interpretation). Meanwhile, N–H stretch absorption from ammoniated surface species has been hypothesized to be a primary contributor to the feature on Ceres (e.g., King et al., 1992; De Sanctis et al., 2015). The lack of absorption around 3.0–3.2 $\mu$m in the Trojan spectra leads us to argue against the presence of fine-grained water frost or ammoniated species as a distinct surface element on Trojans.

We have also included the spectra of two active objects: the Rosetta spectrum of the Jupiter-family comet 67P/Churyumov–Gerasimenko (67P; Poch et al., 2020) and the recently published JWST/NIRSpec spectrum of the active outer main belt asteroid 238P/Read (238P; Kelley et al., 2023). 67P lacks the hydrated silicate feature at 2.7 $\mu$m while having a complex of absorption bands spanning 2.8–3.6 $\mu$m that is attributed to a combination of ammoniated salts and organics. In contrast, the spectrum of 238P (shown separately in the bottom panel of Figure 3) has a single broad 2.7–3.6 $\mu$m band that is quite comparable to the feature seen in the less-red Trojan spectra, albeit with the reflectance minimum situated at somewhat longer wavelengths. Although 238P is much smaller ($\sim$0.6 km) than the Trojans there were observed by JWST, the similarity in spectral shapes is noteworthy. 238P is more dynamically akin to typical outer main belt asteroids than to the Jupiter-family comets, indicating that it was not recently implanted into the inner solar system from the Kuiper belt (Haghighipour, 2009; Hsieh et al., 2016). Nevertheless, it is still possible that the origin of 238P lies in the outer Solar System, as dynamical instability models predict a nonnegligible contingent of primordial trans-Neptunian planetesimals among the outer main belt asteroid population (e.g., Vokrouhlický et al., 2016). Future spectroscopic surveys of more small outer main belt asteroids may uncover other objects with Trojan-like 2.7–3.6 $\mu$m absorption bands.

Given the predictions from dynamical instability models that place the origin of the Trojans in the outer solar system, we now turn to KBOs for further insight into the 3 $\mu$m feature. Prior to JWST, only a handful of KBOs had been spectroscopically studied at wavelengths longer than 3 $\mu$m. During the first year of science operations, JWST observed more than 70 KBOs, providing a treasure trove of 1–5 $\mu$m reflectance spectra spanning all dynamical classes and a wide range of sizes. For comparison with our Trojan spectra, we analyzed publicly available NIRSpec observations of a pair of KBOs that were obtained as part of General Observers Program #2418 (PI: N. Pinilla-Alonso; Pinilla-Alonso et al. 2024). Photometric surveys of KBOs have shown that objects smaller than 1000 km in diameter can be broadly divided into two groups based on their visible colors (e.g., Fraser & Brown, 2012; Wong & Brown, 2017; Fraser et al., 2023). The two KBO targets we selected—2004 NT33 (420 km; Vilenius et al. 2014) and 2004 TY364 (510 km; Lellouch et al. 2013)—are respectively from the so-called red and very-red subgroups within the hot classical population.

In addition, we included two Centaurs (i.e., inward-scattered KBOs that cross the giant planet region on unstable orbits) that were observed as part of Program #2418 and published in Licandro et al. (2024): Okyrhoe and 2003 WL7 (35 and 103 km, respectively; Duffard et al. 2014). These Centaurs have perihelion and aphelion distances of $q=5.8$ au and $Q=8.4$ au, and $q=14.9$ au and $Q=20.1$ au, respectively. All four observations were obtained with the low spectral resolution prism grating (0.6–5.3 $\mu$m; $\Delta\lambda/\lambda$ $\sim$ 30–300), and we processed the observations following the same methodology that was outlined in Section 2. The prism-mode spectrum of the G-type solar analog star SNAP-2 (Program #1128; PI: N. Luetzgendorf; publicly available data) was used to produce the reflectance spectra, which we plot in Figure 4.

The spectra of 2004 NT33 and 2004 TY364 differ broadly in several ways. 2004 NT33 shows strong water ice absorption bands at 1.5, 2, and 3 $\mu$m, including a distinct feature at 1.65 $\mu$m and the Fresnel reflection peak at 3.1 $\mu$m, both of which are indicative of the crystalline form of water ice (see Mastrapa et al. 2013 for a review of amorphous and crystalline water ice on solar system objects). Meanwhile, the spectrum of 2004 TY364 is considerably less water ice rich, with only marginal detections of the 1.5 and 2 $\mu$m absorption features and no signs of the Fresnel peak. It shows a deep and broad 2.7–3.3 $\mu$m absorption feature and an additional rounded spectral band at 3.3–3.6 $\mu$m that corresponds to aliphatic organics (see Section 3.2). The pair of narrow features between 2.6 and 2.8 $\mu$m are overtones of the $\nu_{3}$ band of CO₂ ice (e.g., Quirico & Schmitt, 1997; Baratta & Palumbo, 1998), which have been detected in the JWST spectra of many KBOs (de Prá et al., 2024; Pinilla-Alonso et al., 2024). Moving inward to the Centaurs, we find that the spectrum of 2003 WL7 is largely consistent with 2004 NT33, showing clear water ice absorption bands, albeit with somewhat reduced depths. In contrast, the water ice features on Okyrhoe are strongly attenuated, and the shallow 3 $\mu$m band has a more asymmetric profile that gently rises toward the continuum at 3.6 $\mu$m.

The four KBO and Centaur spectra shown in Figure 4 are broadly representative of the range of 3 $\mu$m band shapes seen on similarly sized bodies throughout the trans-Neptunian region (Licandro et al., 2024; Pinilla-Alonso et al., 2024). Placing these spectra alongside our results for the Trojans, we note the strong similarity in the general shape and width of the broad 3 $\mu$m absorption features to the observed feature on Eurybates, Patroclus, and Polymele, with consistent short-wavelength edges located at roughly 2.7 $\mu$m. This comparison suggests that the same molecular vibration — namely, the fundamental O–H stretch mode (e.g., Mastrapa et al., 2009) — is responsible for the 3 $\mu$m absorption across all of these objects. In the case of KBOs and Centaurs, the primary contributor to the 3 $\mu$m band is water ice (both amorphous and crystalline), which is known to be widely distributed across the populations based on detections of the 1.5 and 2 $\mu$m bands (e.g., Barkume et al., 2008; Barucci et al., 2011).

The identity of the OH-bearing component on the Trojans’ surfaces is less clear. The absence of the 2 $\mu$m water ice band precludes a significant water ice fraction across the surface, consistent with previous observations (Emery & Brown, 2003; Yang & Jewitt, 2007). This finding also aligns with the results from thermophysical modeling of Trojans, which demonstrate that water ice should be depleted down to $\sim$10 m depth on average, though the sublimation front at polar regions may be considerably shallower, up to a depth $\sim$ 10 cm (Guilbert-Lepoutre, 2014). While a coating of fine-grained water frost similar to the proposed component on Themis can effectively suppress the 1.5 and 2 $\mu$m absorption features (Rivkin & Emery, 2010), it produces an OH band at 3.1 $\mu$m that is too narrow to match the observed feature on Trojans. Instead, larger micron-sized grains of water ice are necessary to reproduce the broad rounded shape and the short-wavelength edge at 2.7 $\mu$m that are seen on Trojans, KBOs, and Centaurs, as well as the comets Tempel 1 (Sunshine et al., 2006), Hartley 2 (Protopapa et al., 2014), and 238P (Kelley et al., 2023).

Exposure of more pristine subsurface material through impact gardening has been proposed as a possible avenue by which otherwise unexpected water ice features may emerge on small dark bodies in the middle solar system. Spatially resolved observations of the irregular Saturnian satellite Phoebe, which has a similar size to the largest Trojans and is thought to have originated from the outer solar system, have shown that while water ice absorption is manifested across the entire surface, they are most prominent within impact basins and are correlated with patches of higher optical geometric albedo (Clark et al., 2005; Buratti et al., 2008; Fraser & Brown, 2018).

Extending this argument to the Trojans, which have experienced a similar level of collisional evolution as the main belt asteroids (Marchi et al., 2023), surface turnover from cratering impacts may periodically expose small localized patches of water-rich interior material amid the predominant water-ice-poor dark regolith. However, it is unclear if the present day impactor flux is sufficient to maintain pristine water ice on the surface, given the significantly higher rate of sublimation loss at the higher temperatures of the Trojan region (Guilbert-Lepoutre, 2014; Lisse et al., 2022). We note that isolated patches of water ice have been detected in spatially resolved spectra of the comets Tempel 1 (Sunshine et al., 2006) and 67P (Barucci et al., 2016), which experience much higher average surface temperatures than the Trojans. In those cases, the water ice is concentrated in regions that are topographically shielded from insolation or freshly exposed due to physical processes such as mass wasting. It follows that analogous surface configurations on the Trojans may preserve water ice for extended periods. In the context of this proposed scenario, the difference between the 3 $\mu$m spectra of less-red and red Trojans would indicate that the former have intrinsically more water ice in their subsurface than the latter and suggests distinct initial surface compositions. The future Lucy flybys will provide spatially resolved optical and near-infrared photometry and spectroscopy of the Trojan targets’ surfaces and allow us to assess whether or not localized regions of exposed water ice are present on these objects.

An alternative explanation for the observed 3 $\mu$m OH band on the less-red Trojans is organic refractory material formed by the irradiation of carbon-bearing ices. Laboratory experiments simulating the effect of solar processing on H₂O $+$ CH₃OH $+$ NH₃ ice mixtures (Urso et al., 2020) produced complex residues that display a broad, asymmetric absorption band extending from 2.7–2.8 $\mu$m to 3.6–3.8 $\mu$m, in good agreement with the absorption feature seen on Trojans. This interpretation has also been presented to explain the absorption feature observed on 238P (Kelley et al., 2023). In this scenario, the OH band around 3 $\mu$m is produced through the dissociation of H₂O and CH₃OH. Additional bands that contribute to extending the overall absorption feature to longer wavelengths are due to C–H and N–H stretch modes.

The KBO–Centaur–Trojan comparison has important implications for our understanding of solar system dynamical evolution. Numerical integrations that simulate the effect of current dynamical instability models on the primordial trans-Neptunian disk invariably show that the inward-scattered planetesimals, some of which ultimately become captured into resonance by Jupiter, pass through an intermediate stage when they lie on highly eccentric giant-planet-crossing orbits (e.g., Morbidelli et al., 2005; Nesvorný et al., 2013; Roig & Nesvorný, 2015). Centaurs therefore serve as an ideal proxy for assessing how Trojans may have evolved during their inward migration.

The spectrum of Okyrhoe is unique among the body of published JWST spectra of KBOs, with its extremely muted ($<20\%$) 3 $\mu$m OH band and the characteristic asymmetric shape of its broad 2.7–3.6 $\mu$m absorption. Its spectrum also shows the strongest resemblance to the spectra of less-red Trojans out of all the small body spectra we have presented. Okyrhoe has one of the closest current orbits among the Centaurs hitherto observed by JWST, and we hypothesize that the difference between Okyrhoe and the more distant Centaur 2003 WL7, which has a spectrum that is much more consistent with that of typical KBOs, is indicative of differential evolution of Centaurs as a function of their received insolation. Within the framework of dynamical instability models of solar system evolution, we can apply this argument to the entire Trojan population, which has been located at 5 au for $\sim$4 Gyr. It follows that the cumulative thermal processing of Trojan surfaces may have further attenuated the 3 $\mu$m OH band beyond the level apparent on Okyrhoe, producing the final spectra we see today.

It is important to note that the KBOs observed by JWST so far are significantly larger than the Trojans and Centaurs. As such, we are unable to disentangle possible size–dependent trends in the initial bulk composition and subsequent surface evolution of these objects directly. Future JWST observations of additional low-perihelion Centaurs and KBOs smaller than 300 km in diameter will enable a more robust assessment of the proposed evolutionary trajectory that links the KBOs and Centaurs to the Trojans and provide a definitive empirical test of the purported outer solar system origin of Trojans.

Moving to slightly longer wavelengths, we find a distinct absorption band centered around 3.4 $\mu$m. In contrast to the aforementioned 3 $\mu$m band, this feature is more prominent on the red objects Orus and Leucus than it is on the less-red targets. To examine the shape of these features better, we fit polynomial continuum functions to the spectra from 3.1 to 4.2 $\mu$m, excluding the region between 3.3 and 3.7 $\mu$m. Figure 5 shows the continuum-subtracted spectra, where the 3.4 $\mu$m band is clearly discernible on all targets except for Polymele; it is possible that the relatively poor signal-to-noise ratio of the Polymele spectrum may be obscuring a weak absorption in this wavelength range that has a similar depth to the analogous features seen on the other less-red Trojans. To characterize these bands quantitatively, we first divided the spectra by the local linear continuum fits, anchored to the edges of the band ($\sim$3.25 and 3.57 $\mu$m). We then used a Gaussian to model the band and followed the same 10,000-iteration Monte Carlo procedure as with the 3.0 $\mu$m band. The resulting band center and depth values are reported in Table 3.

Analogous spectral features have been observed across a wide range of solar system small body populations, including main belt asteroids (e.g., Campins et al., 2010; Rivkin & Emery, 2010; Licandro et al., 2011; Takir & Emery, 2012; De Sanctis et al., 2018), the Jupiter-family comet 67P (Raponi et al., 2020), regular and irregular Jovian and Saturnian satellites (e.g., McCord et al., 1997; Cruikshank et al., 2008, 2014; Clark et al., 2012; Brown & Rhoden, 2014), and KBOs (e.g., Grundy et al., 2016; Stern et al., 2019; Brown & Fraser, 2023). In Figure 5, we plot the continuum-subtracted spectra of Ceres, Themis, 67P, Okyrhoe, and 2004 TY364. All of these spectra show a distinct 3.3–3.6 $\mu$m absorption feature that is characteristic of C–H stretch modes from aliphatic organics. Several individual component bands in this region are resolved in the high-precision Rosetta spectrum of 67P (Raponi et al., 2020), including the asymmetric stretching modes of the aliphatic CH₃ (3.38 $\mu$m) and CH₂ (3.42 $\mu$m) groups and the blended symmetric stretching band of CH₂ and CH₃ (3.47 $\mu$m). The presence of a distinct absorption feature at 3.3 $\mu$m in the spectrum of 67P and the relative broadening of Ceres’ organic absorption band toward shorter wavelengths have been attributed to the presence of aromatic hydrocarbons in addition to the aliphatic organics (e.g., De Sanctis et al., 2019; Raponi et al., 2020). No evidence for aromatic hydrocarbons can be found in the Trojan, KBO, and Centaur spectra.

Laboratory experiments have demonstrated that aliphatic organic compounds are readily produced through ion bombardment and ultraviolet irradiation of ice mixtures containing various combinations of H₂O, CH₄, CH₃OH, CO, NH₃, and N₂ (e.g., Strazzulla et al., 2001; Moore & Hudson, 2003; Palumbo et al., 2004; Cruikshank et al., 2005; Materese et al., 2014; Urso et al., 2020). An alternative pathway for forming aliphatic organics is through global aqueous alteration, a scenario that has been proposed for even small primitive objects, such as the parent bodies of the carbonaceous chondrite meteorites (e.g., Schulte & Shock, 2004; Vinogradoff et al., 2018). In addition to yielding the 3.3–3.6 $\mu$m absorption bands observed throughout the middle and outer solar system, these processes also readily reproduce the reddened and darkened surfaces that are characteristic of Trojans, KBOs, Centaurs, and irregular satellites.

Within this interpretation, the systematically stronger aliphatic organic bands on Orus and Leucus suggest that the primordial surface composition of red Trojans was significantly richer in CH-bearing species than the less-red Trojans. The complex interplay between relative ice abundance, grain size, and radiation dose that is evident from these past experiments, as well as the multiplicity of possible chemical pathways that produce the same aliphatic organic compounds, makes it impossible to link the observed organic absorption on Trojans to a specific initial ice mixture. While initial surface ice inventories incorporating hypervolatile species such as N₂ and CH₄ necessitate a cold outer solar system origin, the ability of mixtures containing CH₃OH to reproduce the broad 2.7–3.6 $\mu$m spectral feature means that a middle solar system origin for Trojans cannot be fully excluded, as methanol ice can persist on airless surfaces at 5–20 au on kiloyear to megayear timescales (e.g., Wong & Brown, 2016; Lisse et al., 2022).

In the spectrum of Ceres, the 3.4 $\mu$m band is accompanied by an additional major absorption feature at 3.7–4.0 $\mu$m. Past modeling of Ceres’ surface has revealed that this pair of bands is indicative of organic compounds mixed with a significant amount of carbonate-bearing species, which absorb strongly at both 3.4 and 3.9–4.0 $\mu$m (De Sanctis et al., 2015). The longer-wavelength band is particularly prominent within the bright regions in Occator Crater, where the absorption takes on a distinctive triangular shape with a reflectance minimum around 3.95 $\mu$m (De Sanctis et al., 2016; Carrozzo et al., 2018). The intrinsic scatter in the Trojan spectra throughout this wavelength range is poorer than at shorter wavelengths, and there are signs of increased correlated noise features. We do not detect any hints of a broad dip in reflectance and place upper limits of a few percent on the presence of a carbonate absorption at 3.7–4.0 $\mu$m.

The reflectance spectrum of Eurybates contains a distinct absorption band at 4.25 $\mu$m that is not seen on any other Trojan. This wavelength corresponds to the $\nu_{3}$ vibrational mode of CO₂ ice (e.g., Sandford & Allamandola, 1990; Hansen, 1997). Analogous spectral features are ubiquitous across the middle and outer solar system, from giant planet satellites (e.g., McCord et al., 1998; Cruikshank et al., 2010; Pinilla-Alonso et al., 2011) to Centaurs and KBOs (e.g., Cruikshank et al., 1993; Harrington Pinto et al., 2022; de Prá et al., 2024; Pinilla-Alonso et al., 2024). Figure 6 shows a sample of reflectance spectra displaying 4.25 $\mu$m absorption, plotted alongside the continuum-subtracted spectrum of Eurybates; the high spectral resolution JWST/NIRSpec spectra of Europa and Ganymede were taken from Trumbo & Brown (2023) and Trumbo et al. (2023), respectively. We analyzed the CO₂ band shape on Eurybates using a similar methodology to our analysis of the 3.4 $\mu$m band (Section 3.2). After subtracting the best-fit linear continuum through the edges of the band at $\sim$4.15 and 4.33 $\mu$m, we carried out a Monte Carlo analysis using a Gaussian band profile. The band center and depth are reported in Table 3.

On the cold surfaces of KBOs and Centaurs, this band, centered around 4.268 $\mu$m, is attributed to pure CO₂ ice (e.g., de Prá et al., 2024). However, at the typical temperatures of the Jupiter Trojans, CO₂ ice is extremely unstable, sublimating on day-long timescales (e.g., Lisse et al., 2022). Any exposed CO₂ ice on the surface would be ephemeral, such as in the case of 67P, where isolated patches of CO₂ ice in the southern hemisphere were found to persist during its cold season, when the local region was in shadow, before sublimating away on week-long timescales upon exposure to sunlight (Filacchione et al., 2016). Another example illustrating the instability of CO₂ ice in the Trojan region is the recent JWST observation of the active Centaur 39P/Oterna, which revealed CO₂ gas emission while the target was at 5.82 au (Harrington Pinto et al., 2023). It follows that the CO₂ on the surface Eurybates must be bound to some refractory component. Type II CO₂ clathrates (CO₂ bound in a crystalline water ice structure) are a salient possibility and have been forwarded as an explanation for the 4.27 $\mu$m band seen on Phoebe (Cruikshank et al., 2010). Laboratory studies have shown that the CO₂ band center for type II clathrates drifts to shorter wavelengths with increasing temperature, reaching $\sim$4.265 $\mu$m at the typical surface temperatures of Trojans (Blake et al., 1991). Meanwhile, shorter-wavelength absorption features may be due to mineral-bound CO₂ (Hibbitts et al., 2000, 2003) or irradiation-formed carbonic acid (Jones et al., 2014), which have band centers near 4.25 $\mu$m. Our measured band center for Eurybates ($4.258\pm 0.002$ $\mu$m) is broadly consistent with both of these possibilities. CO₂ can also be produced as a component of the organic-rich residue that forms from the irradiation of carbon-rich ice mixtures — the same process that yields the 3.4 $\mu$m aliphatic organic features described in Section 3.2 (Urso et al., 2020).

The absence of a discernible CO₂ feature on the other Trojan targets — both less-red and red — presents an interesting conundrum. Eurybates is the only member of a collisional family among the five targets observed with JWST, so we posit that its visible CO₂ absorption is related to its impact history. It is important to note that the estimated age of the Eurybates family (1–4 Gyr; Marschall et al. 2022) is significantly longer than typical space weathering timescales for Trojans provided in the literature, which are on the order of $10^{2--3}$ yr (e.g., Melita et al., 2015). Therefore, with respect to irradiation-related alteration of surface materials, Eurybates is considered mature. It follows that the differences between Eurybates and the other Trojans must be due to fundamental distinctions in the physical and/or chemical properties of the surfaces.

Here, we once again use the unique spectrum of Okyrhoe to develop a hypothesis for this phenomenon. While Okyrhoe shows a muted 3 $\mu$m feature with a depth that is intermediate between that of Trojans and KBOs, its spectrum lacks a discernible CO₂ absorption band. A subset of Centaurs has been observed to display cometary activity, with the activity concentrated among low-perihelion objects (e.g., Jewitt, 2009, 2015; Wong et al., 2019b). The exact trigger for the onset of Centaur activity is still not fully understood, but a leading hypothesis points toward the transition between amorphous to crystalline water ice and the concomitant release of trapped volatiles (e.g., Capria et al., 2000; Jewitt, 2009; Wierzchos et al., 2017). While Okyrhoe has never been observed to be active, its relatively close orbit places it in a region where the amorphous-to-crystalline transition can take place (Guilbert-Lepoutre, 2012). Physical modeling of cometary activity has demonstrated that fine-grained refractory material is entrained in the outgassed volatiles and can be redeposited across the surface (e.g., Jewitt, 2002). The obscuration of Okyrhoe’s surface due to past activity may explain the distinctly muted 3 $\mu$m OH band and the absence of 4.25 $\mu$m CO₂ absorption in its spectrum.

If Trojans initially formed as icy planetesimals in the primordial trans-Neptunian region, as current dynamical instability models suggest, they likely became active during their inward migration from the outer solar system. The increased temperatures and space weathering would have led to thermal processing of the initially volatile-rich surfaces. Due to the observed prevalence of CO₂ ice across the KBO population (de Prá et al., 2024; Pinilla-Alonso et al., 2024), it is possible that all Trojans developed surfaces rich in bound CO₂ (e.g., through the aforementioned irradiation process). Meanwhile, the activated outgassing at these higher temperatures would have caused simultaneous physical obscuration of the spectral features across the population. In the special case of Eurybates, the later family forming impact destroyed the deposited lag material on the parent body and exposed material from deeper layers, thereby providing an unobscured view of the 4.25 $\mu$m bound CO₂ band. Analogous JWST observations of other members of the Eurybates family will enable us to determine whether the CO₂ feature is indeed a distinctive feature of the collisional fragments. Recent photometric observations have confirmed a second collisional family within the Trojans — the Ennomos family (Wong & Brown, 2023). Comparisons of the two families could assess whether there are systematic differences in the way CO₂ is bound to the surfaces of Trojans.