Measurement of CH$_3$D on Titan at Submillimeter Wavelengths

TL;DR

This study reports the first detection of monodeuterated methane (CH$_3$D) in Titan's atmosphere using ALMA, providing new insights into Titan's methane D/H ratio and demonstrating the potential for spatially resolved observations.

Contribution

First radio/submillimeter detection of CH$_3$D in Titan's atmosphere, enabling measurement of D/H ratio and opening avenues for spatially resolved studies with ALMA.

Findings

Detected CH$_3$D at 4.6σ and 5.7σ significance levels.

Measured CH$_3$D volume mixing ratio as 6.157×10⁻⁶ in Titan's stratosphere.

Derived D/H ratio in CH$_4$ as (1.033±0.081)×10⁻⁴, consistent with previous measurements.

Abstract

We present the first radio/submillimeter detection of monodeuterated methane (CH$_3$D) in Titan's atmosphere, using archival data from of the Atacama Large Millimeter/submillimeter Array (ALMA). The $J_K=2_1-1_1$ and $J_K=2_0-1_0$ transitions at 465.235 and 465.250 GHz ($\sim0.644$ mm) were measured at significance levels of $4.6\sigma$ and $5.7\sigma$, respectively. These two lines were modeled using the Non-linear optimal Estimator for MultivariatE spectral analySIS (NEMESIS) radiative transfer code to determine the disk-averaged CH$_3$D volume mixing ratio = $6.157\times10^{-6}$ in Titan's stratosphere (at altitudes $\gt130$ km). By comparison with the CH$_4$ vertical abundance profile measured by Cassini-Huygens mass spectrometry, the resulting value for D/H in CH$_4$ is $(1.033\pm0.081)\times10^{-4}$. This is consistent with previous ground-based and in-situ measurements from the Cassini-Huygens mission, though slightly lower than the average of the previous values. Additional CH$_3$D observations at higher spatial resolution will be required to determine a value truly comparable with the Cassini-Huygens CH$_4$ measurements, by measuring CH$_3$D with ALMA close to Titan's equator. In the post-Cassini era, spatially resolved observations of CH$_3$D with ALMA will enable the latitudinal distribution of methane to be determined, making this an important molecule for further studies.

Figures

Full text

1 Introduction

Methane (CH4), one of the primary constituents of Titan’s atmosphere, was first detected by [kuiper_44] and has since been studied through myriad ground- and space-based observations. The photo- and ionic chemistry of CH4 and Titan’s most abundant atmospheric constituent, N2, in the upper atmosphere are responsible for the large quantity of trace organic species discovered during the Voyager-1 and Cassini eras, and are important for the formation of haze, clouds, and Titan’s methane-based hydrological cycle. However, the source of Titan’s atmospheric CH4 reservoir is still unknown, as are the processes by which it may be replenished.

Titan’s CH4 volume mixing ratio was found to be $5.65\times 10^{-2}$ near the surface at the location of the Huygens landing site ($\sim 10^{\circ}$ S) with the Gas Chromatograph Mass Spectrometer (GCMS), decreasing with altitude to the tropopause ($\sim 45$ km); an average stratospheric value of $1.48\times 10^{-2}$ was measured between $\sim 75-140$ km [niemann_10]. This measurement is compatible with the initial determination of stratospheric CH4 abundance derived from the 7.7 $\mu$m $\nu_{4}$ band of CH4 by the Cassini Composite Infrared Spectrometer (CIRS), found to be $(1.6\pm 0.5)\times 10^{-2}$ [flasar_05]. Subsequent observations of the CH4 $\nu_{4}$ band were used to probe Titan’s atmospheric temperature with CIRS throughout the Cassini mission by holding the GCMS CH4 profile as latitudinally invariant. However, studies with the Keck II Near Infrared Spectrometer (NIRSPEC) and the Cassini Visual and Infrared Mapping Spectrometer (VIMS) suggested possible variations in Titan’s atmospheric CH4 [penteado_10a, penteado_10b]. Through the combination of two separate CH4 bands in focal planes 1 and 4 of Cassini/CIRS, [lellouch_14] found the CH4 mole fraction to vary between $\sim 1-1.5\%$ in the lower stratosphere at 12 locations from $70^{\circ}$ N – $80^{\circ}$ S during Titan’s northern winter ($\sim 2005-2010$). This roughly symmetric distribution of CH4 may persist throughout a Titan year ($\sim 29.5$ years), and presents important implications for photochemical and dynamical models of Titan’s atmosphere.

Monodeuterated methane (CH3D) was tentatively identified on Titan by [gillett_75] in observations of the 8.6 $\mu$m band with the Kitt Peak National Observatory (KPNO), and confirmed by [kim_82] in combination with data from the Infrared Telescope Facility (IRTF) and the Voyager-1 Infrared Radiometer Interferometer and Spectrometer (IRIS). Later observations with these facilities [owen_86, debergh_88, coustenis_89b, orton_92, penteado_05], the Infrared Space Observatory (ISO; [coustenis_03]), and through both components of the Cassini-Huygens mission [coustenis_07, bezard_07, niemann_10, abbas_10, nixon_12] have further constrained Titan’s atmospheric deuterium-to-hydrogen ratio (D/H). With an average value measured to be $1.36\times 10^{-4}$ during the Cassini era (see [nixon_12], and references therein), Titan is closer to Earth in terms of atmospheric deuterium enrichment than the giant planets, leading to possible constraints on the formation and sustainability of its atmosphere ([mousis_02b]; [bezard_14], and references therein).

Due to the dependence on atmospheric temperature for many CH4 observations in the IR, measurements of Titan’s CH3D abundance may provide a second method for determining possible CH4 variations with latitude, while further constraints on Titan’s D/H will help unravel the history of its formation [mandt_09]. Previous searches for CH3D in the interstellar medium using sub-mm facilities were unsuccessful [pickett_80, womack_96], or provided a tenuous detection of the $J=1-0$ transition [sakai_12]. However, the advent of the Atacama Large Millimeter/submillimeter Array (ALMA) and Titan’s relatively large atmospheric CH3D content provide a new method to study this molecule with high spatial and spectral resolution from the ground. Here, we detail the first definitive detection of CH3D at sub-mm/radio wavelengths. Though these measurements were produced from short ALMA flux calibration observations of Titan and do not utilize the full capabilities of the array, they demonstrate the means by which to study Titan’s CH4 distribution and D/H in the post-Cassini era.

2 Observations

Titan is often observed as a flux calibration object for ALMA observations, resulting in many short ($\sim 157$ s) observations of the moon from 2012 onward. As of 2015, ALMA Band 8 ($\sim 385-500$ GHz) has been available for use, enabling a search for the $J_{K}=2_{1}-1_{1}$ and $2_{0}-1_{0}$ and transitions of CH3D at 465.235 and 465.250 GHz ($\sim 0.644$ mm). The $J=1-0$ and $J=3-2$ transitions at 232.644 GHz (1.289 mm) and $697.690-697.781$ GHz ($\sim 0.430$ mm), respectively, may also be observable with ALMA in Bands 6 and 9. However, the former has a lower line strength than the $J=2-1$ transitions, and was not detected in the dataset analyzed in [thelen_18] to model the nearby CO $J=2-1$ transition at 230.538 GHz. As of the time of writing, the only 2 observations of Titan at frequencies covering the CH3D $J=3-2$ transitions were taken with the lowest frequency resolution usable with ALMA (31250 kHz), and thus the CH3D lines were unresolved (if present). Similarly, observations containing the $J=2-1$ frequencies use low spectral resolution settings or the Atacama Compact Array (ACA), with the exception of an observation on 2015 May 02 at UT 03:57:08, for ALMA Project Code $\#$2013.1.00227.S (where Titan was used for flux calibration). We detected both $J=2-1$ transitions of CH3D at 4.6$\sigma$ and 5.7$\sigma$ in these data from May 2015, shown in Fig. 1. Despite the somewhat coarse spectral resolution inherent to these data (3904 kHz), we resolve both spectral lines of CH3D, which are unresolved in additional archival data at 15625 kHz resolution. The analogous $J=2-1$ transitions of the 13C-substituted form of CH3D at 464.838 and 464.854 GHz (directly abutting the HC3N $\nu_{7}=1$ line, Fig. 1) were not detected in these data, despite the previous detection of this molecule in the IR [bezard_07].

We followed the methodology described in previous works utilizing ALMA flux calibration data of Titan (e.g. [cordiner_14, lai_17, thelen_18, thelen_19]) to obtain disk-averaged spectra and calibrated image cubes. Using the Common Astronomy Software Applications (CASA) package version 5.1.2, we re-ran the ALMA data calibration scripts available with raw data from the ALMA Science Archive222http://almascience.nrao.edu to ensure that spectral lines in Titan’s atmosphere are not flagged (such as HC3N and CO), and to include the updated flux model for Titan (see ALMA Memo $\#$594333http://library.nrao.edu/public/memos/alma/main/memo594.pdf). Imaging was completed using the Hogböm clean algorithm with a pixel size = $0.1\times 0.1^{\prime\prime}$ and a flux threshold of 10 mJy – roughly twice the noise level. The beam full width at half maximum (FWHM) = $0.767\times 0.491^{\prime\prime}$ for this observation, which is comparable to Titan’s angular size on the sky ($\sim 0.7-1.0^{\prime\prime}$, depending on distance and the inclusion of Titan’s substantial atmosphere). Integrated flux maps of the CH3D lines and both nearby HC3N transitions are shown in Fig. 2. Due to the relatively low signal-to-noise ratio of the CH3D and HC3N $\nu_{7}=1$ lines in this observation and the large beam size compared to Titan’s disk, these data are unsuitable for nuanced interpretation of latitudinal variations in Titan’s atmosphere. Despite this, the spatial distribution of both HC3N maps are consistent with contemporaneous datasets analyzed by [cordiner_17a] and [thelen_19], and the discrepancies in flux found between $\nu=0$ and vibrationally excited HC3N lines discussed in [cordiner_18].

3 Radiative Transfer Modeling

We extracted a disk-averaged spectrum of Titan by defining a mask of pixels containing at least $90\%$ of Titan’s continuum flux, as in [lai_17]. Using the method in previous studies (e.g. [molter_16, thelen_18]) to extract flux from within $2\times$ the ALMA PSF resulted in a dilution of the disk-averaged CH3D lines in these data. The resulting spectrum is shown in Fig. 1. We then converted the disk-averaged flux density into radiance units (nW cm*-2* sr*-1* / cm*-1*) as described in Appendix A of [teanby_13] using 24 line of sight emission angles to model the flux from the full disk of Titan and the atmosphere up to 1200 km above the surface. The array of field-of-view averaging points increases in density towards the limb of Titan to account for limb brightening, similar to previous disk-averaged models of Titan’s atmosphere in the sub-mm [teanby_13, thelen_18]. Although additional emission angles may be required to accurately model higher spatial resolution data (e.g. [thelen_19]), summing over 24 annuli was sufficient to model the disk-averaged radiance from these data. Radiances were converted using Titan’s distance of 9.038 AU. The sub-observer latitude was $24.55^{\circ}$ and the Doppler shift was $-7.5$ km s*-1*; these values were obtained from the JPL Horizons ephemerides generator444https://ssd.jpl.nasa.gov/horizons.cgi. We then generated initial radiative transfer models using the Non-linear optimal Estimator for MultivariatE spectral analySIS (NEMESIS) software package [irwin_08] to determine any offsets in the continuum level.

Titan’s continuum was modeled using collisionally-induced absorption from CH4, H2, and N2 pairs (including isotopes), with coefficients from [borysow_86a, borysow_86b, borysow_86c, borysow_87, borysow_91]; and [borysow_93]. As Titan’s continuum in the sub-mm is largely dependent on the tropopause temperature, we constructed a disk-averaged temperature profile using contemporaneous ALMA measurements from [thelen_18] in the stratosphere and above, and an interpolation of tropospheric temperatures found using data from the Huygens Atmospheric Structure Instrument (HASI; [fulchignoni_05]) and those obtained through Cassini radio occultations [schinder_12]. This resulted in a small ($\textless 5\%$ of the initial radiance), constant offset between the data and model, rectified by multiplying the data by a factor of 1.047. This scaling factor is consistent with previous studies using ALMA flux calibration data (see [thelen_19], and references therein), and is due to minor discrepancies between our Titan model and the one used in the CASA reduction package for ALMA data (discussed further in [thelen_18]).

To obtain an accurate radiative transfer model of the CH3D lines emitted in Titan’s atmosphere, we ran NEMESIS in line-by-line mode over a small spectral range, shown in Fig. 3. CH3D spectral line and partition function parameters were obtained from the Cologne Database for Molecular Spectroscopy (CDMS; [muller_01, womack_96, bray_17]). Line broadening and temperature dependence coefficients were obtained from the HITRAN 2012 database [rothman_13]. Our NEMESIS model also consists of nearby species that may influence the CH3D line shape and nearby continuum, and an appropriate temperature-pressure profile for Titan during northern summer. We thus include N2 and CH4 vertical abundance profiles from [niemann_10] and [teanby_13]; CO with a constant volume mixing ratio of 50 ppm (see [serigano_16], and references therein); HC3N, using the 2015 disk-averaged abundance profile found in [thelen_19]. Though nearby emission lines of C2H3CN and C2H5CN do not strongly affect the line shape of the CH3D lines and are undetected in the modeled spectral range of these data, these species were included for completeness using abundance profiles from [lai_17]. The vertical temperature and abundance profiles included in our model are shown in Fig. 4.

Due to the somewhat coarse spectral resolution settings of this observation and the relatively low signal-to-noise ratio of both CH3D lines (compared to the nearby HC3N line, for example), we opted to find a best-fit spectrum using a simple scaling model (i.e. a constant multiplicative factor) of an a priori CH3D profile. All other modeling parameters, primarily the temperature and additional gas abundance profiles, were held constant, and the CH3D profile was set to retain the initial shape of the a priori input (that is, NEMESIS did not retrieve a continuous vertical profile). Our initial CH3D profiles were found by multiplying in situ CH4 data obtained with the Huygens probe [niemann_10] by various D/H ratios. These include measurements made during the Cassini era from previous ground-based studies [penteado_05, debergh_12], the Huygens GCMS [niemann_10], and with Cassini/CIRS [coustenis_07, bezard_07, abbas_10, nixon_12], and cover a range of D/H ratios from $(1.13-1.59)\times 10^{-4}$. Spectral models using the D/H ratios = $1.17\times 10^{-4}$ and $1.59\times 10^{-4}$ from [coustenis_07] and [nixon_12], respectively, are shown in Fig. 3 (magenta and blue spectra); the corresponding CH3D abundance profiles are shown in Fig. 4 (black lines). Models using all a priori abundance profiles converge on a single best-fit spectrum shown in Fig. 3 (gold), with scaling factors producing a common CH3D profile shown in Fig. 4 (gold line). Our retrievals were most sensitive between $\sim 100-200$ km, where we find the CH3D abundance to decrease from ($6.455-6.157$)$\times 10^{-6}$.

4 Discussion and Conclusions

Using archival ALMA flux calibration data of Titan from 2015, we have produced the first definitive detection of CH3D in the sub-mm. While integrated flux maps of the ground ($\nu=0$) and vibrationally excited ($\nu_{7}=1$) HC3N transitions produced from the same image cube (Fig. 2B, C) show spatial distributions consistent with previous studies of Titan with ALMA [cordiner_17a, cordiner_18, thelen_19], the integrated flux map of CH3D (Fig. 2A) shows a lack of significant variation between the poles (less than $\sim 30\%$). Though this is consistent with a constant CH3D profile (and that of its parent molecule, CH4) with latitude, the low spatial resolution and S/N of the CH3D lines prohibit the interpretation of possible latitudinal variations seen in the higher resolution data acquired with Cassini/VIMS and CIRS [penteado_10b, lellouch_14]. We thus measured the disk-averaged abundance of CH3D using the NEMESIS radiative transfer code to be a constant value = $6.157\times 10^{-6}$ above $\sim 130$ km, where our measurements are most sensitive. When taken with the CH4 profile found by the Huygens GCMS [niemann_10], our CH3D abundance yields a D/H = ($1.033\pm 0.081)\times 10^{-4}$.

The D/H ratio found above is within the errors of previous measurements from ISO, IRTF, KPNO, and the Cassini-Huygens mission [coustenis_03, penteado_05, coustenis_07, niemann_10, debergh_12], though generally lower than the average of measurements made during the Cassini era ($1.36\times 10^{-4}$; [nixon_12]), and from Voyager-1 ($1.5\times 10^{-4}$; [coustenis_89b]). The addition of our measurement results in a new weighted mean D/H = ($1.203^{+0.057}_{-0.054})\times 10^{-4}$ for observations made during the Cassini/Huygens mission, or utilizing GCMS results. The error bars on our measurement are relative, and only reflect the retrieval errors of our best-fit synthetic CH3D spectrum. As we cannot simultaneously detect CH4 with ALMA, any D/H measurements rely on previous data obtained in the IR by ISO, Voyager-1, Cassini, various ground-based facilities, or in situ data from the Huygens probe. As the GCMS abundance profile of CH4 was determined at the probe’s landing site ($\sim 10^{\circ}$ S), disk-averaged measurements of Titan with ALMA may not yield a completely comparable D/H to previous, self-consistent measurements with the Cassini-Huygens mission. As shown during the Voyager-1 mission and with the ISO [coustenis_89b, coustenis_03], the measurement of Titan’s D/H ratio is often not straightforward without a well constrained measurement of CH4. These early results were determined using CH4 abundances = 1.8–1.9$\%$, which would significantly lower our D/H measurement ($\sim 8.1\times 10^{-5}$) if used instead of the Huygens/GCMS results. Further, as there is only one dataset on the ALMA archive (at the time of writing) with the spectral settings and resolution capable of resolving both $J=2-1$ transitions of CH3D, we were unable to stack measurement sets to produce a higher S/N detection.

As such, the D/H value we report should be taken with caution, as there are many potential sources of error that are difficult to quantify in relatively low spatial resolution ALMA data – including variations in temperature and CH4 abundance with latitude. However, despite our relatively low disk-averaged value of D/H compared to Cassini-Huygens measurements, it remains larger than the D/H values derived from CH4 abundances for Jupiter ($2.2\times 10^{-5}$, [lellouch_01]) and Saturn ($1.6\times 10^{-5}$, [fletcher_09]) by almost an order of magnitude and is more in line with terrestrial D/H measurements in Earth’s oceans (D/H for Vienna Standard Mean Ocean Water is $\sim 1.56\times 10^{-4}$; see, e.g. [alexander_12]). Thus, our measurement is consistent with previous observations and models that indicate Titan’s CH4 reservoir did not evolve from a Saturnian D/H value (e.g. through photochemistry or serpentinization reactions within Titan’s interior; [cordier_08, mousis_09b]). Instead, Titan was possibly formed from an aggregate of planetesimals that were initially enriched in deuterium from interstellar CH4. The isotopic exchange between infalling gas phase CH4 and H2 in the solar nebula eventually resulted in an enhanced D/H value, which was then preserved as CH4 condensed or was trapped within water-ice clathrates in the Saturnian subnebula [mousis_02a, mousis_02b]. Titan’s D/H value was then modified through photochemistry, diffusion, atmospheric escape, and the outgassing of CH4 from the interior [cordier_08, mandt_09], finally resulting in near an order of magnitude greater deuterium enrichment compared to Saturn as we find here.

Despite the uncertainties of our D/H measurement, the detection of CH3D in the sub-mm allows for independent observations of the spatial variation in atmospheric CH4 without the large dependence on simultaneous temperature measurements (as are required in many IR observations). As noted above, the lack of ALMA flux calibration observations covering the $J=2-1$ and the adjacent transitions of CH3D presents a need for targeted study of Titan with longer integration times and adequate spectral resolution. With ALMA’s high spatial resolution capabilities, further studies of CH3D may enable mapping of CH4 down to 100’s of km allowing for a follow-up to the study by [lellouch_14]; higher S/N data will also result in CH3D abundance measurements from a larger range of altitudes. Finally, measuring Titan’s CH3D abundance near the equator allows for a more constrained D/H when determined using data from the Huygens probe.

5 Acknowledgments

AET and MAC were funded by the National Science Foundation Grant $\#$AST-1616306. CAN and MAC received funding from NASA’s Solar System Observations Program. CAN was supported by the NASA Astrobiology Institute. PGJI acknowledges the support of the UK Science and Technology Facilities Council.

This paper makes use of the following ALMA data: ADS/JAO.ALMA$\#$2013.1.00227.S, 2013.1.01010.S, and 2012.1.00688.S. ALMA is a partnership of ESO (representing its member states), NSF (USA) and NINS (Japan), together with NRC (Canada) and NSC and ASIAA (Taiwan) and KASI (Republic of Korea), in cooperation with the Republic of Chile. The Joint ALMA Observatory is operated by ESO, AUI/NRAO and NAOJ. The National Radio Astronomy Observatory is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc.