Hypervolatiles in a Jupiter-family Comet: Observations of 45P/Honda–Mrkos–Pajdušáková Using iSHELL at the NASA-IRTF

The recent commissioning of iSHELL (Rayner et al. 2012, 2016), which replaced the heritage CSHELL spectrograph and became available for use through proposal peer review beginning in 2016 October, has opened a new realm for the study of native ice abundances in comets. This powerful new instrument¹¹ features both high-resolution long-slit spectroscopy and imaging/active guiding in the near-IR (λ = 1.1–5.4 μm), with eXecutable User Interfaces (XUIs) for each patterned after SpeX, the IRTF moderate-resolution spectrograph/imager.

The iSHELL long-slit echelle spectrograph incorporates a 2048 × 2048 pixel Hawaii H2RG array detector and delivers spectral resolving power (λ/Δλ) up to ∼7.5 × 10⁴ with the narrowest (0375-wide) slit. This slit maps a spatially uniform monochromatic source (e.g., an arc lamp or sky emission line) onto 3 spectral pixels (FWHM), thus providing an effective 0125 pixel^–1 plate scale at the entrance slit. Additional slit widths are available, including 4'' for minimizing slit losses and thereby achieving accurate absolute flux calibration. For our observations of 45P, we used a 075 (6 pixel) wide slit. This delivers λ/Δλ ≈ 3.8 × 10⁴ for a uniform monochromatic source; our comet spectra were consistent with this or slightly higher values (see Section 3.2).

Through its cross-dispersing capability, iSHELL simultaneously measures signal in many (>10) consecutive echelle orders (Figure 1). In the spatial dimension (i.e., along the slit), its cross-dispersing optics produce a (modestly) wavelength-dependent plate scale on the array of 016–017 pixel^–1 for the orders included in our study. Multiple slit lengths are available; our observations were conducted at λ > 2.8 μm, allowing use of a 15'' slit. This permitted detailed spatial analysis of emission profiles (Section 5.5).

Zoom In Zoom Out Reset image size

The free spectral range is matched to the array at λ ≈ 4 μm, meaning that iSHELL is spectrally continuous for shorter wavelengths (i.e., there are no interorder "gaps" in wavelength), making it unique among IR echelle spectrographs. A given L or L-prime instrument setting spans a total range of 10%–11% of the selected wavelength (e.g., as seen in Figure 1). At longer wavelengths (∼4.5–5.3 μm), two identical cross-dispersing gratings are required to achieve full spectral coverage. These gratings are tilted slightly relative to each other in the dispersion direction, such that one extends to the blue end of each encompassed order and the other to the red end, with approximately 85% overlap in their wavelength coverage.

The iSHELL imager has a 42'' circular field of view and a pixel scale of approximately 01. Filter imaging is performed in specific wavelength bands, independent of the wavelength regime being used to obtain spectra. The slit-viewing camera allows for guiding on targets while obtaining spectra. As with SpeX, active (or "auto") guiding is achieved through periodic measurement of the photo-center for "spillover" light (i.e., that not passed through the entrance slit), provided that the target is sufficiently bright and the background sky is sufficiently stable. The centroid position is relayed to the Telescope Control System to correct pointing in real time as needed. Active guiding was not possible for our (daytime) observations of 45P (see Section 3.1).

The hypervolatiles CO and CH₄ are substantially underrepresented in taxonomic studies of parent volatile composition in JFCs. The difficulty associated with these measurements is illustrated in Figure 2, in which histograms (panels (a) and (b)) show the very few measurements of CH₄ and CO in JFCs (these panels are taken from Figure 2 of Dello Russo et al. 2016a). Panel (c) illustrates the high degree of accuracy required to separate lines of cometary CH₄ from their (opaque) telluric counterparts, even for Δ_dot = 10–15 km s⁻¹ and in the case of a relatively bright comet having a substantial abundance ratio CH₄/H₂O; the example shown is C/2001 A2 LINEAR (CH₄/H₂O = 1.5 × 10⁻²), obtained with NIRSPEC at Keck II (λ/Δλ ≈ 24,000; Gibb et al. 2007). The higher spectral resolving power of iSHELL, together with the daytime observing capability of the IRTF, ameliorates this difficulty, although disentangling cometary emission from telluric absorption remains problematic for small Doppler shifts. Near perihelion, 45P had highly favorable Δ_dot (Table 1), displacing lines of CH₄ and CO well away from their absorbing counterparts in the telluric atmosphere.

Zoom In Zoom Out Reset image size

Table 1.  Log of IR Spectral Observations of Comet 45P/H-M-P with iSHELL^a

2017 January UT Date	Setting IDb	Δνc (cm−1)	UT Start–End	Rh (au)	Δ (au)	Δdot (km s−1)	Molecule(s) (orders)d	Tinte (s)	Slit PA (deg)f
06/07	M2		23:04–00:37	0.552	0.567	−35.56		1051	74.5
		2159.6–2141.5, 2140.2–2122.2					CO, H2O (111, 110)
07	Lp1		00:56–01:48	0.552	0.565	−35.36		1186	74.5
		3043.5–3019.1, 3005.1–2980.6					CH4, OH* (157, 155)
		3005.1–2980.6					C2H6,CH3OH,CH4,OH* (155)
		3005.1–2980.6, 2851.0–2827.9					CH3OH, OH* (155, 147)
		2851.0–2827.9, 2793.1–2770.6					H2CO, OH* (147, 144)
07	L1		02:28–02:58	0.552	0.564	−35.19		830	74.5
		3465.8–3438.0					H2O (179)
	BS 8709g
07/08	L2		22:35-00:18	0.558	0.546	−34.98		1779	74.0
		3389.0–3361.8					H2O (175)
		3312.4–3285.6					HCN, NH3, C2H2 (171)
		3389.0–3361.8, 3312.4–3285.6					NH3 (175, 171)
08	M2		00:37–01:28	0.558	0.545	−34.86		1751	74.0
		2159.6–2141.5, 2140.2–2122.2					CO, H2O (111, 110)
08	Lp2		02:09–02:54	0.558	0.544	−34.70		1186	74.0
		2793.1–2770.6					H2CO, OH* (144)
	BS 8709g

Notes.

^aThe slit dimensions used for 45P were 075 × 15''. The spatial scale on the spectrometer array is approximately 016–017 pixel⁻¹ in the orders analyzed for our study. Our results are not sensitive to the exact value used; we adopted the value 0167 pixel⁻¹ in calculating production rates and abundance ratios (Table 2). ^bSpectral settings were chosen to optimize the sampling of molecular composition in comet 45P. ^cApproximate frequency range covered within each targeted order. ^dEncompassed molecules/radicals (OH*) that are addressed in this paper. Echelle order(s) in which emissions occur are listed in parentheses. ^eTotal on-source integration times, by setting. ^fSlit position angle on the sky, measured clockwise from north as usual. On both dates this corresponds to the extended heliocentric radius vector (i.e., the projected Sun–comet direction). ^gIR flux standard star observed on each date, measured in all settings used for 45P. BS 8709 is spectral type A3 V, L/Lp-magnitude 3.05, and M-magnitude 3.06. The slit dimensions used for BS 8709 were 40 × 15'', chosen to maximize the fraction of stellar flux transmitted through the slit.

Download table as:  ASCIITypeset image

We targeted 45P shortly prior to (on UT 2016 December 30–31) and following (on UT 2017 January 6–8) its perihelion passage. On all dates 45P was located <30° from the Sun, and therefore all observations were conducted during daytime. The pre-perihelion dates were weathered out, but we obtained high-quality post-perihelion spectra. Here we report results from two days, bridging UT 2017 January 6/7 and 7/8 (Table 1). For convenience these will henceforth be referred to as (UT) January 7 and 8, respectively.

We first imaged Venus (to establish telescope pointing), then a bright star near our target, and finally 45P using a broadband K filter. Only this and shorter-wavelength filters were available at the time of our observations of 45P. Broadband and narrowband L and L-prime and narrowband M filters have since been added.

For the small solar elongation of 45P, the background sky level was relatively high. We found that short-term background fluctuations were comparable to or greater than the comet signal, and this precluded using the iSHELL auto-guider. Instead, we established nonsidereal telescope tracking rates that were adequate to maintain the comet at a constant location on the array—within the seeing disk, approximately 09–1'' and 12–15 on January 7 and 8, respectively—for the duration of a spectral observing sequence (amounting to 3–5 minutes of clock time). We acquired spectra using our standard ABBA sequence of telescope nodding, with approximately 30–40 s of integration time in each of the four scans, which were stored as individual frames. Both A and B beams were in the slit, spatially separated by half its length (75) and positioned symmetrically with respect to its midpoint (see Figures 1 and 3(a)). We oriented the slit along the projected Sun–comet direction on the dates of our observations (Table 1), which permitted comparing release of material into sunward- versus anti-sunward-facing hemispheres (see Figure 7, and discussion in Section 5.5).

Zoom In Zoom Out Reset image size

At the end of each ABBA spectral sequence, we moved the telescope 10'' north (approximately perpendicular to the slit) and re-imaged the comet; this permitted assessing its position more clearly. We then repositioned it at the fiducial A-beam location and began the next spectral sequence. This process was repeated until adequate signal-to-noise ratio was accumulated in targeted cometary emissions, based on running stacks of raw difference frames (A – B) such as that shown in Figure 1. The comet was bright enough to recognize continuum and/or brighter emission lines in most individual A–B spectral difference frames, as displayed in the spectrometer data viewer XUI.

Our methodology for processing spectra is detailed in the refereed literature (Bonev 2005; DiSanti et al. 2006, 2014; Villanueva et al. 2009; Radeva et al. 2010) and so will not be repeated here, except for aspects pertaining uniquely to the layout of iSHELL orders on the array. In contrast to the tilting (or slanting) of orders in long-slit spectral data obtained with other high-resolution infrared spectrometers (specifically, NIRSPEC/Keck II, CSHELL/IRTF, CRIRES/VLT, and IRCS/Subaru), the curvature ("bowing") of iSHELL orders required isolating these and spatially resampling (i.e., straightening) through use of a second-order polynomial (a parabolic) function. Once accomplished, the subsequent steps of (1) spectral resampling through registration of observed sky emission lines along the slit, (2) obtaining the approximate spectral calibration through comparison of sky line positions with a modeled telluric radiance spectrum, and (3) improving spectral calibration and establishing spectral resolving power and column burdens of terrestrial atmospheric absorbers were accomplished as discussed previously. All spectra of 45P shown in Figures 3–5 represent the signal contained in 15 spatial pixels (amounting to a 075 × 25 aperture) centered on the peak comet signal, taken to be the approximate line of sight to the nucleus. These 15-row spectral extracts were performed on processed orders, using the combined A- and B-beam signal. Figure 4(c) shows a two-dimensional (spatial-spectral) representation of the combined signal from order 155, separate beams of which are shown in raw form in Figure 1.

For all comet settings, we also obtained spectra of the bright IR flux standard star BS 8709 using the 4''-wide slit (Table 1). Comparisons between observed stellar spectra and modeled atmospheric transmittance were performed (step 3 above). We used active guiding on BS 8709; thus, its spectral resolving power depended primarily on the stellar point-spread function and hence on atmospheric seeing. Our values for BS 8709 consistently clustered around λ/Δλ = 2 × 10⁴, considerably larger than that expected for uniform monochromatic illumination of the 4'' slit (∼(6–8) × 10³).

Contributions from continuum and gaseous emissions in the comet spectra were established as described previously (e.g., DiSanti et al. 2016); the procedure is illustrated in Figures 3–5. The fully resolved transmittance function was convolved to the spectral resolving power of the data (approximately 4.0 × 10⁴) and scaled to the observed comet continuum. Cometary emission lines were isolated by subtracting the modeled continuum, and this difference, representing the observed net emission spectrum, was compared with molecular fluorescence models.

In practice, all models (continuum and molecular) were fit simultaneously to each observed comet spectrum so as to minimize rms deviations in the overall residuals (corresponding to observed spectrum minus modeled continuum plus molecular fluorescence), shown at the bottom of each spectral panel in Figures 3–5 together with the measured ±1σ stochastic noise envelope (see Section 3.4.1). This fitting utilized an algorithm developed previously for interpreting spectra of Mars's atmosphere (Villanueva et al. 2008). Its application to comet spectra is detailed in Appendices A.1 and A.2 of DiSanti et al. (2016).

We applied the column burdens of absorbers measured for BS 8709 to the spectra of 45P. We then varied the spectral resolving power so as to minimize the rms deviations in the overall comet residuals. This involved comparison of models with observed spectra, utilizing both telluric absorption features and cometary emissions. (At the resolution of iSHELL, cometary emission lines are not resolved, but instead are sampled at the instrumental resolution.) This led to values of λ/Δλ consistent with or slightly higher than that measured for uniform, monochromatic illumination of the 075-wide slit (3.8 × 10⁴). For some spectra, we found that values as high as ∼4.5 × 10⁴ were required to minimize rms deviations in the cometary residuals. We attribute this to nonuniform illumination of the slit by the cometary coma, the brightness of which peaked for lines of sight passing close to the nucleus. We tested the sensitivity of our results to λ/Δλ, determining molecular production rates and abundance ratios for values ranging from 3.8 to 4.5 × 10⁴. Over this range of modeled spectral resolving power, our results remained constant to well within 1σ uncertainty. For this reason all production rates and abundance ratios reported here are based on λ/Δλ = 4.0 × 10⁴.

We obtained "nucleus-centered" production rates (Q_nc) using a well-established formalism (Dello Russo et al. 1998; DiSanti et al. 2001; Bonev 2005; Villanueva et al. 2011a). This relates observed emission line flux to (modeled) fluorescence efficiency (g-factor) multiplied by atmospheric transmittance at each Doppler-shifted line frequency and includes molecular photodissociation lifetime together with geometric parameters. Additional details are found in Section 3.2.2 of DiSanti et al. (2016), and papers detailing the fluorescence models used in our analysis are cited by molecule in Section 3.2 of Paganini et al. (2014). Individual line g-factors depend on rotational temperature (T_rot), which for our present study is consistent with 90 K, as determined from measurement of H₂O lines (see Figure 5(a), and Section 3.5.1) in a spectral region first identified from spectral analysis of comet 153P/Ikeya–Zhang (Dello Russo et al. 2004).

The Q_nc were then multiplied by an appropriate growth factor (GF), determined through the well-documented "Q-curve" method for analyzing spatial profiles of emissions and dating to our study of carbonyl sulfide in C/1995 O1 (Hale-Bopp; Dello Russo et al. 1998). This primarily corrects for atmospheric seeing—which invariably suppresses signal along lines of sight passing close to the nucleus—while also accounting for potential slight drift of the comet during the course of the ABBA spectral sequences. The product Q_nc × GF is taken to be the "total" (or "global") molecular production rate, denoted Q_tot. Our results for 45P are summarized in Table 2.

Table 2.  Molecular Production Rates and Abundances in 45P/H-M-P^a

2017 UT Date	Setting/Order	Species	Γb	Qncc 1026 mol s−1	GFd	Qtote 1026 mol s−1	Qx/${Q}_{H2O}$ f
Jan 6/7	M2/111	H2O	6.06	172 ± 9.0	(1.62 ± 0.09)	280 ± 19
		CO		1.04 ± 0.09		1.69 ± 0.16	0.61 ± 0.05
	M2/110	H2O	5.57	162 ± 11	(1.63 ± 0.10)	263 ± 20
		CO		0.96 ± 0.11		1.56 ± 0.18	0.56 ± 0.06
		CO					0.59 ± 0.04g
		H2O		171 ± 8.9c	1.62 ± 0.07	277 ± 18
Jan 7	Lp1/157	OH*	8.80	110 ± 31	(2.16 ± 0.54)	204 ± 58
		CH4		0.80 ± 0.06	(1.91 ± 0.16)	1.49 ± 0.13	0.77 ± 0.07
	Lp1/155	OH*	8.55	88.4 ± 12.6	(1.70 ± 0.19)	165 ± 24
		CH4		0.83 ± 0.11		1.55 ± 0.20	0.81 ± 0.11
		C2H6		0.54 ± 0.03	(1.87 ± 0.10)	0.99 ± 0.07	0.52 ± 0.04
		CH3OH		4.01 ± 0.24		7.49 ± 0.52	3.90 ± 0.33
	Lp1/147	OH*	8.43	115 ± 15	(1.87 ± 0.16)	213 ± 28
		CH3OH		3.48 ± 0.22	(1.82 ± 0.20)	6.44 ± 0.45	3.35 ± 0.28
		H2CO		0.48 ± 0.08		0.88 ± 0.15	0.46 ± 0.08
	Lp1/144	OH*	7.85	118 ± 21	(1.96 ± 0.27)	217 ± 40
		H2CO		0.30 ± 0.06		0.55 ± 0.11	0.28 ± 0.06
		CH4					0.79 ± 0.06g
		CH3OH					3.59 ± 0.31g
		H2CO					0.34 ± 0.08g
		OH*			(1.82 ± 0.12)h
		H2O		104 ± 9c	1.85 ± 0.06h	192 ± 18
Jan 7	L1/179	H2O	8.25	99.1 ± 5.1	2.28 ± 0.07	226 ± 14
Jan 7/8	L2/175	H2O	16.0	165 ± 9.5	(1.94 ± 0.24)	320 ± 44
		OH*		173 ± 13		337 ± 49
		NH3		<2.06		<4.03	<1.24
	L2/171	OH*	15.1	130 ± 21		257 ± 48
		HCN		0.080 ± 0.020		0.156 ± 0.042	0.049 ± 0.012
		C2H2		0.126 ± 0.043		0.245 ± 0.088	0.076 ± 0.026
		NH3		1.06 ± 0.39		2.06 ± 0.80	0.64 ± 0.24
		H2O		165 ± 9.2c	1.94 ± 0.24	325 ± 34
Jan 8	M2/111	H2O	6.80	143 ± 7.8	(2.36 ± 0.11)	310 ± 31
		CO		1.09 ± 0.12		2.35 ± 0.32	0.72 ± 0.07
	M2/110	H2O	6.51	172 ± 10	(2.00 ± 0.13)	372 ± 38
		CO		0.86 ± 0.07		1.87 ± 0.21	0.57 ± 0.04
		CO					0.61 ± 0.09g
		H2O		150 ± 7.9c	2.17 ± 0.18	326 ± 32
Jan 8	Lp2/144	OH*	7.88	183 ± 18c	2.77 ± 0.40	508 ± 89
		H2CO		0.669 ± 0.055		1.86 ± 0.31	0.37 ± 0.04

Notes.

^aValues assume a rotational temperature of 90 K, based on relative intensities among H₂O lines in a "nucleus-centered" extract (see Figure 5(a) and related text in Section 3.5.1). All uncertainties represent 1σ, and the upper limit for NH₃ in order 175 of the L2 setting represents 3σ. ^bCalibration factor [10⁻¹⁸ W m⁻² (cm⁻¹)⁻¹/(ADU s⁻¹)], from IR flux standard BS 8709 (Table 1). ^cProduction rate based on a nucleus-centered aperture having angular extent 075 × 25 (Section 3.2). The weighted mean Q_nc for H₂O within each iSHELL setting is in italics, in the last line of entries (see Section 4). ^dSlit-loss factor, for converting Q_nc to total (i.e., global) production rate (Q_tot; see note "e"). Within each iSHELL setting, GF was measured for one or more simultaneously observed emission lines having high signal-to-noise ratio. Italicized values in parentheses are those measured for specific molecules; their weighted mean value (listed with the corresponding mean Q_nc for H₂O) is assumed to apply to all co-measured species. ^eTotal production rate. Uncertainties in Q_tot incorporate those in GF and Q_nc, and also an assumed uncertainty of ±5% in Γ. ^fMolecular abundance relative to H₂O = 100. For all species, this is expressed as the ratio of Q_nc to the mean Q_nc for simultaneously measured water (H₂O and/or OH*), listed in italics (note "c"). Entries in bold italics represent weighted mean abundances for molecules measured in multiple echelle orders. ^gAbundances in bold italics are weighted mean values, for species falling in multiple orders. These are CO, CH₄, CH₃OH, and H₂CO on January 7 and CO on January 8. ^hFor Lp1 on January 7, the mean GF for OH* in the four orders is listed (1.82±0.12). The overall GF for this setting (1.85±0.06) incorporates this plus those measured for CH₄, C₂H₆, and CH₃OH.

Download table as:  ASCIITypeset image

3.3.1. Dependence on Gas Outflow Speed

The highly favorable geocentric radial velocity of 45P near perihelion (Δ_dot ≈ −35 km s⁻¹) combined with the sensitivity and high spectral resolution of iSHELL permitted robust measurements of both CO and CH₄ (Table 2). We expect our results for 45P, together with similar measurements of future targets, to establish meaningful statistics for abundance ratios of these two substantially underrepresented hypervolatiles, particularly in JFCs as illustrated unequivocally in Figures 2(a)/(b) (also see Figures 6(a)/(b) and discussion in Section 5.1).

3.4.1. CO

We obtained secure detections of CO measured simultaneously with H₂O (Figure 3), on both January 7 and 8. In Figure 3(a), lines of H₂O are marked (circles) in the two orders (110 and 111) containing the strongest observed lines of CO, with additional H₂O apparent in the two other orders shown. Many more H₂O lines are covered in this setting (M2), which in total encompasses (most of) orders 99–114.

Extracted spectra reveal emissions from both CO and H₂O superimposed on the cometary continuum (Figures 3(b), (c)). The modeled continuum is represented by the gold trace overlaying each spectrum. Below these are fluorescence models (color-coded and labeled by molecule), shifted in frequency according to the Δ_dot of 45P, multiplied by the modeled monochromatic atmospheric transmittance at each Doppler-shifted line frequency, and scaled to the observed emission line intensity in excess of the continuum. The red trace is the overall modeled spectrum (continuum + molecular), and the bottom trace (labeled "residuals") is observed spectrum minus overall modeled spectrum. Figures 4 and 5 show analogous representations for other instrument settings/orders.

Zoom In Zoom Out Reset image size

Abundances of CO in JFCs (plus 2P/Encke) previously obtained by remote sensing number only five. Of these, one is a (3σ) upper limit (in 2P/Encke; Radeva et al. 2013) and one represents a range of values observed with the Hubble Space Telescope (in 103P/Hartley 2; Weaver et al. 2011). The remaining three measurements have ±1σ uncertainties that are very large compared with our measurement of CO in 45P (see Figure 6(a), and Section 5.1).

3.4.2. CH₄

Our study emphasizes measurement of the hypervolatiles CO and CH₄; however, only by characterizing a suite of native volatiles can the chemistry of 45P be framed in a more complete context. Robust measurement of molecular production rates requires knowledge of rotational temperature (T_rot) in the coma. We accomplished this through targeting water near 2.9 μm.

3.5.1. Measurement of Rotational Temperature in Comet 45P

We sampled a spectral region encompassing H₂O lines (Figure 5(a)) that together span a large range of rotational energies (approximately 50–550 cm⁻¹) and provide a robust measure of rotational temperature for realistic values found in cometary comae (values of T_rot among comets range from ∼30 to 150 K). This region was first identified in studies of Halley-family comet 153P/Ikeya–Zhang, using a single CSHELL setting that encompassed nearly the same spectral extent as that shown in Figure 5(a) (Dello Russo et al. 2004). The value of T_rot we measured for 45P is consistent with 90 K.

Zoom In Zoom Out Reset image size

We adopt this value of T_rot for all molecules reported here. To validate this assumption, we also measured T_rot for CH₃OH (Villanueva et al. 2012; DiSanti et al. 2013), isolating the net signal in its ν₃ band (in order 147 of the Lp1 setting) by subtracting (modeled) contributions from continuum, H₂CO, and OH* emissions (see Section 3.5.2 and Figure 5(c)). This provided a value (86⁺¹¹/₋₉ K) consistent with the T_rot we measured for H₂O (89 ± 3 K), albeit less well constrained owing primarily to the smaller range of energies sampled (∼10–200 cm⁻¹) and also to the high density of lines for CH₃OH leading to spectral confusion (see, e.g., discussion in Section 3.1 of Gibb et al. 2012). Similarly, from C₂H₆ ν₇ emission (Villanueva et al. 2011b) in Lp1 order 155 (Figure 4(b)) we measured T_rot = 98⁺¹⁸/₋₁₅ K, even less well constrained than T_rot for CH₃OH but still consistent with our value for H₂O.

We were unable to obtain meaningful measures of T_rot for other species, because their emissions were too weak and/or because their lines did not sample a sufficient range of rotational energies to obtain an accurate determination. Nonetheless, as an additional validation to our approach we note that in most cases rotational temperatures among species derived from IR observations of a given comet are in agreement, and that slight changes in T_rot result in only very small differences in production rates and hence abundance ratios relative to H₂O (Gibb et al. 2012). For these reasons, we take 90 K to be the most probable value for T_rot in 45P and base our results for all molecules on this temperature.

3.5.2. Additional Molecules

Our secure detections of CO and CH₄ in 45P do not simply augment statistics regarding their abundances in JFCs. Rather, they are critical for establishing statistically meaningful abundances for these hypervolatiles in this significantly underrepresented population of comets, as illustrated by the very few previous measurements in JFCs obtained by remote sensing (five for CO, and four for CH₄; Figures 6(a) and (b)).

CO was depleted relative to its median value measured among OCCs. Methane was consistent with its median value among OCCs and seemingly enriched in 45P compared to the few JFCs measured to date. Although the abundance ratios of CO and CH₄ are fairly similar in 45P (0.6%–0.8%), the emerging (and evolving) overall compositional taxonomy among all measured comets (primarily OCCs) suggests they are not correlated (as opposed to anticorrelated). This was first demonstrated by Gibb et al. (2003) based on a sample of eight OCCs, and more recently in a compilation of 19 OCCs, for which the formal Pearson product–moment correlation coefficient (r) was <0.5 (Dello Russo et al. 2016a). Too few JFCs have been measured to assess a meaningful comparison between their CO and CH₄ abundances. The noncorrelated nature of CO and CH₄ abundance ratios in OCCs suggests that thermal processing may not be the dominant factor in determining hypervolatile abundances.

This may also be true for 45P. Its substantial abundance of CH₄ demonstrates that this low sublimation temperature (T_subl) species can be preserved in a JFC, perhaps even trapped in less volatile ice to survive the thermal effects of repeated (and frequent) passages to well within 1 au of the Sun. This supports retention of a significant primitive record. The depleted abundance of CO (relative to its median value among OCCs; Figure 6(a)) may indicate its loss through evolutionary processing or, given its similarly low T_subl to that of CH₄ (see Table 3, note a), its efficient conversion to CH₃OH through H-atom addition reactions on grain surfaces during the formative epoch (Section 5.2.2), demonstrated through irradiation of laboratory ices at very low temperatures (<25 K; Watanabe et al. 2004; Oba et al. 2010).

5.2.1. Hypervolatiles: CO, CH₄, and C₂H₆

5.2.2. Additional Species

5.2.3. Comparisons with Rosetta Measurements of JFC 67P/Churyumov-Gerasimenko

A primary means of addressing sources of radical species in cometary comae is comparing measured abundances from optical and IR observations. Comet 45P was classified as "typical" in terms of its abundance ratios C₂/OH and C₃/OH, and also its CN/OH and NH/OH (A'Hearn et al. 1995). Comparing these to our measured abundance ratios is useful for assessing potential "parent–daughter" relationships, for example, C₂ from C₂H₂, CN from HCN, and NH from NH₃. Our value for NH₃/H₂O in 45P is consistent with those measured among both JFCs and OCCs, whereas C₂H₂/H₂O is consistent with that among JFCs but below that among OCCs. Based on this (admittedly) somewhat simplistic reasoning, it is therefore plausible that the abundance we measured for NH₃/H₂O (0.64% ± 0.24%) can account for the NH/OH measured by A'Hearn et al. (1995; 0.59% ± 0.34%), notwithstanding the large uncertainties associated with both measurements. Nonetheless, in contrast, both C₂/OH (0.21% ± 0.04%) and CN/OH (0.13% ± 0.03%) are considerably higher than the respective 45P values for C₂H₂/H₂O (0.076% ± 0.026%) and HCN/H₂O (0.049% ± 0.012%), suggesting the need for one or more additional sources of both C₂ and CN.

Our results for 45P also contribute information on the volatile composition of comets at small heliocentric distance. This makes 45P the first JFC for which such a broad suite of parent volatiles have been measured at R_h < 0.8 au. Such measurements are rare for both JFCs and OCCs, due largely to a combination of sensitivity limits and observational circumstances. In late 2013, an opportunity afforded itself with the dynamically new Sun-grazing OCC C/2012 S1 (ISON), for which compositional measurements extended inward to R_h < 0.5 au. HCN, H₂CO, NH₃, and C₂H₂ in ISON exhibited increased abundance ratios within R_h ∼ 0.8 au (Dello Russo et al. 2016b; DiSanti et al. 2016). In contrast, 45P had mixing ratios for H₂CO and NH₃ with respect to H₂O that were consistent with their respective median values among OCCs and JFCs, C₂H₂ was consistent with its median JFC value, and HCN was depleted (Figure 6(c)). The mixing ratios we measure for HCN, H₂CO, NH₃, and C₂H₂ in 45P are smaller than the few measurements obtained for OCCs within 0.8 au of the Sun. Our measurements of 45P near perihelion represent an important data point; however, more measurements are needed at small R_h (for both JFCs and OCCs) to test potential abundance variations with R_h.

Figure 7 compares emission profiles from two settings on January 7, the day on which our observations sampled the largest number of molecules having strong emissions. For all spectral observations we oriented the slit along the projected Sun–comet direction (PA 745). It is important to note that both the M2 and Lp1 settings represent independent snapshots of cometary activity, averaged over the time intervals specified in Figure 7(a) (and in Table 1). Over the approximately 1.5 hr interval between the (mid-sequence) times of the M2 and Lp1 observations, the dust continuum changed considerably: the M2 continuum was relatively symmetric about the nucleus (taken to be the zero-point on the x-axis), whereas the Lp1 continuum was clearly enhanced in the anti-sunward-facing hemisphere. The difference in dust coma morphology between M2 and Lp1 may be related to rotation of the nucleus, for which a period of ∼7.5 hr was estimated from radar observations (Lejoly & Howell 2017, CBET 4357(2)).

Zoom In Zoom Out Reset image size

This justifies comparison of spatial profiles measured simultaneously, because these are affected identically, irrespective of changes resulting from potential differences in slit-loss correction factor among settings, or from nucleus rotation. We therefore focus on comparing the distribution of parent volatiles within each iSHELL setting.

Because of the low CO abundance in 45P, we combined the five brightest lines in the M2 setting (R1 and R2 in order 111, and P3–P5 in order 110), and then binned (smoothed) the resulting profile by 2 in the spatial dimension to improve signal-to-noise ratio (Figure 7(b)). Similarly, in the Lp1 setting (Figure 7(c)) we combined nine lines of OH* spanning the four orders analyzed and smoothed the resulting profile by 2. The profile of OH* emission is critically important in that it faithfully tracks the distribution of parent H₂O molecules in the coma (for details, see Bonev et al. 2006, and references therein).

Figure 7(b) shows that during the M2 observations H₂O tracked the continuum fairly closely, whereas CO was significantly asymmetric in the sunward-facing hemisphere and its "secondary" peak (near 500–600 km sunward) was much stronger than the enhancement seen for co-measured H₂O at a similar projected position. In the Lp1 setting, H₂O and CH₃OH tracked the continuum, whereas C₂H₆ and (more so) CH₄ showed sunward asymmetry. These results may be indicative of an activated region that preferentially released more highly volatile species into the sunward-facing hemisphere, and if so, that the degree of asymmetry increased with decreasing native ice sublimation temperature. It also appears to have persisted for at least the nearly 3 hr interval encompassed by the combined measurements in the two settings. Alternatively, the secondary source may have been more pronounced during the M2 observations, and that this was responsible for the highly prominent emission profile intensity for CO offset sunward from the nucleus.

5.5.1. Possible Effects of Spatial Profiles on Molecular Abundance Ratios

Our (global) water production rates in Table 2, using emissions from H₂O and/or OH* measured within each iSHELL setting, generally agree (to within their respective 1σ uncertainties). However, measured GFs increased with time on both days, consistent with degradation in atmospheric seeing arising from direct sunlight warming the telescope environment and structure over the course of each day. This makes it necessary to compare global production rates (Q_tot) for H₂O measured in different settings, when assessing the evolution in gas production in 45P.

On the first day (bridging UT January 6/7), the value of global Q(H₂O) from the M2 observations ((2.77 ± 0.18) × 10²⁸ molecules s⁻¹) was higher than the mean of those from Lp1 and L1 (order 179 only), (2.19 ± 0.07) × 10²⁸ s⁻¹. This suggests a decrease in gas production by 26% ± 9% over approximately a 3 hr period. On the second day (bridging UT January 7/8), global values of Q(H₂O) increased by a factor of 1.56 ± 0.29 between the interval UT22:35–01:28 (based on the mean from L2 and M2 settings, (3.25 ± 0.23) × 10²⁸ s⁻¹) and UT02:09–02:54 (from the Lp2 setting, (5.08 ± 0.89) × 10²⁸ s⁻¹; Table 1). These Lp2 observations had the largest GF measured for any setting on either day; however, we note that Lp2 also had the largest nucleus-centered water production rate for any setting on either day. Given the rotation period estimated from radar measurements of 45P (∼7.5 hr; Lejoly & Howell 2017, CBET 4357(2)), it is conceivable that its activity varied on timescales of hours; however, a more rigorous study than our observations permit is necessary to test such variations.

Overall, the mean water production was larger by 46% ± 11% on January 7/8 than on January 6/7. Whatever the cause of these variations in measured water production (both on each day and day to day), this argues strongly for assessing volatile abundance ratios relative to simultaneously measured water production, as was done for our study and which, as stated previously, provided consistent values for species measured on both days (CO and H₂CO; Section 4).