Parent Volatile Outgassing Associations in Cometary Nuclei: Synthesizing Rosetta Measurements and Ground-based Observations

Comets, as remnants of the solar system’s formation, vary in volatile-refractory content. In situ comet studies, such as the Rosetta mission to 67P/Churyumov–Gerasimenko, provide detailed volatile composition insights, while ground-based studies offer broader comet samples but in fewer species. Comparing 67P’s volatile correlations during the 2 yr Rosetta mission with those from remote sensing gives insights into volatile distribution in the nucleus and factors influencing their release. Our goal is to identify associations between volatiles seen from the ground and those in 67P. Given 67P’s seasonal variations, we segmented the Rosetta mission around 67P into six epochs, reflecting different insolation conditions. It has been suggested that there are at least two different ice matrices, H2O and CO2 ice, in which the minor species are embedded in different relative abundances within them. We employed various methodologies to establish associations among volatiles, such as volatile production rates, spatial distributions, patterns in mixing ratio, and local outgassing source locations. We note that different techniques of grouping molecules with respect to H2O and CO2 may yield different results. Earth’s atmosphere blocks CO2; however, due to observed differences between H2O and C2H6 from the ground and between H2O and CO2 from comet missions, C2H6 is suggested to be a CO2 proxy. Our study delves into cometary coma molecular correlations, highlighting their associations with H2O and CO2 matrices and advancing our understanding of the early solar system comet formation and evolution.


Introduction
Comets represent a diverse population of small, volatile-rich remnants from the era of solar system formation.They encapsulate a broad spectrum of volatile and refractory materials, reflecting the diverse conditions present during their formation in the protosolar nebula between 5 and 30 au (or more) from the Sun, before migration of the giant planets ejected them into their current stable dynamical reservoirs, the Oort cloud and the scattered Kuiper disk (Nesvorný et al. 2017;Vokrouhlický et al. 2019).Volatiles within the comet interior are protected by an overburden of material and thus may largely retain their original composition and reflect the conditions present in the protoplanetary disk where (and when) they formed (see Bockelée-Morvan et al. 2004;Mumma & Charnley 2011;Müller et al. 2022).Thus, comet composition measurements provide a plausible link to the initial conditions and subsequent evolution of the early solar system (e.g., see Öberg et al. 2015;Willacy et al. 2022).The nuclei of active comets are obscured by the gas and dust surrounding them, making it impossible to observe the nucleus directly unless visited by a spacecraft.Spacecraft visits to comets to date show that their surfaces are covered with refractory materials with very little exposed ice (e.g., Luspay-Kuti et al. 2018;Combi et al. 2020).The primary volatile composition of the coma has been used as a proxy for the chemical composition of the nucleus, but the accuracy of this assumption depends on how volatiles are stored in and released from the nucleus and the nature of chemical reactions in the inner coma.Only six comets have been visited by spacecraft, and only Rosetta remained in orbit and had the instrumentation to characterize the temporally resolved chemical makeup of a single comet (67P/Churyumov-Gerasimenko, hereafter 67P) in detail.Therefore, the information on a large and diverse population of comets necessarily relies on remote sensing.Physical and chemical correlation trends have been suggested for both individual comets and comet populations (e.g., Dello Russo et al. 2016b;Luspay-Kuti et al. 2018).The distributions of volatiles in the coma as measured at different spatial scales by remote sensing and in situ measurements suggest that minor species (e.g., CH 3 OH, CO, H 2 CO, CH 4 , and OCS) may be embedded within the ice matrices of the most abundant species, polar H 2 O or nonpolar CO 2 .In this paper, we are investigating the possible synergy between findings on volatile associations from in situ measurements in comet 67P with those derived from groundbased measurements.Therefore, we are delving deep to answer an essential question: are volatiles associated with common or distinct outgassing sources?In Section 2, we go over methods of establishing associations or differences in volatile release, which provides clues to how ices are stored in comet nuclei.In Section 3, we review the trends observed in comets to date based on these methods.Finally, in Section 4, we discuss results and present challenges that exist in understanding how volatiles are associated (or separated) in cometary nuclei.

Methods to Establish Associations
When comparing ground-based and Rosetta results, we infer associations in volatile release and storage based on multiple criteria.In this section, we review these various approaches to exploring associations in outgassing behavior using both remote sensing and spacecraft in situ measurements.

Short-term Temporal Variability in Measured Production Rates
Due to the rotation of a likely locally heterogeneous nucleus and the illumination of different parts of it, the volatile production rates of different species can follow the same diurnal variability pattern and change on timescales of a few hours and/or days.Short-term variability in production rates has been observed in many comets to date.If nonuniform shortterm production rate variations are observed, then the relative mixing ratios (with respect to H 2 O for most cases) could be variable.For instance, in comet 67P, some volatiles (e.g., CH 4 ) showed diurnal variations that differed from those for most other volatiles (e.g., Luspay-Kuti et al. 2015, 2019;Bockelée-Morvan et al. 2016;Fink et al. 2016).In comet 9P/Tempel 1, some short-term variability in production rate was observed due to the effect of the impact (e.g., Mumma et al. 2005).Comet 103P/Hartley 2 (hereafter Hartley 2) exhibited significant short-term variability in activity and heterogenous outgassing, where the overall gas and dust production increased by ∼60% over the course of ∼5-5.5 hr, with C 2 H 6 increasing more rapidly than other species (Dello Russo et al. 2011;Kawakita et al. 2013).In comet 73P, the bulk composition of different fragments was the same; however, variations in parent volatiles were observed in fragments B and C (see Villanueva et al. 2006;Dello Russo et al. 2007).Simultaneous measurements of multiple species can be used to establish which molecules follow similar patterns in variability (see Figures 1,2,3,and 6 and  Rotation of the nucleus can reveal chemical heterogeneity as different parts rotate in and out of solar radiation.The obliquity of the rotation axis determines which latitudes are heated throughout a comet's orbit.Seasonal variations have been observed in multiple short-period comets such as 67P (Läuter et al. 2020), 19P/Borrelly (Schleicher et al. 2003), and 2P/Encke (A' Hearn et al. 1985;Roth et al. 2018).In some cases, comets exhibit complex rotation, where most parts of the nucleus might be illuminated on short time spans (e.g., Hartley 2).Space missions to comets, such as Rosetta to comet 67P and EPOXI to comet Hartley 2, revealed distinct and common sources for different volatiles (e.g., see Protopapa et al. 2014;Fink et al. 2016;Biver et al. 2019;Lai et al. 2019;Laüter et al. 2022).
Volatile abundances may in some cases be subject to evolutionary, seasonal, and heliocentric effects; for instance, in C/2009 P1 (Garradd), the production rate of CO increased even after the comet passed perihelion (Bodewits et al. 2014;McKay et al. 2015), perhaps owing to a new layer of the comet's nucleus being exposed to solar irradiation, whereas the production rate of H 2 O followed the predicted heliocentric dependence (decreasing as the comet passed perihelion; see Bodewits et al. 2014;McKay et al. 2015).These nonuniform production rates and mixing ratios of CO were observed in comet C/2009 P1 (Garradd) by both groundbased studies (McKay et al. 2015) and space-based studies from the High-Resolution Instrument Infrared Spectrometer on board the Deep Impact Flyby spacecraft (Feaga et al. 2014).A similar effect was observed in 2I/Borisov, where the CO/H 2 O ratio increased over time (Bodewits et al. 2020;Cordiner et al. 2020).The depleted CH 3 OH reported during 2P/Encke's 2017 apparition (Roth et al. 2018) compared with its enriched abundance in 2003 (Radeva et al. 2013), plus other compositional differences observed in 2017 compared to 2003, may have also resulted from seasonal effects in 2P/Encke's nucleus (Roth et al. 2018).However, the observed differences (specifically in short-period comets) can be due to evolutionary effects and/or the dependences of volatile production rates on heliocentric distance (e.g., see Section 4.2.1 in Roth et al. 2018).These effects must be considered when placing individual comets into the context of a comet population and investigating correlations between species.

Patterns in Chemical Mixing Ratios (Relative Abundances)
across Different Comets/Populations (Taxonomies) Based on early observations of a few comets at near-IR wavelengths, three basic taxonomic classes based on volatile composition were put forward: organics-depleted, organicsnormal, and organics-enriched (Mumma & Charnley 2011).However, additional observations have revealed a more complex sorting with comet composition comprising more of a continuum or hybrid of properties in relation to these original basic taxonomic classes (i.e., comets can be depleted in certain parent volatiles while enriched in others; Villanueva et al. 2006;Gibb et al. 2012;Radeva et al. 2013;Dello Russo et al. 2016a;Roth et al. 2017Roth et al. , 2018;;Saki et al. 2021;Harrington-Pinto et al. 2022).Recently, more detailed taxonomic classifications and volatile relationships in comets indicating this hybrid classification have been suggested (see Dello Russo et al. 2016a andLippi et al. 2021 for more details).

Spatial Distribution (Spatial Profile) of Volatiles
It is possible to map the distribution of volatiles in some cometary comae for species that are both abundant enough and have sufficiently strong transitions.For instance, after centering a comet within a long slit, near-IR spectroscopy can map the intensity of light along the slit at wavelengths of particular molecular emissions.Multiple emissions from the same species can be coadded to obtain this distribution along the slit (the spatial profile) of that volatile species (Dello Russo et al. 2016aRusso et al. , 2022)).At millimeter and submillimeter wavelengths, interferometry can directly image the 3D spatial-spectral distribution of molecules (Cordiner et al. 2014;Roth et al. 2021;Biver et al. 2022).These spatial profiles can then be used to investigate and compare how these molecules are released into the comet's coma from its nucleus.The shape of volatile spatial profiles can reveal whether specific volatiles are directly sublimating from the nucleus (and can be tied back to the same location) or released into the coma from extended sources through the dissociation of larger molecules or the evaporation of grains in the coma (Dello Russo et al. 2016a;Bonev et al. 2021;Biver et al. 2022).In general, ices sublimating directly from the nucleus exhibit a spatial profile that peaks in intensity, presumably at or near the nucleus, and then falls off with an r −2 dependence, where r is the nucleocentric distance.On the other hand, spatial profiles of molecules produced by photolysis or extended sources in the coma fall off more slowly with a flatter spatial distribution (see Section 3.3.3).Comparison with coma models allows distinguishing between the release of icy grains in the coma (as EPOXI found in comet Hartley 2 for water) versus direct production from ices in the nucleus (as Rosetta found for water in comet 67P; Biver et al. 2019).It is important to note that distinguishing volatiles originating from the nucleus from those emitted by icy grains becomes challenging when grain lifetimes are short or if spatial resolution is poor.This can be true either when the comet is at smaller heliocentric distances, where the lifetimes of ices within the solar radiation field are reduced, or when the comet is at large geocentric distances, where the spatial scales per pixel are larger.In either case, the extent of the icy coma may be confined to a small area relative to the field of view (see Protopapa et al. 2018 for more details).In addition, volatiles emitted from icy grains typically have a smaller falloff in temperature with distance from the nucleus than those originating from the nucleus (Bonev et al. 2021;Fougere et al. 2012).
The spatial distribution of parent molecules is related to how they are stored as ices in the nucleus.Thus, volatile ices that are associated with each other in the nucleus should tend to have similar spatial distributions in the coma, whereas dissimilarities may reflect a segregation of these nucleus sources.The existence of two ice phases for polar molecules such as H 2 O and CH 3 OH and nonpolar molecules such as CO 2 and C 2 H 6 has been put forth as one possible explanation for the differences observed in volatile spatial distributions in the coma (see Section 4.4 for details).We also note that linking the observed coma morphology to the distribution of active areas on the nucleus is a difficult task (even for in situ measurements) because jetlike features can be a result of diffuse or even uniform activity over topographic concave areas, such as the ones seen on comet 67P (Shi et al. 2018).Furthermore, the projection of coma structures can exhibit jetlike features that can vary based on the perspective of the observer (see Shi et al. 2018).

Location of Local Volatile Outgassing Sources
Owing to the large orbital obliquity of 67P (∼52°), the illumination of the surface changes significantly throughout the comet's orbit, which causes strong seasonal variations in outgassing (Hässig et al. 2015;Keller et al. 2015;Le Roy et al. 2015;Luspay-Kuti et al. 2015).Latitudes and longitudes of the sources responsible for outgassing of different volatiles on the surface of comets can be used to investigate if volatiles are subliming from common or distinct sources on the nucleus.Space missions to comets such as 67P and Hartley 2 revealed distinct and common sources for different volatiles such as H 2 O and CO 2 (e.g., A' Hearn et al. 2011;Protopapa et al. 2014;Fink et al. 2016;Biver et al. 2019;Lai et al. 2019;Combi et al. 2020;Laüter et al. 2022).However, because of the growing mean free path for molecular collisions, it is a difficult task to directly connect observed species to specific sources on the surface from in situ measurements, but only to general areas (Marschall et al. 2019).To trace local measurements to global production rates and nucleus outgassing sources, accurate outflow models are needed.Läuter et al. (2020) used an analytical model for the expansion of a collisionless gas into space with an optimization procedure to constrain a large number of emission sources, whereas Fougere et al. (2016) and Combi et al. (2020) similarly constrained emission sources, accounted for solar illumination, and added Monte Carlo modeling for direct sublimation of major volatiles, including molecule-molecule collisions in 67P, which could then be compared with remote sensing by VIRTIS and MIRO at any given time.

Volatile Associations Inferred from Rosetta and Groundbased Measurements
To investigate the associations of parent volatiles on long and short timescales from ROSINA measurements in comet 67P and present and compare correlations reported in the literature along with the effects of seasonal insolations on production rates of different species, we have divided 67P's data into six epochs.These epochs are chosen to encompass times before, after, and between the two equinoxes and the perihelion (see Tables 1 and 2).To determine the correlation between the abundances of different molecules, we used three different statistical measures: R 2 , slope (S), and Pearson correlation (r).R 2 measures the proportion of variance in the dependent variable (molecular production rate) in relation to the independent variable (H 2 O or CO 2 production rate).A higher R 2 value indicates a stronger correlation between the variables.In fitting a line to the production rate of a volatile species versus H 2 O or CO 2 , the slope represents the change in the dependent variable (molecular production rate) for a unit change in the independent variable (H 2 O or CO 2 production rate).A positive slope indicates a positive correlation between the variables, while a negative slope indicates a negative correlation.The Pearson correlation measures the strength and direction of the linear relationship between two variables.A correlation coefficient (r) of 1 indicates a perfect positive correlation, while a correlation coefficient of −1 indicates a perfect negative correlation, or anticorrelation.A correlation coefficient of 0 indicates no correlation between the variables.In our study, we found that the correlation of different molecules with either H 2 O or CO 2 varied depending on the epoch (see Section 3.2 and Table 2).By considering the correlation factor (R 2 ) values alongside the Pearson correlation coefficients and regression slopes, we can garner a more comprehensive understanding of the nature and magnitude of these relationships.Such an approach enables us to draw more informed conclusions regarding the mechanisms driving the release of gases from comet 67P.

H 2 O and CO 2 Ices as the Bases for Chemical Associations
In 67P, the release of H 2 O and CO 2 is generally not correlated, except during the southern summer, and at these times, all gases have a moderate to strong correlation with each other (see Table 1 and Figures 1 and 2; also see Fougere et al. 2016;Gasc et al. 2017;Läuter et al. 2020;Luspay-Kuti et al. 2022 Note.R h is the heliocentric distance.SL is the subsolar latitude of comet 67P.Sublimation temperatures of ices are presented for each molecule in parentheses (Palumbo et al. 1995;Ferrante et al. 2008;Yamamoto & Ashihara 1985; see Table 1 in Meech & Svoren 2004).The conflict between some of the papers (e.g., epochs 1 and 6 for HCN) could be due to using different techniques and instruments and the fact that these epochs consist of portions with water still driving cometary activity and portions with water cooled off, as our chosen intervals include a large change in the heliocentric distance.S23: this work (based on R 2 ; the correlation factors for all molecules between the two equinoxes are high with respect to both H 2 O and CO 2 , indicating that most species are correlating with one another in epochs 3-5, and therefore we chose the higher of the two values).G17:  Rubin et al. 2015;Keller et al. 2015;Hu et al. 2017;Keller et al. 2017;Fulle et al. 2018Fulle et al. , 2020)).
Based on the methodologies presented in Section 2, we suggest that there are at least two different ice matrices, H 2 O and CO 2 ice, in which the minor species are embedded in different relative abundances (also see Luspay-Kuti et al. 2015;Gasc et al. 2017;Hu et al. 2017;Luspay-Kuti et al. 2019;Läuter et al. 2020;Luspay-Kuti et al. 2022).For this reason, we compare all other volatiles with respect to either H 2 O or CO 2 .Note.Linear regression, slope of the line, and Pearson correlation are shown as R 2 , S, and r, respectively.The correlation factors for all molecules between the two equinoxes are large with respect to both H 2 O and CO 2 , indicating that most species are correlating with one another in epochs 3-5.We chose the higher of the two values of R 2 for the derived long-term associations presented in Table 1.The significance of these correlation depends on the sample size.In our analysis and based on the data from Läuter et al. (2020), we are in a sample size regime where the entire mission data are statistically significant; therefore, we are using the entire mission when referring to association between species.

Associations among Parent Species with H 2 O and CO 2 in 67P
Rosetta orbited comet 67P before, during, and after its perihelion passage (2015 August 13.09) and revealed that the observed associations among species, and their correlation or lack thereof, varied depending on the hemisphere predominating the volatile release (e.g., Biver et al. 2019;Combi et al. 2020).This was largely dictated by 67P's orbital tilt and position at the time of measurement and the specific hemisphere under consideration.Changes in the associations of certain volatiles relative to H 2 O and CO 2 have been previously reported in the literature (e.g., Luspay-Kuti et al. 2015;Gasc et al. 2017;Luspay-Kuti et al. 2019;Läuter et al. 2020;Luspay-Kuti et al. 2022).Figures 2(A Our findings, summarized in Tables 1 and 2, also highlight discrepancies in the observed associations of volatiles with respect to H 2 O and CO 2 in 67P due to variations in heliocentric ranges or statistical techniques used in different studies.A) and (B), there is probably a more complex relationship between the molecules, especially if these are not connected to the identical source regions on the nucleus (Biver et al. 2019;Laüter et al. 2022).The subsequent sections will explore the association of individual molecules with H 2 O and CO 2 .1. CO. Combi et al. (2020) reported that prior to 67P's first equinox (2015 May 10), the coma abundance CO (and CO 2 ) showed poor correlation with H 2 O.However, after the first equinox, as the southern hemisphere became illuminated, their correlation increased.2015,2019) and is also grouped under the CO 2 -correlated category by Läuter et al. (2020).In our analysis, the average production rate of C 2 H 6 showed a slightly better correlation with H 2 O than CO 2 in epochs 1 and 3 but an overall better correlation with CO 2 than H 2 O in the entire mission (see Tables 1 and 2).9. H 2 S, OCS, and CS 2 .Gasc et al. (2017) showed that H 2 S correlates well with CO 2 near and post-second equinox.Läuter et al. (2020) grouped OCS, CS 2 , and H 2 S under the CO 2 -correlated category by their power-law component.While our analysis showed that a better correlation exists between these sulfur-bearing molecules and H 2 O in some of the epochs, the entire Rosetta-ROSINA mission data indicate a higher positive correlation with CO 2 than H 2 O (see Tables 1 and 2).10.C 2 H 5 OH.Based on the power-law component shown in Läuter et al. (2020), C 2 H 5 OH is grouped under the CO 2 -correlated category.The overall production rate of C 2 H 5 OH shows a higher correlation with CO 2 than H 2 O for the entire mission.11.H 2 CO.H 2 CO shows an overall higher correlation with H 2 O than CO 2 (see Tables 1 and 2).Similarly, it is grouped under H 2 O-correlated gases by Läuter et al. (2020).

Associations among Species from Ground-based Measurements
Knowledge of volatile production rates (molecules s −1 ) and mixing ratios (with respect to H 2 O, C 2 H 6 , etc.) forms the basis of ground-based studies of cometary chemistry.The existence (or lack thereof) of correlation trends between relative abundances of parent molecules observed in ground-based studies depends on the comet population class.This includes comets with a vast variety of chemical properties, from different reservoirs-either the Kuiper Belt (Jupiter-family comets, JFCs; and Centaurs) or the Oort cloud (Oort cloud comets, OCCs)-and with different degrees of subsequent evolution (which is still not well understood) due to solar irradiation.In ground-based observations, the correlation of volatiles can also be evaluated based on their spatial distributions in the coma, which may reflect how they were stored in the nucleus.This can involve distinguishing between volatiles released directly from ices in the comet's nucleus and those that are released or formed from other (extended) sources.However, it is important to note that most ground-based observations are snapshot measurements, and the interpretation depends on excitation and radiative transfer models.This can cause discrepancies between near-IR, radio, and optical cometary studies, which must be considered when placing these measurements into context.

Treatment of CO 2 from Ground
Because CO 2 lacks a stable dipole moment and has a substantial presence in the Earth's atmosphere, its direct detection is only possible from space.Thus, the production rate and abundance correlations for CO 2 in comets that are solely observed by groundbased instruments are typically estimated based on the correlation observed in comets visited by spacecraft or observed using spacebased telescopes, generally at infrared wavelengths (e.g., AKARI and the NEOWISE space telescope; Ootsubo et al. 2012).CO 2 can be indirectly measured from space at UV wavelengths (1900-2300 Å) through detection of CO Cameron bands that are excited by electron impact and photodissociation of CO 2 (Weaver et al. 1994;Raghuram & Bhardwaj 2012).The only ground-based method of inferring CO 2 production in comets is through optical observations of forbidden oxygen emission, if the Doppler shift is large enough to separate cometary and telluric oxygen emission (for more details, see Festou & Feldman 1981;McKay et al. 2012;Decock et al. 2013;Raghuram & Bhardwaj 2014;McKay et al. 2015McKay et al. , 2016;;Harrington-Pinto et al. 2022).However, the oxygen emission can be produced by H 2 O as well (a challenge in using this method to infer CO 2 production; see Sections 2.1 and 2.2 in Harrington-Pinto et al. 2022 for details).Because of this difference between CO 2 detection techniques and ground-based observations of comet parent volatiles and the consequence that measurements are rarely obtained at the same time, interpretation is subject to both methodology differences in determining production rates and accounting for temporal variability.

Taxonomies and Relative Abundances
Investigation of chemical composition across comet populations can reveal valuable insights about the existence of possible trends among comets.Dello Russo et al. (2016b) studied the trends in volatile production rates and mixing ratios in comet populations and showed that CH 3 OH, HCN, C 2 H 6 , and CH 4 abundances (relative to H 2 O in percentage) were correlated with each other.H 2 CO and NH 3 abundances were correlated with each other but not correlated with other presented volatiles.The C 2 H 2 abundance showed weak correlations with H 2 CO, NH 3 , and C 2 H 6 ; however, it had a stronger correlation with HCN abundance.The abundance of hypervolatile CO showed a weak correlation with other hypervolatile CH 4 (which is more sensitive to thermal processing; see Roth et al. 2018Roth et al. , 2020)).
Using the mixing ratios of volatiles in 20 comets observed with the Keck telescope, Lippi et al. (2021) investigated the association between volatiles using a Spearman correlation.They found that hydrocarbon pairs (C 2 H 6 -CH 4 , CH 4 -C 2 H 2 , and C 2 H 6 -C 2 H 2 ) show a high value of the Spearman correlation with respect to one another.CO showed a medium correlation with NH 3 and CH 3 OH (similar to NH 3 -C 2 H 2 ), while the correlation between CO mixing ratios and hydrocarbon ranges from 0.5 to 0.8 (see Table 5 in Lippi et al. 2021 for details).Although Lippi et al. (2021) did not find correlations between HCN and hydrocarbons, Dello Russo et al. (2016b) found a high positive correlation between HCN and hydrocarbons.This difference in findings could be due to the use of different data sets and approaches or the still-small sample sizes of the individual databases.
It is also important to note the shortcomings to developing a chemical taxonomy of comets when determining relationships.In the case of high-resolution near-IR and radio measurements, the number of comets sampled is still small (∼45) compared to the >200 comets characterized in the optical and near-UV.However, whereas IR and radio observations can sample parent volatiles, optical and near-UV measurements are restricted to daughters of often uncertain parentage.As the Rosetta results of 67P have shown, ground-based taxonomic studies can only measure a very small subset of volatiles present within a comet nucleus, and as such, global interpretations are necessarily limited.
Product species such as OH, NH, CN, CH, C 2 , C 3 , and NH 2 have been observed in comets in the optical and UV regions, with variation reported among the population; for example, ∼30% of comets are depleted in C 2 /CN ratio.This trend is supported by optical spectral surveys at both high and low resolution (see Fink 2009;Cochran et al. 2015).Studies of fragment species in the optical and near-UV wavelengths (∼0.3-0.9 μm) showed that comets can be classified into seven compositional classes (Scheicher & Bair 2014;Bair et al. 2020), either typical or carbon chain-depleted, with several subgroups of the depleted type, as defined by A' Hearn et al. (1985).The study of daughter species, such as ionized particles in the coma or tail, can provide valuable information about the comet's environment and interactions with the solar wind.There are diverse chemical and physical processes involved in the production of daughter species (e.g., CN and C 2 ), and an individual daughter product may have multiple possible precursors (Feldman et al. 2004).Thus, abundances in the optical/UV may not directly reflect the composition of the nucleus, emphasizing the continued need to study parent (directly released from the nucleus) volatile composition (see Section 4.1 and Dello Russo et al. 2016b).
Observations in millimeter and submillimeter wavelengths using radio facilities (e.g., Nancay; the Atacama Large Millimeter/ submillimeter Array, ALMA; and IRAM) can provide measurements of production rates and the spatial distribution of molecules such as H 2 CO, NH 3 , HCN, CO, and CS (e.g., see Figure 3 in Biver et al. 1997 andFigure 5 in Crovisier et al. 2016).Multiple investigations have been conducted to classify comets observed at radio wavelengths (e.g., Crovisier et al. 2009;Biver et al. 2022).These studies found that, with the exception of CO, which is more abundant in OCCs, there is no clear distribution of volatiles that would allow for a taxonomy based on measurements at radio wavelengths (see Figure 10  Exploring the spatial profiles of volatiles in comets uncovers pronounced differences in the behavior of certain volatiles with respect to one another.For instance, H 2 O and C 2 H 6 , which are generally the easiest molecules to obtain spatial distributions in cometary comae from the ground at IR wavelengths, commonly show distinct spatial profiles (e.g., Dello Russo et al. 2016b;Bonev et al. 2021;Khan et al. 2021;Roth et al. 2021;Dello Russo et al. 2022).These differences can imply that these volatiles have been stored in different ice matrices within the nucleus.Spatial profiles of some volatiles (e.g., H 2 CO, HCN and its isomer HNC, CS, and CO) are accessible through emission lines at radio frequencies (e.g., see Cordiner et al. 2014).The importance of radio interferometry in obtaining spatial maps for molecules released from different sources has been well demonstrated in the literature (e.g., Boissier et al. 2007;Cordiner et al. 2014Cordiner et al. , 2017;;Roth et al. 2021).Various radio interferometric maps can yield spatial information on the same scales as near-IR measurements (e.g., Plateau de Bure, Noema, ALMA) and distinguish differences seen in the spatial distribution of volatiles (e.g., see Crovisier et al. 2016;Biver et al. 2022).For instance, H 2 CO has been observed and mapped with radio telescopes in many comets to date (e.g., C/ 2012 S1 (ISON), C/2015 ER61, and C/2012 F6 (Lemmon)).These radio maps show that molecules such as HNC and H 2 CO are progressively released in the coma from extended sources and could be correlated with dust production as well, whereas HCN seems to come directly from the nucleus ices.If extended sources and icy grains exist in the coma, the rotational temperature can peak off the nucleus, providing evidence that icy grain vaporization can be a major source of heating in the coma (e.g., 46P/Wirtanen, Bonev et al. 2021 Fougere et al. 2012).CO emission lines are present at near-IR, millimeter, and submillimeter wavelengths and can provide important information about the distribution of volatiles at larger R h , where H 2 O is not fully activated (e.g., C/2016 R2; Cordiner et al. 2022).However, the detection of CO at near-IR wavelengths requires a high geocentric velocity to Doppler shift its emissions away from their strong terrestrial counterparts, a limitation that CO measurements at radio wavelengths do not have.
The differences and similarities in the spatial profiles of parent species observed simultaneously and at different wavelengths such as CO, HCN, and H 2 CO can increase the number of molecules and comets with these measurements and help our understanding of how different release mechanisms influence the distribution of volatiles in cometary comae.

Overlap between 67P and Ground-based Measurements
Comet 67P has two lobes and a high axial tilt (∼52°).Therefore, it experiences strong seasonal effects.Gasc et al. (2017) showed that CO 2 , CO, H 2 S, CH 4 , and HCN abundances had a strong south-north heterogeneity for the entire period of their observations (R h ∼ 2.0-2.7 and 3.1-3.5 au postperihelion).Therefore, the hemisphere over which these measurements were taken can make a difference in interpreting the bulk nucleus chemistry (see Gasc et al. 2017 andLuspay-Kuti et al. 2019).In 67P, before and after the two equinoxes, the spatial distributions of volatiles (e.g., H 2 O and CO 2 ) look distinctive and with the smallest illumination; CO 2 and CO primarily originate from the southern hemisphere (M.6 in Fougere et al. 2016).Observations of Hartley 2 by EPOXI were limited in temporal coverage during the flyby, but similar compositional heterogeneity in CO 2 and H 2 O gases and ice was seen between the small and large lobes.
Most ground-based cometary production rate measurements are snapshots and biased toward a limited range of heliocentric distance mostly around ∼1 au, the brightest comets, and for the easiest to detect molecules such as H 2 O, C 2 H 6 , HCN, and CH 3 OH.Thus, a relative paucity of measurements exists for a large range of heliocentric distances, fainter comets including JFCs, and more difficult-to-detect molecules (e.g., NH 3 and C 2 H 2 ).The associations derived from ground-based IR measurements in Dello Russo et al. (2016b) and Lippi et al. (2021) are based on global coma abundances, which lack the spatial resolution to sense the heterogeneous outgassing to the degree seen in comets 67P and Hartley 2 by spacecraft.
Table 3 provides an overview of the molecules observed and the different association methodologies employed in groundbased measurements compared to the Rosetta ROSINA observations of comet 67P.This table offers a comprehensive comparison of the two data sources, enabling a better understanding of the overlap and disparities between them.The associations of observed species from the Rosetta mission to comet 67P include various methodologies, such as long-and short-term production rates and power-law relationships with respect to R h .On the other hand, ground-based measurements have primarily focused on chemical abundances, spatial profile analyses, seasonal, evolutionary effects, and short-term production rate variability.For instance, Dello Russo et al. (2016b) et al. (2017) did not find any correlation between outgassing properties and the sublimation temperature of the corresponding pure ices (see Table 1).MIRO observations show that NH 3 , CH 3 OH, and HCN ices with similar sublimation temperatures (78, 99, and 95 K, respectively; i Based on the short-term variation of volatiles observed in many comets (e.g., see Dello Russo et al. 2016b;Biver et al. 2018Biver et al. , 2022;;Roth et al. 2020;Lippi et al. 2021;Saki et al. 2021;Khan et al. 2023).

Gasc
Yamamoto & Ashihara 1985) have distinct spatial distributions and mixing ratios and sometimes no correlation in their behavior (e.g., Biver et al. 2019; see also Table 1).Thus, the sublimation temperatures of the pure ices are found to be uncorrelated with the slope of the decrease or increase of the production rates or coma content for the volatile species.Similar conclusions were found for Hartley 2, as HCN and CH 3 OH (with similar vacuum sublimation temperatures) had different spatial distributions, with CH 3 OH and H 2 O showing similar antisunward enhancements in their spatial distributions, while HCN and hypervolatile C 2 H 6 showed a more sunward enhancement (see Dello Russo et al. 2011;Mumma et al. 2011;Bonev et al. 2013;Kawakita et al. 2013).

Discussion
Cometary coma measurements of the composition of primary volatiles that are subliming directly from the nucleus have been used as a proxy of the bulk nucleus abundance.However, the Rosetta mission to comet 67P and ground-based measurements of relative abundances across multiple apparitions of some comets (e.g., 2P/Encke and 21P/Giacobini-Zinner) or taxonomic studies covering a sample of comets within the same class have shown that snapshot measurements of primary volatiles may not directly reflect ice abundances.These results may suggest that the composition of cometary ices within a comet nucleus may typically be heterogeneous, with measured coma abundances dependent on particular active areas on the nucleus.For instance, due to its high axial tilt, cometary coma measurements of 67P showed strong seasonal effects across its orbit.Rosetta revealed 67P's contact binary nucleus, adding more complexity to the volatile measurements tying back to the bulk nucleus composition.The tumbling motion of comet Hartley 2 revealed by the EPOXI spacecraft caused different areas of the comet to be activated that would be otherwise hidden from ground-based observations.The variations seen in coma measurements can include (but are not limited to) dependence on heliocentric distance, as some volatiles are not fully subliming at large heliocentric distances (e.g., H 2 O), and evolutionary processes preferentially depleting some species relative to others, which may be most visible in the JFC class due to their short orbital periods and multiple close solar passages.
All space-based missions have been focused on JFCs due to their easier accessibility compared to OCCs, which come to perihelion or closest approach with high velocities and random inclinations, therefore making cometary space missions to OCCs quite challenging.On the other hand, the brighter OCCs are generally easy to measure with ground-based observations.Among the molecules studied in the near-IR, there is a paucity in measurements of certain harder-to-detect species (e.g., H 2 CO, C 2 H 2 , and OCS; see Dello Russo et al. 2016b) in both JFCs and OCCs, making developing a taxonomy and investigating the correlation between these species a challenging task.Among these, the sulfur-bearing molecule OCS has been sampled in only three JFCs and seven OCCs so far (Saki et al. 2020;Faggi et al. 2023).Atomic sulfur emissions require observations in the UV/near-UV wavelengths (which specifically requires spacebased observatories, as the UV is inaccessible from groundbased facilities) and the radiative transfer modeling of the atomic sulfur transitions.More measurements of sulfur-bearing molecules will be required in order to fully investigate the correlation between these species.Furthermore, a multiwavelength campaign to observe all volatiles in all bandpasses depends not only on the brightness and productivity of comets but also on the availability of instruments capable of performing these types of measurements, which is quite challenging at heliocentric distances of more than 2 au.State-of-the-art observatories such as JWST, with high sensitivity, can detect emission features at larger heliocentric distances and for fainter targets (e.g., see Kelley et al. 2023).

Outliers
Most comets become active, and their sublimations begins, as they get closer to the inner parts of the solar system (R h < ∼3 au).The majority of comets that form the current taxonomies were observed at ∼1 au.Thus, there is a paucity of measurements and investigation of correlations of parent volatiles of comets closer to the Sun (see Biver et al. 2011;Dello Russo et al. 2016b;Lippi et al. 2021).For example, DiSanti et al. (2016) found that HCN became enriched in comet C/2012 S1 (ISON) at a small heliocentric distance (0.26% at R h = 0.43 au) relative to measurements at a larger heliocentric distance (0.07% at R h = 0.82 au).Roth et al. (2018) observed a similar HCN trend in comet 2P/Encke with an increase from 0.09% at R h ∼ 1.2 au in 2003 (Radeva et al. 2013) to 0.17% at R h = 0.45 and 0.11% at R h = 0.53 au.Other volatiles such as H 2 CO, C 2 H 2 , and NH 3 also show enrichment as comets get closer to the Sun (Dello Russo et al. 2016b).Therefore, in order to fully understand the possible trends in chemical mixing ratios and search for correlations among species, more measurements of all parent volatiles at smaller heliocentric distances are required.Moreover, no parent volatile measurements have been reported in some comets, such as comet 96P/ Machholz 1, with extreme anomalous composition in fragment species that does not fit in any taxonomic class (Schleicher 2008).Therefore, investigation of a correlation between parent species in such comets will have a high potential impact.Some short-period comets exhibit variability across multiple apparitions; for instance, comets 2P/Encke and 21P/Giacobini-Zinner showed variability across apparitions (see Radeva et al. 2013;Roth et al. 2018;Faggi et al. 2019;Roth et al. 2020).One possible explanation could be thermal processing, so that one layer of cometary material was removed, and a new, compositionally different layer was exposed during the new apparition, leading to the observed variability in composition.Therefore, to understand the role of thermal processing and its effects on the measured composition and thus correlation between species, more observations of short-period comets on consecutive apparitions are required.In some long-period comets, such as comet C/2016 R2 (PanSTARRS), the dominant ice was CO (McKay et al. 2019).Interstellar comet 2I/Borisov showed the same CO dominance compared to the rest of its volatiles (Bodewits et al. 2020).Therefore, more observations of CO-rich comets are necessary to fully understand their compositions, diversity, and correlation of their molecular species.

Fragment Species
JFCs and OCCs could have experienced different thermal processing due to the repeated close passages of JFCs compared to OCCs, thus leading to a possible depletion of some volatiles in JFCs (e.g., see Roth et al. 2018).This effect can be more visible on volatiles with the lowest vacuum sublimation temperatures (e.g., CO and CH 4 ).However, a study of fragment (daughter) species in a much larger sample of comets at optical wavelengths shows no correlation between depletion of carbon-chain species such as C 2 and C 3 and dynamical age, suggesting that the differences seen in JFCs and OCCs are primordial (A'Hearn et al. 1995;Dello Russo et al. 2016b).Therefore, the observed differences in parent species may also be primordial and may reflect different processing experienced by volatiles and dust prior to incorporation into cometary nuclei, rather than thermal processing after nucleus agglomeration.
Optical studies have revealed that most measured daughter species show little variation with heliocentric distance (A'Hearn et al. 1995;Fink 2009;Langland-Shula & Smith 2011;Cochran et al. 2012;Dello Russo et al. 2016b).One exception is an increase in C 2 /CN with decreasing heliocentric distance measured in some comets (A'Hearn et al. 1995;Opitom et al. 2015aOpitom et al. , 2015b see Table 11 in Dello Russo et al. 2016b).In many comets, there is a significant source of C 2 and/or CN from grains, and the importance of that source is variable within the comet population (see Table 11 in Dello Russo et al. 2016b).Therefore, optical measurements on a larger number of comets cannot be used to infer general information about their parent molecule's mixing ratios, and the significance of infrared and optical measurements must be assessed on a comet-by-comet basis (Dello Russo et al. 2016b).Even though this means that the mixing ratios of daughter species cannot inform the mixing ratios of parent volatiles in many cases (C 2 -and CN-bearing ices from more complex species such as C 4 H 2 , CH 2 C 2 H 2 , CH 3 C 2 H, and HC 3 N are normally at least an order of magnitude less abundant than C 2 H 2 and HCN in comets; see Fray et al. 2005;Hölscher 2015), optical and infrared comparisons can provide information on the sources of C 2 and CN and allow inferences on the prevalence and composition of organic grains.Therefore, coordinated multiwavelength observations of comets are necessary to better evaluate the associations between parent and daughter species.

Ammonium Salts
Ammonium salts (NH 4 + X − ) can form with the interaction of ammonia with molecules such as HCN, HNCO, and HCOOH at the low temperatures presented in the comet formation regions in the protoplanetary disk midplane.The main nitrogen-bearing species in comets, NH 3 , NH 2 , HCN (and its isomer HNC), CH 3 CN, and HC 3 N, have emission features in the near-IR and radio wavelengths (see Biver & Bockelee-Morvan 2015, Dello Russo et al. 2016b;Bockelée-Morvan & Biver 2017;Lippi et al. 2021;Biver et al. 2022), and the detection or stringent upper limits of these molecules have been reported in many comets to date (e.g., Dello Russo et al. 2016b;Biver et al. 2018Biver et al. , 2021;;Saki et al. 2021;Biver et al. 2022;Dello Russo et al. 2022;Khan et al. 2023).Ammonium salts are hard to detect in general, as they are unstable in the gas phase and their infrared signature is often hidden.While nitrogen-bearing molecules such as NH 3 are expected to increase in abundance as comets approach the inner parts of the solar system (e.g., C/2016 S1 (ISON); Dello Russo et al. 2016b;DiSanti et al. 2016) due to the high sublimation temperatures of the salts, which sequester these species into ammonium salts, it should be noted that the photochemical properties of NH 3 in comets are not yet fully understood.Therefore, a firm association with ammonium salts has not been demonstrated in comets other than 67P, and caution is needed in interpreting the nitrogen depletion in comets (see Biver et al. 2022 for a detailed discussion).

Polar versus Nonpolar Species and Pristine Ice-dust Formation
It is believed that ices can be separately stored in the nucleus based on molecular polarity, which may be revealed by their distributions in the coma.Polar and nonpolar ices have been observed in astronomical environments such as the interstellar medium (ISM), for example, through polar and nonpolar CO absorption features (see Figure 10 in Boogert et al. 2015).Ground-based studies of some comets, such as Hartley 2, C/2007 W1 (Boattini), C/2013 V5 (Oukaimeden), and C/ 2009 P1 (Garradd), suggested two phases of ices in their nuclei, one polar-bonded (e.g., associated with H 2 O and CH 3 OH) and another dominated by nonpolar bonds (e.g., CO 2 , CH 4 ; Villanueva et al. 2011;Paganini et al. 2012;Villanueva et al. 2012;DiSanti et al. 2014DiSanti et al. , 2018)).HCN, on the other hand, is a polar molecule, and yet it is often associated with the nonpolar hydrocarbons (Villanueva et al. 2011;Paganini et al. 2012;Villanueva et al. 2012;DiSanti et al. 2014DiSanti et al. , 2018)).During the Rosetta mission to comet 67P, early investigations found a consistent pattern between the production rates of O 2 (nonpolar) and H 2 O. Fluctuations in O 2 production closely mirrored those observed in H 2 O (Bieler et al. 2015;Luspay-Kuti et al. 2018;Combi et al. 2020), a strong indication that they are correlated, even though O 2 is more volatile (Fougere et al. 2016).Rubin et al. (2020) showed that the production rates of N 2 (which is also nonpolar and has a similar volatility as O 2 ) were not correlated with H 2 O and have a different profile and variation compared with H 2 O. Therefore, the polarity or nonpolarity of a species may not always appear to play a significant role in the observed correlations.
Formation of polar and nonpolar ices would need to occur in the earliest stages of protostellar disk formation, deep within cold, dark molecular clouds comprised of about 99% gas and 1% dust by mass (see Gibb 2001 for more details).The dust grains in these regions consist of a silicate core, possibly with a carbonaceous component, over which an ice mantle may be deposited when the temperature is low enough for gas phase molecules to condense (less than ∼150 K).The ice mantle may be composed of two layers (see Figure 1.1 in Gibb 2001); the innermost ice layer is dominated by polar molecules, primarily H 2 O ice, with smaller amounts of molecules such as CH 3 OH, CO, CO 2 , NH 3 , and CH 4 trapped in the water-ice matrix.At lower temperatures (<30 K), more volatile nonpolar molecules like CO 2 and potential hypervolatiles like N 2 , O 2 , and CO can freeze to the dust grains.In general, molecules will evaporate once the grains reach or exceed the sublimation temperature for that molecule.However, it is possible to trap nonpolar molecules like CO and CH 4 , which ordinarily sublimate below 30 K, in a water-ice (polar) matrix, where they can be retained at higher temperatures (see Whittet 2002 and references therein).Some molecules can be in both the polar and nonpolar mantles.CO, for example, has a dipole moment but is primarily in the nonpolar mantle mixed with CO 2 .At higher temperatures, polar CH 3 OH can also be in a mixed mantle with nonpolar CO 2 .Being in an ice mantle induces a small dipole moment in otherwise nonpolar molecules such as CO 2 (Gibb 2001;Whittet 2002).CH 4 is a nonpolar molecule but has been observed to be in the polar mantle before (probably because it forms via hydrogen addition reactions on the grain as water is forming via the same hydrogen addition reactions but to oxygen rather than carbon).For more information on interstellar dust and ice, see Gibb (2001), Whittet (2002), Boogert &Ehrenfreund (2004), andBoogert et al. (2015).
It is not clear whether protostellar envelope ices can survive when they accrete into the protoplanetary disk.However, depending on the distance to the central star, full or partial destruction of ices and subsequent reformation may have occurred (Mumma & Charnley 2011;Boogert et al. 2015).The Stardust mission and the detection of crystalline silicates revealed that refractory dust that was formed closer to the Sun where the temperature is too high for any ice to survive can move radially outward, accumulating ices along the way and thus diluting any presolar components in the outer comet formation zones (Brownlee 2014).Boogert et al. (2015) discuss the differences between the compositions of cometary and young stellar object envelope ices (see Section 6.5 of Boogert et al. 2015).C-and N-bearing species in comets are underabundant on average.For C, the abundance distributions overlap, which may imply that the early solar system resembled a low-mass young stellar object with less C in the ices (Öberg et al. 2011;Boogert et al. 2015).Low cometary CO and CO 2 may also reflect sublimation of the most volatile, nonpolar components upon accretion from the envelope to the protoplanetary disk (Visser et al. 2009a(Visser et al. , 2009b)).However, neither argument holds for N; the NH 3 abundance distributions do not overlap, and NH 3 is only present in the polar (least volatile) ices (Boogert et al. 2015).
Direct observations of ices in the protoplanetary disk midplane, where comets are formed, are not possible, since these regions are optically thick at IR and millimeter wavelengths.At longer wavelengths, such as the decimetric wavelengths sampled by the Next Generation Very Large Array (ngVLA), the dust is optically thin, enabling observations into the comet formation region.The ngVLA also has the requisite sensitivity and high angular resolution to map out the distributions of organics, such as nitrogen-bearing CH 3 CN, in the disk's midplane (Öberg et al. 2018).It is also possible that ices in the disk surface may be linked to the midplane by turbulence (Willacy et al. 2022).High spatial resolution observations have shown that these ices can experience thermal processing (see Section 5.4 of Boogert et al. 2015).Further observations are needed to search for evidence of energetic processing, but chemical models that include vertical disk turbulence have indicated that the ice composition throughout the disk may be significantly affected by energetic processes in the disk surface and radial transport (Ciesla & Sandford 2012).

Correlation of Sulfur-bearing Molecules
Identifying the molecules responsible for trapping atomic sulfur in molecular clouds and effectively hiding the element in protoplanetary disks remains a critical goal for planetary science and in general for astrophysics (see Kama et al. 2019).Less than 1% of the ISM sulfur abundance can be accurately measured in gas form, requiring clathration or chemical processing to become incorporated into refractories and ices.Measuring sulfur-bearing molecules and their correlations in cometary comae is an excellent way to provide measurements of the end state of this process.Atomic sulfur measurements are common for comets in the UV, although their interpretation is not straightforward due to the required radiative transfer modeling.The in situ measurements by the ROSINA instrument showed that 67P's measurements of atomic sulfur relative to water could not be easily traced to any sulfur-bearing species, therefore prompting more questions about the sources of elemental sulfur in comets (Calmonte et al. 2016;Biver et al. 2022;Noonan et al. 2023).Due to their strong transitions, multiple sulfur-bearing molecules are commonly studied in comets from remote observations.In the UV, S, S 2 , and CS have been detected in multiple comets so far (e.g., Meier & A'Hearn 1997;Reylé & Boice 2003;Noonan et al. 2023); in the near-IR, OCS has been detected in multiple comets with transitions at ∼4.8 μm (Saki et al. 2020); and at radio wavelengths, species such as H 2 S, SO, SO 2 , H 2 CS, OCS, and CS have strong transitions (Boissier et al. 2007; see Table 3 in Biver et al. 2022).These molecules have been measured in dozens of comets (e.g., see Boissier et al. 2007, Table 3 in Biver et al. 2022, Table 5 in Le Roy et al. 2015, Table 1 in Bockelée-Morvan et al. 2004, and Table A2 in Calmonte et al. 2016 for a list of detected sulfur-bearing species and their range of abundances in comets).
The correlation of sulfur-bearing molecules to both H 2 O and CO 2 at different epochs of the Rosetta mission makes it difficult to tie them to a specific region of the nucleus or icy matrix (see Figures 2 and 3).Läuter et al.'s (2020) power-law exponent for sulfur-bearing species showed better correlation with H 2 O at certain times (−290 to −180 days to perihelion) and with CO 2 at other times (100-160 and 190-380 days postperihelion) in 67P's orbit (e.g., H 2 S in epochs 3 and 5 and OCS in epoch 4; see Figure 2(A) versus 2(B)).One possible explanation for the slightly better correlation with H 2 O early in the mission (epochs 1-3; see Figures 3 and 4) and a shift toward CO 2 postperihelion could be depletion of CO 2 in the upper layers of the comet.As CO 2 has a sublimation temperature that renders it volatile out to large heliocentric distances, the upper layers of the nucleus that were depleted in CO 2 during the previous perihelion passage began to sublimate primarily H 2 O during the beginning of Rosettaʼs escort.As the thermal budget was largely devoted to sublimating H 2 O in the early phases of the mission, and sulfur-bearing molecules that had been present in the upper layers following the previous perihelion passage would have sublimated with the CO 2 during the outbound portion of the orbit, the correlations between the sulfur-bearing molecules and CO 2 in the early phases of the mission are weaker (see Figures 3 and 4).As shown in Figure 4, the sulfur-bearing molecules more closely follow H 2 O and CO 2 variations on the pre-and postperihelion dates, respectively.This is similar to what has been suggested by Luspay-Kuti et al. (2022) for the correlation of O 2 with H 2 O and CO 2 in the pre-and postperihelion dates, respectively.
Interstellar chemistry would suggest that OCS should be more closely correlated with CO and CO 2 (see Section 5.2 in Saki et al. 2020); however, a strong correlation with H 2 O could imply a different formation pathway in protoplanetary disks.Given that OCS and CS 2 both show a strong correlation with H 2 O production, oxidation of CS 2 to produce OCS in the protoplanetary disk may be likely.Ward et al. (2012) showed that thermal oxygen atoms are quite efficient at converting CS 2 ice to OCS at ISM conditions, in fact, so much so that it is difficult to maintain any CS 2 reservoir.However, this may not be the case at protoplanetary disk densities, where the lack of sulfur-bearing molecule detections leaves chemical models poorly constrained (Semenov et al. 2018;Noonan et al. 2023).
Another substantial sulfur atom donor in the formation process could be SO 2 , which is not evaluated in this study.This would likely rely on oxidation of sulfur atoms by H 2 O to form SO 2 ; however, as noted in Saki et al. (2020), this may be one possible way to limit OCS formation in H 2 O gas-rich areas of the protoplanetary disk, where comets are anticipated to have formed.The CS 2 /H 2 O ratio may be a strong indication of the formation location in the disk, a sign of the necessary temperature and UV influence to produce OCS while maintaining a CS 2 reservoir.The correlation of H 2 S and CO 2 appears to be more primordial than the OCS and CS 2 link to H 2 O, as the irradiation of H 2 S and CO and CO 2 ices is expected to be the main chemical source of CS 2 .This H 2 S correlation could be indicative of the remaining primordial connection following the irradiation and formation of oxidized sulfur molecules, or it may simply be the result of the similar volatility of the compounds at inner solar system temperature ranges.
The Space Telescope Imaging Spectrograph on the Hubble Space Telescope measured the sulfur-bearing species CS in comet 67P during the 2021 apparition (Noonan et al. 2023) and discovered significant variations from both Rosetta measurements (Calmonte et al. 2016) and radio observations (Coulson et al. 2020;Biver et al. 2021).Higher levels of CS/H 2 O (derived from CS and OH) were seen compared to in situ measurements, challenging the previous assumption that CS is a direct product of CS 2 dissociation, as it was also shown in radio millimeter and submillimeter observations of comets.This suggests there might be other sulfur-bearing molecules that are contributing to the production of CS, or, as indicated in Calmonte et al. (2016), since CS could not be resolved from CO 2 , there might be significant CS gas subliming from the nucleus (see Noonan et al. 2023 for more details).
The correlation factor in production rates (molecules s -1 ) of sulfur-bearing molecules with respect to both H 2 O and CO 2 follows an increase from epoch 1 to 2, decreasing from epoch 2 to 3 for OCS and H 2 S with respect to CO 2 and H 2 O, respectively, while it stays relatively the same or increases for the rest (Figures 2(A) and (B)).OCS versus H 2 O remains relatively constant except in epochs 1 and 6, where it is lower.Epochs 4-6 illustrate a relatively constant or general decrease in the correlation factor of production rates for all sulfurbearing molecules (see Figures 2(A) and (B)).However, as shown in Figure 3, the correlation factor in the mixing ratios of sulfur-bearing molecules with CO 2 (e.g., OCS/H 2 O versus CO 2 /H 2 O in percentage) show a constant increase from epoch 1 to 6 except for H 2 S, which shows a higher correlation factor in epoch 3.This contrasts the variable correlation factor of sulfur-bearing mixing ratios with that of H 2 O (e.g., CS 2 /CO 2 versus H 2 O/CO 2 in percentage) seen from epoch 1 to 6.This increasing correlation is indicative of a more primordial material sublimating from layers of the comet as the thermal wave reaches the layers richer in sulfur-bearing species.

Summary
We conducted a comprehensive correlation analysis of parent volatiles in comet 67P during six different epochs of the Rosetta mission and compared them with ground-based observations of comets.It has been suggested that there are at least two different ice matrices, H 2 O and CO 2 ice, in which the minor species are embedded in different relative abundances (e.g., Luspay-Kuti et al. 2015;Gasc et al. 2017;Hu et al. 2017;Luspay-Kuti et al. 2019;Läuter et al. 2020;Luspay-Kuti et al. 2022).Thus, it is reasonable to compare production rates, relative abundances, and spatial distributions of minor volatiles with respect to both CO 2 and H 2 O.We employed various methodologies to establish associations among volatiles (listed in Section 2)).We note that different techniques of grouping molecules with respect to H 2 O and CO 2 may yield different results, as observed in certain cases (see Table 3).In that context, our investigation yielded the following.
1. Seasonal variations in 67P.67P exhibited strong seasonal variations due to its high axial tilt (∼52°).The time during which volatile composition measurements were obtained along its orbit affected the observed correlations among species, suggesting that these relationships may vary in different active areas of 67P.1). 4. Volatiles complex relationships.Some species, such as sulfur-bearing molecules (e.g., H 2 S), followed H 2 O closely at some heliocentric distances and CO 2 at others (see Table 1 and Figure 2).This, combined with seasonal effects suggest that there is probably a more complex relationship between these species measured in 67P (e.g., see Biver et al. 2019;Läuter et al. 2020; and Figure 2). 5. Abundance associations with R h dependence in 67P.
Power  et al. (2022), there are treatments that can be employed to estimate CO 2 production rates from the ground (e.g., CO Cameron bands), these methods present challenges and are often difficult to carry out.8. Spatial distributions of species within the comet population.
Spatially resolved measurements in radio and near-IR provided insights into volatile distributions in the coma.NH 3 , H 2 CO, and CH 3 OH displayed similar profiles to H 2 O in multiple comets, while C 2 H 6 , HCN, CH 4 , CO, CS, and C 2 H 2 exhibited differences.Consequently, these species likely exhibited profiles similar to CO 2 (with the caveat that CO 2 cannot be measured from the ground).9. Polarity.Polar (e.g., H 2 O and CH 3 OH) and nonpolar (CO 2 and CH 4 ) species, which are observed in cometary coma, showed differences in spatial profile when compared with one another.Therefore, it has been suggested that they are stored in separate ice matrixes.However, certain volatiles, such as HCN (polar), are often associated with the nonpolar hydrocarbons, challenging this classification based on polarity.10.Snapshot measurements.Ground-based facilities may suggest that volatiles such as CH 3 OH, H 2 CO, NH 3 , and C 2 H 2 better followed variations in H 2 O production rates, whereas CO and CH 4 better followed C 2 H 6 (and, by proxy, CO 2 ).We note that this behavior is not universal and could be different from comet to comet but might be attributed to the hypervolatility of these species.
Overall, our findings suggest that the production rates of different molecules in comet 67P are not only influenced by the H 2 O and CO 2 ice matrix but also by other factors, such as seasonal variations.The use of R 2 , slope (S), and Pearson correlation (r) allowed us to rigorously analyze and quantify the relationships between these variables, provided valuable information for understanding the complex processes at work in comet 67P, and made a detailed comparison with groundbased measurements.

Figure 1
Figure1shows the weak correlation between CO 2 and H 2 O during the entire Rosetta mission.This lack of correlation suggests that the release of CO 2 and H 2 O are in general independent of one another.We fit a line to the production rate of H 2 O versus CO 2 for the highest, lowest, and average production rate values presented inLäuter et al. (2020) and perform a different statistical measure (see Figure1).Separate source regions of H 2 O and CO 2 have been observed in other comets as well (e.g., 9P/Tempel 1 with the Deep Impact mission;Feaga et al. 2007;Farnham et al. 2013; and Hartley 2 with the EPOXI mission).In comet Hartley 2, EPOXI/DIXI revealed that H 2 O had multiple sources on the nucleus, while strong CO 2 emission from the smaller lobe dragged icy grains along into the coma, from which water sublimed, resulting in an extended H 2 O source and the hyperactivity of Hartley 2 (A'Hearn et al. 2011;Protopapa et al. 2014).
) and (B) show our derived correlation coefficients (R 2 ) of different species with respect to H 2 O and CO 2 (data fromLäuter et al. 2020).The differences in correlation coefficients are due to varying activity levels, seasonal effects, and sublimation processes at different heliocentric distances.A similar power-law relationship was found by Läuter et al. (2020) and by Biver et al. (2019) using MIRO mapping data.
Figures A1-A12 in the Appendix show the change of production rates of parent volatiles (averaged over a range of heliocentric distances; see available data in Läuter et al. 2020) with those of H 2 O and CO 2 at the same epoch (see the Appendix).These figures are interactive 3D; therefore, we refer readers to the online version of this manuscript rather than a static PDF.As suggested by Läuter et al. (2020), Biver et al. (2019), and Figures 2(

Figure 1 .
Figure 1.The overall low correlation between CO 2 and H 2 O (R 2 = 0.48).Data are the production rates based on ROSINA measurements from 2014 August to 2016 September taken from Läuter et al. (2020).The color bar shows the log of C 2 H 6 production rates (another volatile that is systematically observed in comets and has been used as a base for volatile mixing ratios) during the entire mission.Each circle represents the average values within each measurement, while the error bars show the highest and lowest production rates within the same measurement.

Figure 2 .
Figure2.Variations seen in the correlation factor (R 2 ) of the average production rates of each parent species as measured at different epochs (see Table2). (A) With respect to H 2 O. (B) With respect to CO 2 .Data are taken fromLäuter et al. (2020), who present average, highest, and lowest production rate values taken at different R h ranges.Regions were designated as strong (R 2 > 0.9), good (0.9 > R 2 > 0.8), moderate (0.8 > R 2 > 0.5), and weak (R 2 > 0.5) correlation.

Figure 4 .
Figure 4. ROSINA measurement of production rates of H O, CO 2 , and sulfur-bearing molecules plotted against days around perihelion.Note how the sulfur-bearing molecules better follow H 2 O in preperihelion and CO 2 in postperihelion.The error bars indicate the highest and lowest values, while the points indicate the average values within those measurements (see Läuter et al. 2020 for details).

Figure A4 .
Figure A4.The production rates of CH 4 .

Figure A6 .
Figure A6.The production rates of CS 2 .

Figure A8 .
Figure A8.The production rates of H 2 S.

Figure A10 .
Figure A10.The production rates of NH 3 .

Figure A12 .
Figure A12.The production rates of OCS.
for more details).Coma imaging using scanning spectrometers such as VIRTIS on board the Rosetta spacecraft indicates that H 2 O and CO 2 have different distributions and are released from different spots on the nucleus

Table 2
Different Statistical Measures of Each Parent Volatile in Comet 67P with Respect to H 2 O and CO 2 at Different Epochs Läuter et al. (2020)spect to H 2 O. (B) With respect to CO 2 .Data are taken fromLäuter et al. (2020), who present average, highest, and lowest production rate values taken at different R h ranges.Regions were designated as strong (R 2 > 0.9), good (0.9 > R 2 > 0.8), moderate (0.8 > R 2 > 0.5), and weak (R 2 > 0.5) correlation.67Pat3.34au preperihelion (see Figures5 and 6and Table 2 in Luspay-Kuti et al. 2015).Later, Läuter et al. (2020) included CH 3 OH in the H 2 O-correlated group of gases based on the heliocentric power law in the outbound portion of 67P's orbit.Our analysis indicates that CH 3 OH is better correlated with H 2 O than CO 2 in comet 67P. 5. NH 3 .Gasc et al. (2017) showed that NH 3 correlates with H 2 O at certain times (2.4-2.63 and 3.20-3.50au) and with CO 2 in others.Läuter et al. (2020) also placed NH 3 in the H 2 O-correlated group of gases based on their study of the power-law components of the relationship between production rates and heliocentric distance.Our analysis showed similar results to that of Läuter et al. (2020); however, a strong correlation exists between NH 3 and CO 2 in certain epochs when 67P experienced its southern summer (see Table 2).6. HCN.Gasc et al. (2017) observed a flat distribution (correlation) with respect to H 2 O for HCN and showed that HCN correlates well with CO 2 near and post-second equinox.However, Luspay-Kuti et al. (2015) showed that on the inbound portion of the orbit (∼3.3-3.2 au), HCN better correlates with H 2 O than with CO 2 , suggesting the possible heterogeneity of 67P's nucleus.Läuter et al. (2020) grouped HCN with CO 2 based on the power-law component.Our analysis also confirmed that HCN is better correlated with CO 2 than H 2 O. 7. CH 4 .Luspay-Kuti et al. (2015) found a weak correlation between CH 4 and CO 2 before the inbound equinox, whereas a stronger correlation existed between CH 4 and H 2 O.In contrast, CH 4 showed a better correlation with CO 2 in the outbound portion of the orbit (near and post-second equinox; Gasc et al. 2017; Luspay-Kuti et al. 2019).Läuter et al. 2 O (Luspay-Kuti et al. Crovisier et al. 2009)d Figure2inCrovisier et al. 2009).Measurements of volatile spatial distributions in the coma are mostly limited by comet productivity and the strengths of molecular emissions.The easiest-to-detect species with the strongest emissions can be detected in most observable comets (e.g.,H 2 O and C 2 H6 in the near-IR, HCN and CH 3 OH in radio, and CN in optical/UV).However, studies including a larger suite of molecules, such as C 2 H 2 , OCS, CS, CO, and NH 3 , are limited to the brightest and most productive comets.Therefore, detailed spatial studies have been done only for a limited number of comets, for instance, C 2 H 2 in Hartley 2 (Dello Russo et al. 2011), NH 3 and C 2 H 2 in C/2012 S1 (ISON) (Dello Russo et al. 2016a), C 2 H 2 in 46P/Wirtanen (Bonev et al. 2021), OCS in C/2002 T7 (LINEAR) (Saki et al. 2020) and C/1995 O1 (Hale-Bopp) (Dello Russo et al. 1998), CS in C/2015 ER61 (PanSTARRS) (Roth et al. 2021), CO in C/2016 R2 (PanSTARRS) (Cordiner et al. 2022), and NH 3 and C 2 H 2 in C/2014 Q2 (Lovejoy) (Dello Russo et al. 2022).
Russo et al. 2011; 3 OH; see Table1) Bonev et al. 2013;Kawakita et al. 2013tended in the sunward direction, which was different from that of H 2 O (see DelloRusso et al. 2011; Mumma et al. 2011;Bonev et al. 2013;Kawakita et al. 2013).In the case of comet Hartley 2, H 2 O is the common measured molecule between EPOXI and ground-based observations (Dello 6 in comet C/2007 W1 (Boattini) exhibited antisymmetry, resembling the profile of HCN but different from the symmetric profiles of H 2 O and CH 3 OH (see Figure 5 in Villanueva et al. 2011).Similarly, in comet Hartley 2, the spatial profiles of C 2 H 6 and HCN (with similar vacuum sublimation concluded that HCN/H 2 O has a moderately high degree of Pearson correlation to C 2 H 6 /H 2 O in ground-based infrared measurements.This can indicate that HCN and C 2 H 6 ices are connected in the nucleus, so the HCN/C 2 H 6 relative abundance is relatively constant among comets.More importantly, ground-based measurements have limited data for certain molecules such as C 2 H 2 , CO, and CH 4 , indicating a need for more comprehensive observations.
2. Associations in long-term monitoring in 67P.Long-term measurements showed better associations of CH 3 OH, H 2 CO, NH 3 , and O 2 with H 2 O, while C 2 H 5 OH, C 2 H 6 , CH 4 , CO, CS 2 , H 2 S, HCN, and OCS exhibited stronger associations with CO 2 .3. Short-term variations in 67P.Short-term monitoring of volatiles indicated that species such as CH 3 OH, O 2 , NH 3 , and H 2 S better followed H 2 O, while CO, H 2 S, CH 4 , HCN, and C 2 H 6 exhibited better correlations with CO 2 (see Table -law component analysis revealed that CH 3 OH, O 2 , H 2 CO, and NH 3 were better associated with the H 2 O ice matrix, whereas CO, H 2 S, C 2 H 6 , CH 4 , HCN, C 2 H 5 OH, OCS, and CS 2 were better associated with CO 2 .6.Chemical abundances within the comet population.Groundbased radio and infrared observations indicated that species such as CH 3 OH, CH 4 , HCN, and C 2 H 2 were better associated with H 2 O, while species such as CS, CO, OCS, NH 3 , and H 2 CO likely correlated better with CO 2 .7. The blocking of CO 2 by Earth's atmosphere.Earth's atmosphere effectively blocks CO 2 emissions.The observed differences between H 2 O and C 2 H 6 in comets from ground-based measurements and between H 2 O and CO 2 from space missions to comets (Rosetta and EPOXI) have led to the use of C 2 H 6 as a proxy for CO 2 .Although, as discussed in Section 3.3.1 and in Harrington-Pinto