Observations

The observations were made at the Instituto de Radioastronomie Milimetrica (IRAM) 30 m telescope at Pico Veleta, Spain at 1–8 h UT on 10 Jan 2008 using the HERA camera and the VESPA spectrometer. The CH₃OH 5(1,1)-4(1,1) line at 239.7463 GHz was observed with 80 MHz passbands with a telescope beam size of ~10.5′′ full width at half maximum ( FWHM). The observations were made by frequency-switching over a narrow 3.45 MHz interval to maintain flat baselines using 80 kHz channels in VESPA. Figure 1 shows the location of Enceladus w.r.t. Saturn at the time of observations. The angular size of Enceladus was 0.08′′ with Saturn 8.6 AU from the Earth. The ring opening angle was −6.9°. We note that the only possible contaminant within the beam was the moon Mimas towards the end of the observing period.

Fig. 1. Location of Enceladus and Saturn during the CH₃OH observations on 10 January 2008 (1–8 h UT). The C- (innermost ring; light blue), B- (centre ring; green) and A-rings (outer ring; blue) are shown, along with the centre of the E-ring (dotted line), which also corresponds to Enceladus' orbital radius.

The telescope could track Enceladus, but the acquisition software was only able to track the velocity of Saturn. Saturn was offset by ≥30′′ during the observations to prevent spectral contamination of the Enceladus data. Velocity shifts to place the data in the Enceladus rest-frame were then made in reduction software for each observation. In total, 48 observations were made, with the drift between observations being <0.1 km s⁻¹ on average (worst case of 0.4 km s⁻¹). The applied shift varied from −9.2 km s⁻¹ at the start to −12.6 km s⁻¹ at the maximum elongation followed by reversal back to −10.6 km s⁻¹ at the end of the track. Hence, any artefacts associated with particular spectrometer channels would be smeared over a few km s⁻¹ in the co-added data, and so not be able to create a false-positive narrow line.

Details of the data reduction (i.e. despiking, baseline-fitting and baseline-subtraction) are given in the Appendix (see Fig. A1). The spectra are automatically calibrated to a T*_A antenna temperature scale. However, we must correct the antenna temperature scale for telescope inefficiency. The efficiencies were not measured during our run, but the HERA User Manual (V2.0; Nov 2009) stated the beam efficiency is 0.52 at 230 GHz and the forward efficiency is 0.90. We calculate a main-beam temperature T _MB = F _eff/B _eff T*_A, i.e. T*_A is divided by 0.58 to give T _MB.

In addition to the CH₃OH spectrum, we also include publicly available observations taken by Herschel (originally from project ID OT2_elellouc_3, PI Lellouch) of the ortho-H₂O 2₁₂–1₀₁ line at 1669.9 GHz, taken on 27 June 2012 at ~22:00 UT over a period of ~0.33 h using the HIFI instrument. The observations were taken in the fast DBS (dual beam switch) raster mapping mode, where we extracted the spectrum using only the observation centred directly on Enceladus. The telescope beam FWHM is ~13′′ at ~1660 GHz. During this observation, Enceladus was at similar elongation from Saturn as the observations from IRAM, where the rings were edge-on.

Models

Even though methanol has been found within the plume components (Waite et al., Reference Waite, Lewis, Magee, Lunine, McKinnon, Glein, Mousis, Young, Brockwell, Westlake, Nguyen, Teolis, Niemann, McNutt, Perry and Ip2009), the region in which IRAM can detect methanol must be on a larger scale. The diameter of the telescope beam subtends ~73 000 km (~1.2 R_S), which is much larger than Enceladus' diameter of only ~500 km. Even if the methanol cloud is optically thick and warm (e.g. plume temperatures are ~180 K; Johnson et al. Reference Johnson, Smith, Tucker, Liu, Burger, Sittler and Tokar2006; Cassidy & Johnson Reference Cassidy and Johnson2010), the minimum size we are sensitive to is ~1750 km, or 7 R_Enc. On the other hand, diffuse gas in the larger neutral torus could fill the beam (where the torus has a radial extent of R_Enc ± 1.25 R_S and a vertical extent of ±0.4 R_S; Farmer Reference Farmer2009; Hartogh et al. Reference Hartogh, Lellouch, Moreno, Bockelée-Morvan, Biver, Cassidy, Rengel, Jarchow, Cavalié, Crovisier, Helmich and Kidger2011).

For generality, we consider a cloud of methanol with length l, and use the non-local thermodynamic equilibrium (non-LTE) radiative transfer code RADEX to find solutions for the radiation temperature (T _R). Details of RADEX and the full parameter space of the models that fit the CH₃OH observation can be found in the Section‘Full RADEX parameter space’. For clouds with size scales smaller than the telescope beam (i.e. l < 73 000 km), we must account for beam dilution, which causes the main-beam temperature to be lower than the expected radiation temperature by T _MB = η _dil T _R, where η _dil = (l²/l² + l²_beam) and l_beam is the beam diameter.

For the radiative transfer models, we set the gas kinetic temperature T _gas = 100 K (e.g. Hartogh et al. Reference Hartogh, Lellouch, Moreno, Bockelée-Morvan, Biver, Cassidy, Rengel, Jarchow, Cavalié, Crovisier, Helmich and Kidger2011) and the line FWHM at 0.43 km s⁻¹ as observed (section‘Results’). While plume temperatures reach ~180 K (Johnson et al., Reference Johnson, Smith, Tucker, Liu, Burger, Sittler and Tokar2006; Cassidy & Johnson, Reference Cassidy and Johnson2010), we expect plumes to cool to Saturn's thermal field (100 K) at the scales traced by our observations (see Hartogh et al. Reference Hartogh, Lellouch, Moreno, Bockelée-Morvan, Biver, Cassidy, Rengel, Jarchow, Cavalié, Crovisier, Helmich and Kidger2011 for modelling of the gas temperature within the neutral torus). From the gas kinetic temperature, we can calculate the non-thermal (turbulent) velocity dispersion of methanol by σ ²_NT  = σ ²_CH3OH − σ ²_T, where σ _NT is the non-thermal line width, σ _CH3OH is the observed line width of CH₃OH (calculated from $ {\rm \sigma}_{{\rm CH}_3{\rm OH}}= \hbox{v}_{\rm FWHM} / \sqrt {8\ln (2))}$ and σ _T is the thermal component of the line width (assuming a 100 K cloud). Removing the thermal component leaves a turbulent velocity dispersion of ~0.1 km s⁻¹, suggesting the methanol gas is dispersing at  ≤ 0.1 km s⁻¹.

Solutions for T _MB were sought for different CH₃OH:H₂O abundance ratios (X), where we relate the H₂O density, n(H₂O), to the CH₃OH column density, N(CH₃OH) by n(H₂O) = (N(CH₃OH)/lX). The CH₃OH:H₂O abundance is constrained as X ≤ 5%, which is the expected C:H₂O budget (estimated from the CO₂:H₂O abundance within Enceladus' plumes; Waite et al. Reference Waite, Lewis, Magee, Lunine, McKinnon, Glein, Mousis, Young, Brockwell, Westlake, Nguyen, Teolis, Niemann, McNutt, Perry and Ip2009). If methane CH₄ is the parent species of CH₃OH, then the preferred CH₃OH:H₂O abundance is X ≤ 1%. In RADEX, the collision partner is H₂, so we set density to be 10 ×  higher to account for the higher mass of the actual collision partner H₂O.

Additionally, we find solutions for the cloud length l that reproduce T_MB(CH₃OH) within the allowed CH₃OH:H₂O abundance. Furthermore, a maximum H₂O column density was applied. The maximum assumes that all of the H₂O molecules vented (estimated to be 10²⁸ molecules s⁻¹; Smith et al. Reference Smith, Johnson, Perry, Mitchell, McNutt and Young2010) will survive during the photodissociation timescale ~2.5 months (e.g. Cassidy & Johnson Reference Cassidy and Johnson2010; Smith et al. Reference Smith, Johnson, Perry, Mitchell, McNutt and Young2010; Hartogh et al. Reference Hartogh, Lellouch, Moreno, Bockelée-Morvan, Biver, Cassidy, Rengel, Jarchow, Cavalié, Crovisier, Helmich and Kidger2011) within the specific region of size l.

Figure 3 depicts the models that match the expected CH₃OH radiation temperature (i.e. the main-beam temperature corrected for beam dilution) within a conservative 50% uncertainty for each cloud length l. We investigate the CH₃OH:H₂O abundance by comparing the methanol and water column densities, N(CH₃OH) and N(H₂O) for each cloud length. We find allowed CH₃OH cloud dimensions l to be 1750–72500 km. At smaller cloud lengths (l < 1750 km), methanol becomes increasingly beam diluted and high densities and column densities are needed to produce the methanol detection. However, CH₃OH also becomes optically thick and higher gas temperatures are needed to adequately reproduce the beam-diluted detection. We find the modelled CH₃OH-to-H₂O abundance is relatively independent of cloud length, ranging from 0.5 to 5%. This abundance is ~50 ×  higher than what is found in the direct vicinity of the plumes (Waite et al., Reference Waite, Lewis, Magee, Lunine, McKinnon, Glein, Mousis, Young, Brockwell, Westlake, Nguyen, Teolis, Niemann, McNutt, Perry and Ip2009), implying chemical processing has occurred.

Fig. 3. Left: RADEX model results for CH₃OH and H₂O column densities and varying cloud lengths that match the CH₃OH observation. Colours correspond to cloud lengths 1750–75 000 km (in increments of 250, 500 and 2500 km at ranges l < 2500 km, 2500 < l < 5000 km and l > 5000 km, respectively). Dashed lines (left to right) correspond to CH₃OH-to-H₂O abundances of 5, 1 and 0.5%. Right: Plasma velocities near Enceladus, based on Fig. 2 in Jia et al. (Reference Jia, Russell, Khurana, Ma, Najib and Gombosi2010). The component X (R=R _Enc) is along the direction of the corotational flow around Saturn and component Y is positive towards Saturn. Velocities are measured w.r.t. Enceladus' orbital velocity. Contours correspond to the highest plasma densities, n~1–4 × 10² cm⁻³.

The cloud size scenarios are summarized in the appendix (Table B1), including the range of H₂O densities from the RADEX models that fit the CH₃OH observation. For smaller cloud lengths (e.g. l = 1750 km), we have H₂O densities ranging from n_RADEX(H₂O) = 1.5 × 10⁹–1.1 × 10¹⁰ cm⁻³. This density range is reasonably in agreement with past Cassini measurements of the E3 and E5 flybys, which found the peak H₂O density to reach ~10⁸–10⁹ cm⁻³ near the plumes (Smith et al., Reference Smith, Johnson, Perry, Mitchell, McNutt and Young2010; Teolis et al. Reference Teolis, Perry, Magee, Westlake and Waite2010). At larger cloud lengths (e.g. l = 72 500 km), we find the H₂O density range to be n_RADEX(H₂O) = 4.5 × 10⁴–1.6 × 10⁵ cm⁻³, slightly above Cassidy & Johnson (Reference Cassidy and Johnson2010), which found neutral densities in the torus ~10⁴ cm⁻³.

We can estimate how far the CH₃OH is able to travel outward from Enceladus from the photodissociation timescales under solar irradiation. If the gas is dispersing at  ≤ 0.1 km s⁻¹ and the photodissociation timescale is t_photo = 1.7 × 10⁶ s (accounting for Saturn's distance at 9.3 AU; Huebner et al. Reference Huebner, Keady and Lyon1992), then we expect a methanol cloud extent up to  ≤ 173 000 km (i.e. ~3 R_S). This means the methanol is easily able to spread into the neutral torus and fill the beam of the telescope. We note that H₂O and CH₄ have even longer photodissociation timescales than CH₃OH and are not a limiting factor.

Therefore, there are two possible explanations for the narrow CH₃OH line observation: methanol is made both by Enceladus and through other chemical pathways within the neutral torus that: (1) spreads outward or (2) remains in a more compact, confined gas cloud that trails Enceladus' orbit by several km s⁻¹.

The velocities at which methanol is seen are consistent with a torus model, in particular with molecules spreading outwards from Enceladus' orbit (Farmer, Reference Farmer2009). At the time of observations, Saturn's rings were tilted such that a beam pointed at Enceladus' position passed preferentially through the far-side of the neutral torus and even broader E-ring (3-9 R_S). Molecules further out than Enceladus will orbit with slower speeds. If we are observing the far-side of the torus or E-ring, molecular line observations will be redshifted w.r.t. Enceladus as it approaches. This is seen in the observations in Fig. 2. The magnitude of the line-of-sight velocity shift w.r.t. Enceladus can be fitted assuming the molecules are in Keplerian rotation in the ring plane, which was tilted w.r.t. Earth at 7° in 2008. The detected methanol must be at an orbit of ~8 R_S to be in the telescope beam, assuming most of the spectral emission was contributed around the time of maximum elongation. However, methanol ejected directly from Enceladus through plumes may only be able to spread by ~3 R_S accounting for a methanol photodissociation time at the distance of Saturn (i.e. out to a radius ~7 R_S accounting for the distance between Saturn and Enceladus; Huebner et al. Reference Huebner, Keady and Lyon1992). Therefore, the detected methanol would have to be a secondary product of chemical processing within the larger neutral torus in this scenario, as discussed below.

The second possibility is that a compact, confined gas cloud trails Enceladus' orbit by several km s⁻¹. An example of a trailing gas cloud can be seen in a model by Jia et al. (Reference Jia, Russell, Khurana, Ma, Najib and Gombosi2010), which was reconstructed from Cassini flyby measurements of the plasma. In Fig. 3, we show a diagram of Enceladus with the plasma velocities and density contours (based on Fig. 2 in Jia et al. Reference Jia, Russell, Khurana, Ma, Najib and Gombosi2010). The plume emerges from the southern pole of Enceladus (to -Z) and streams along the X-axis, which is in the line-of-sight for our observations. Electrons south of Enceladus' plumes form a denser cloud at length l≲1750 − 2000 km (7–8 R_Enc) and trail Enceladus by ≲6 km s⁻¹. We have observed CH₃OH redshifted by a similar +8.1 km s⁻¹ and over a similar-sized region. However, further modelling and observations are needed to explain the mechanism in which gas would be confined by Saturn's magnetosphere.

Additionally, there is independent evidence of a significant (4σ; see section‘Results’) redshifted molecular component from archival ortho-H₂O spectrum taken when Enceladus was at a similar elongation from the Saturn. Since the active plumes are thought to feed the neutral torus in orbit around the Saturn, it is expected that H₂O will fill the Herschel telescope beam and be present at both small and large scales surrounding Enceladus. In particular, this redshifted H₂O component can result from gas spreading (Farmer, Reference Farmer2009) or the presence of a confined cloud of gas lagging the orbital velocity of both the moon and the co-rotating neutral torus, where the latter process depends on the exact velocity of outgassing material at the time. If either gas spreading or a confined cloud are the causes of the redshifted molecular emission, we do not necessarily expect the H₂O and CH₃OH observations to be redshifted by the same amount because H₂O may be tracing different radii than the CH₃OH detection and/or temporal variations observed from the plumes (Hedman et al., Reference Hedman, Gosmeyer, Nicholson, Sotin, Brown, Clark, Baines, Buratti and Showalter2013). Careful monitoring of Enceladus needs to be done in the future to better understand how the plume processes change over time.

Considering that water, hydrogen, oxygen and hydroxyl are abundant in the torus, the rate of methanol production may depend on the availability of methane, which is a ~1.6% CH₄:H₂O abundance (Waite et al., Reference Waite, Lewis, Magee, Lunine, McKinnon, Glein, Mousis, Young, Brockwell, Westlake, Nguyen, Teolis, Niemann, McNutt, Perry and Ip2009), or other hydrocarbons. Using public astronomical networks (KIDAFootnote ¹, UMIST; McElroy et al. Reference McElroy, Walsh, Markwick, Cordiner, Smith and Millar2013), we suggest possible routes to methanol through gas-phase associate detachment:

(1)$$ \hbox{CH}_3 + \hbox{OH}^{-} \rightarrow \hbox{CH}_3\hbox{OH} + \hbox{e}^{-}. $$

While other routes for the creation of methanol are possible, these involve more complex molecules that are less abundant and may not be present in the plume environments.

Coates et al. (Reference Coates, Jones, Lewis, Wellbrock, Young, Crary, Johnson, Cassidy and Hill2010) find that dissociative electron attachment can produce negative water-group ions (e.g. OH⁻, O⁻, H⁻) in Enceladus' plumes from neutral H₂O, where O⁻ and H⁻ are produced by polar photodissociation of H₂O by photons in ~36 − 100 eV. While H⁻ has a relatively short photodissociation timescale (~6 s at Saturn; see Huebner et al. Reference Huebner, Keady and Lyon1992), Cordiner & Charnley (Reference Cordiner and Charnley2014) find that H⁻ quickly reactions with H₂O to produce OH⁻ + H₂ in the comet 1P/Halley environment, where both OH⁻ and O⁻ are abundant within the coma ( ≤ 1000 km from the nucleus). This cometary environment may be similar to Saturn, where Saturn is a soft X-ray source (scattering X-rays from the Sun; Bhardwaj et al. Reference Bhardwaj, Elsner, Randall Gladstone, Cravens, Lisse, Dennerl, Branduardi-Raymont, Wargelin, Hunter Waite, Robertson, Østgaard, Beiersdorfer, Snowden and Karchenko2007) that could form these water-group ions.

While the methyl radical is not directly modelled from the Cassini INMS data (Waite et al., Reference Waite, Combi, Ip, Cravens, McNutt, Kasprzak, Yelle, Luhmann, Niemann, Gell, Magee, Fletcher, Lunine and Tseng2006, Reference Waite, Lewis, Magee, Lunine, McKinnon, Glein, Mousis, Young, Brockwell, Westlake, Nguyen, Teolis, Niemann, McNutt, Perry and Ip2009), it is a key molecule in the production of more complex hydrocarbons (e.g. Titan; Atreya et al. Reference Atreya, Adams, Niemann, Demick-Montelara, Owen, Fulchignoni, Ferri and Wilson2006) and is likely present within Enceladus' plumes or created once material is ejected (Coates et al., Reference Coates, Jones, Lewis, Wellbrock, Young, Crary, Johnson, Cassidy and Hill2010). One possible pathway is through associative attachment from ethylene (CH₂) and water-group ions, where traces of ethylene have been noted in the plume environment (Waite & Magee, Reference Waite and Magee2010):

(2)$$ \hbox{CH}_2 + \hbox{H}^- \rightarrow \hbox{CH}_3 + \hbox{e}^-, $$

Similarly, reactions directly with methane and other molecules abundant in the plume environment can produce the methyl radical. For example, ion-neutral reactions:

(3)$$ \hbox{CH}_4 + \hbox{O}^- \rightarrow \hbox{CH}_3 + \hbox{OH}^-, $$

(4)$$ \hbox{CH}_4 + \hbox{H}_2{\rm O}^+ \rightarrow \hbox{CH}_3 + \hbox{H}_3{\rm O}^+, $$

where H₂O⁺ and H₃O⁺ have been found directly (Tokar et al. Reference Tokar, Johnson, Thomsen, Wilson, Young, Crary, Coates, Jones and Paty2009; altitudes ~200 km). Neutral-neutral reactions with methane are also possible with neutral water-group atoms and molecules present in the E-ring environment:

(5)$$ \hbox{CH}_4 + \hbox{OH} \rightarrow \hbox{H}_2{\rm O} + \hbox{CH}_3, $$

(6)$$ \hbox{CH}_4 + \hbox{O} \rightarrow \hbox{OH} + \hbox{CH}_3, $$

In all of these proposed scenarios, the plasma density appears to be an integral part of the gas-phase chemical production of CH₃OH and enhancement in CH₃OH abundance once the plume material is ejected from Enceladus. This further supports the offsets in velocity we see in the CH₃OH spectrum and the plasma within 2000 km from the south pole of Enceladus (Jia et al., Reference Jia, Russell, Khurana, Ma, Najib and Gombosi2010), where this material is travelling more slowly than the moon. Plasma densities are found to peak at ~10² cm⁻³, indicating an electron abundance ≥10⁻⁷ w.r.t. modelled H₂O (see Table B1), which is similar to electron abundances found in the ISM (w.r.t. H₂; Bergin & Tafalla Reference Bergin and Tafalla2007). As stated above, further observations are needed to constrain possible variability in the plume and surrounding environment, which may lead to changes in density, chemical composition and velocity of the ejected material over time.

We note that Hodyss et al. (Reference Hodyss, Parkinson, Johnson, Stern, Goguen, Yung and Kanik2009) find a potential CH₃OH ice surface feature, though this can also be interpreted as hydrogen peroxide H₂O₂ (Newman et al., Reference Newman, Buratti, Jaumann, Bauer and Momary2007). However, the past work has also shown that CH₃OH tends to be destroyed in ice pathways, making it less likely that CH₃OH is produced from ice on Enceladus' surface (Bergantini et al., Reference Bergantini, Pilling, Nair, Mason and Fraser2014).

Discussion and conclusions

We report on the first ground-based detection of a molecule (CH₃OH) in Enceladus' plumes. Radiative transfer models suggest the origin of methanol can be from a cloud with length ranging from 1750 km up to a size comparable with the neutral torus in the orbit around Saturn. There are two possible explanations for producing the redshifted CH₃OH and H₂O, taking into account the solar photodissociation timescale of CH₃OH and the observed redshifted CH₃OH emission. First, methanol may be a secondary product of chemical processing within the neutral torus and spread further out into the E-ring to a radius ~8 R_S (e.g. Farmer Reference Farmer2009) at velocities smaller than Enceladus' rest-frame velocity (i.e. due to Keplerian rotation). Alternatively, methanol may original from a smaller, beam-diluted region nearer to Enceladus (i.e. l < 2000 km). To further distinguish between these two mechanisms, further modelling and observations are needed.

What is the origin of the methanol cloud surrounding Enceladus? Since methanol is expected to be present in an organic environment (Matson et al., Reference Matson, Castillo, Lunine and Johnson2007), an exciting prospect is that the observed methanol is being produced by living organisms within Enceladus' subsurface ocean. In Earth's oceans, Mincher & Aicher (Reference Mincer and Aicher2016) find that methanol is produced at a rate as high as 0.3% w.r.t. the total cellular carbon. If we assume that methanol production within subsurface oceans on Enceladus is similar to the Earth and all of the carbon measured in Enceladus' vents is expelled by microbes in the subsurface ocean (i.e. C:H₂O~5%; Waite et al. Reference Waite, Lewis, Magee, Lunine, McKinnon, Glein, Mousis, Young, Brockwell, Westlake, Nguyen, Teolis, Niemann, McNutt, Perry and Ip2009), then we would expect a CH₃OH:H₂O abundance within Enceladus' subsurface oceans to be ~0.015% in an organic environment. This result is similar to the CH₃OH abundance measured in the direct vicinity of the vents by Cassini at ~0.01% (Waite et al., Reference Waite, Lewis, Magee, Lunine, McKinnon, Glein, Mousis, Young, Brockwell, Westlake, Nguyen, Teolis, Niemann, McNutt, Perry and Ip2009), indicating that it is possible for the CH₃OH found in the vents to be a biomarker in an extreme case with the carbon predominantly of biological origin.

In contrast at larger scales, our CH₃OH observation suggests that the gas cloud trailing Enceladus has a CH₃OH:H₂O abundance that is ~50 ×  higher than in the vents. Therefore, it is likely that methanol is being produced once the material is ejected from the subsurface ocean, making it improbable as a biomarker signature in this particular case. In the future, caution should be taken when reporting on the presence of supposed biomarkers, in both solar system and exoplanet environments. The most robust method for investigating the complex chemistry, particularly in subsurface ocean environments, is obtaining observations close to the vents.