SPECTROSCOPIC CHARACTERIZATION AND DETECTION OF ETHYL MERCAPTAN IN ORION^*

The spectral millimeter-wave survey of Orion KL carried out with the IRAM 30 m radio telescope presented more than 8000 unidentified lines (Tercero et al. 2010; Tercero 2012). Many of them (nearly 4000) have been identified as lines arising from isotopologues and vibrationally excited states of abundant species (Demyk et al. 2007; Carvajal et al. 2009; Margulès et al. 2009, 2010; Tercero et al. 2012; Motiyenko et al. 2012; Daly et al. 2013; Coudert et al. 2013; Haykal et al. 2014; López et al. 2014). These identifications significantly reduce the number of U-lines and mitigate line confusion in the spectra. Nevertheless, many of those features still remain unidentified. Therefore, a search for new molecular species in that cloud based on precise laboratory measurements continues to be a field of great research activity. Recently, the discovery of methyl acetate and gauche-ethyl formate (Tercero et al. 2013), the search for allyl isocyanide (Haykal et al. 2013), as well as a tentative detection of phenol (Kolesniková et al. 2013) have been reported.

Methanol and ethanol are very well known molecules in many astrophysical environments. The thiol equivalent of methanol, methyl mercaptan (CH₃SH), has been detected toward Sgr B2 by Linke et al. (1979). Previous searches for it toward dark clouds and Orion did not provide positive results (Irvine et al. 1987, 1989). However, this molecule has been recently detected toward the cold first hydrostatic core B1 (Cernicharo et al. 2012) and was also observed toward the hot core G327.3-0.6 by Gibb et al. (2000). Hence, the thiol equivalent of ethanol, ethyl mercaptan (CH₃CH₂SH), could also be present in space. Initial studies of the Stark modulated microwave spectra by Imanov et al. (1967), Hayashi et al. (1970, 1973), Quade et al. (1975), and Nakagawa et al. (1976) provided the first values of the spectroscopic constants for the gauche and trans conformers of ethyl mercaptan. These microwave laboratory data, however, cannot be used to accurately predict their frequencies in the millimeter- and submillimeter-wave domains.

In this Letter, we report new laboratory measurements of ethyl mercaptan in the millimeter- and submillimeter-wave region and its first observation in the interstellar medium. Newly derived spectroscopic constants of its gauche and trans forms have allowed us to detect the gauche conformer toward Orion KL with the IRAM 30 m radio telescope. Trans-ethyl mercaptan has been tentatively detected in this study. In addition, the first detection of methyl mercaptan toward Orion KL is also presented.

A commercial sample of ethyl mercaptan was used without any further purification. Rotational spectra in the 8.7–26.5 GHz region were taken at a temperature of −15°C and a pressure of 6 mTorr with a waveguide chirped-pulse Fourier transform microwave spectrometer at New College of Florida (Reinhold et al. 2011). The waveguide was cooled via a set of home-built cooling loops connected to a chiller with a recirculating fluid (−15°C is the base temperature reached). A 250 ns chirped pulse is generated with an arbitrary waveform generator (Tektronix AWG7101) and subsequently frequency shifted, filtered, and amplified before interacting with a static gas sample in a 10 m coil of WRD-750 waveguide. The resulting molecular free induction decay is amplified and detected in the time domain with an oscilloscope (Tektronix TDS6154C), after which it is Fourier transformed to generate a frequency-domain spectrum. In each spectrum, 2 million free induction decays of 4 μs duration were averaged and Fourier transformed using a Kaiser–Bessel window function. Peaks in the frequency-domain spectrum had an FWHM of ∼800 kHz and so the uncertainty on the line center was set to 80 kHz.

Rotational spectra in the millimeter (60–300 GHz) and submillimeter-wave (500–590 GHz, 625–660 GHz, and 835–880 GHz) regions were recorded at a pressure of approximately 15 mTorr by a recently constructed millimeter-wave spectrometer at the University of Valladolid (A. M. Daly et al. in preparation). The source of radiation is an Agilent E8257D synthesizer (250 kHz–20 GHz) connected to a set of passive or active cascade frequency multipliers (VDI, Inc.). Room temperature measurements were carried out in a free space 360 cm long Pyrex cell. Up to 170 GHz, the optical path length was doubled using a rooftop mirror and a polarization grid. A silicon bolometer was used as the detection element above 800 GHz. At lower frequencies, the signal was detected by either Schottky diodes or the quasi-optical broadband detectors (VDI, Inc.). All spectra were recorded in 1 GHz sections using the frequency modulation technique with second harmonic lock-in detection (modulation depth between 20–50 kHz and modulation frequency of 10.2 kHz for the semiconductor detectors and 90 Hz for the cryogenic detector). Frequency accuracy is estimated to be better than 50 kHz. Rotational spectra of the ³⁴S isotopologue from the microwave up to the submillimeter-wave region were measured in natural abundance.

Ethyl mercaptan is a near prolate asymmetric top molecule which is present in two stable forms: gauche and trans configurations which are defined by the value of the torsion angle α of the –SH group measured from the symmetry plane of the trans configuration (α = 0); see Figure 1(a). For the gauche conformer, two equivalent configurations at approximately α = 120° and 240° can be interchanged by tunneling motion. A schematic potential energy diagram for the –SH group torsion is shown in Figure 1(b). The tunneling process removes the vibrational degeneracy and the vibrational ground state is split into two substates labeled as 0⁺ and 0⁻ with an energy separation ΔE = E⁻ − E⁺ of about 1750 MHz (Quade et al. 1975; Nakagawa et al. 1976). In order to facilitate the assignments of the gauche- and trans-ethyl mercaptan millimeter-wave spectra, the microwave data from Quade et al. (1975) were used for initial predictions of the rotational transitions. Pickett's spectral fitting programs SPCAT and SPFIT (Pickett 1991) were used to predict the spectra as well as to determine the spectroscopic constants. The visualization, processing, and assignments of the rotational spectra were performed using the SVIEW and ASCP programs included in the AABS package (Kisiel et al. 2005).

Zoom In Zoom Out Reset image size

The rotational spectrum of gauche-ethyl mercaptan is formed by two types of transitions, a- and b-type pure rotational transitions (∣μ_a∣ = 1.48(2) D, ∣μ_b∣ = 0.19(10) D; Quade et al. 1975), between the rotational levels within the same torsion substate, i.e., 0^±↔0^±, and c-type torsion–rotational transitions (∣μ_c∣ = 0.59(2) D; Quade et al. 1975) connecting different torsion substates (0^∓↔0^±). Figure 1(d) illustrates a section of the a-type R-branch pure rotational transitions which are the dominant features of the ethyl mercaptan rotational spectrum. The doubling pattern of the c-type transitions resulting from the tunneling process can be seen in Figure 1(e). The frequency separation between the corresponding c-type doublets is close to 2ΔE, which is about 3.5 GHz (see Figure 1(e)).

Several near degeneracies of the 0⁺ and 0⁻ energy levels beginning at low J values have been observed which require accounting for the Coriolis-like terms that connect the 0⁺ and 0⁻ substates. Perturbation-allowed transitions between the 0⁺ and 0⁻ energy levels with ΔK_a = even and ΔK_c = even selection rules arising from the mixing of the 0⁺ and 0⁻ substates with the same K_a quantum number have been observed as well. Finally, more than 2200 distinct transitions involving the rotational quantum numbers J'' and $K_{a}^{\prime \prime }$ up to 88 and 25, respectively, have been analyzed using the effective framework fixed axes two-state Hamiltonian as defined by Quade et al. (1963):

$$\begin{equation} H_{\mathrm{eff}} = H_{\mathrm{rot}}^{+} + H_{\mathrm{rot}}^{-} + H^{\pm } + \Delta E, \end{equation} \tag{ 1 }$$

where $H_{\mathrm{rot}}^{+}$ and $H_{\mathrm{rot}}^{-}$ represent Watson's A-reduced semi-rigid Hamiltonian in I^r representation (Watson 1977) for the 0⁺ and 0⁻ substates, respectively, and H^± is the torsion–rotation interaction Hamiltonian:

$$\begin{equation} H^{\pm } = D^{\pm }(J_{b}J_{c}+J_{c}J_{b}) + Q^{\pm }J_{a} + N^{\pm }J_{b}, \end{equation} \tag{ 2 }$$

where D^±, Q^±, and N^± represent the Coriolis-like coupling constants. Centrifugal distortion corrections up to the eighth order have been included in the analysis. All the perturbed transitions have been successfully treated by means of the Coriolis-like constants defined in Equation (2), with the K dependence of the D^± constant taken into account. For the gauche-³⁴S isotopologue, more than 700 transitions (J'' and $K_{a}^{\prime \prime }$ up to 56 and 18, respectively) have been analyzed in the same manner as the parent species. The spectroscopic constants determined for both isotopologues are listed in Table 1. The measured frequencies are given in Table 2 (the full Table 2 is given in the online journal).

Table 1. Spectroscopic Constants^a of gauche- and trans-Ethyl Mercaptan (A-Reduction, I^r-Representation)

Constant	Unit	Gauche (Parent Species)		Gauche (34S)		Trans
		0+	0−	0+	0−
A	(MHz)	28747.4104 (66)	28747.2715 (65)	28709.699 (24)	28708.964 (26)	28416.7604 (18)
B	(MHz)	5295.1422 (36)	5295.0008 (36)	5175.5685 (36)	5175.4383 (36)	5485.77901 (15)
C	(MHz)	4845.9421 (36)	4845.9689 (36)	4744.7050 (36)	4744.7310 (36)	4881.81770 (15)
ΔJ	(kHz)	3.326369 (20)	3.323582 (20)	3.18351 (18)	3.18084 (18)	3.83217 (27)
ΔJK	(kHz)	−18.39280 (60)	−18.35859 (60)	−17.9512 (15)	−17.8994 (14)	−22.4549 (58)
ΔK	(kHz)	204.1591 (82)	203.9217 (81)	206.8 (14)	205.1 (14)	210.03 (18)
δJ	(kHz)	0.514429 (12)	0.513170 (14)	0.481132 (82)	0.480129 (71)	0.65664 (11)
δK	(kHz)	8.781 (12)	8.555 (12)	8.271 (34)	8.024 (35)	7.342 (20)
ΦJ	(kHz)	0.0029872 (28)	0.0029225 (31)	0.0029872b	0.0029225b	−0.00086 (17)
ΦJK	(kHz)	0.0790 (13)	0.0661 (13)	0.0790b	0.0661b	0.367 (18)
ΦKJ	(kHz)	−1.3341 (45)	−1.2955 (44)	−1.3341b	−1.2955b	−2.717 (106)
ΦK	(kHz)	6.341 (48)	5.171 (46)	6.341b	5.171b	26.7 (80)
ϕJ	(kHz)	0.0012107 (15)	0.0011805 (16)	0.0012107b	0.0011805b	−0.000935 (91)
ϕJK	(kHz)	0.04561 (63)	0.04326 (65)	0.04561b	0.04326b	−0.164 (26)
ϕK	(kHz)	5.858 (95)	4.825 (97)	5.858b	4.825b	28.89 (97)
LJ	(mHz)	⋅⋅⋅	⋅⋅⋅	⋅⋅⋅	⋅⋅⋅	0.000060 (11)
LKKJ	(mHz)	⋅⋅⋅	⋅⋅⋅	⋅⋅⋅	⋅⋅⋅	2.11 (40)
LK	(mHz)	−1.140 (75)	0.673 (71)	⋅⋅⋅	⋅⋅⋅	1374 (110)
lJK	(mHz)	⋅⋅⋅	⋅⋅⋅	⋅⋅⋅	⋅⋅⋅	0.0497 (61)
lKJ	(mHz)	0.0523 (38)	0.0414 (38)	⋅⋅⋅	⋅⋅⋅	4.77(50)
lK	(mHz)	0.403 (21)	0.601 (19)	⋅⋅⋅	⋅⋅⋅	⋅⋅⋅
ΔE	(MHz)	1753.9788 (37)		1730.995 (19)		⋅⋅⋅
D±	(MHz)	8.359 (96)		5.36 (14)		⋅⋅⋅
Q±	(MHz)	−15.00 (38)		−26.51 (57)		⋅⋅⋅
N±	(MHz)	6.30 (11)		⋅⋅⋅	⋅⋅⋅	⋅⋅⋅
$D^{\pm }_{K}$	(MHz)	0.01358 (50)		⋅⋅⋅	⋅⋅⋅	⋅⋅⋅
σfit	(kHz)	39		41		30

Notes. ^aThe numbers in parentheses are 1σ uncertainties in the units of the last decimal digit. ^bFixed to the parent species value.

Download table as:  ASCIITypeset image

Table 2. Laboratory Measurements and Astronomically Detected Lines of Ethyl Mercaptan

J'	$K_{a}^{\prime }$	$K_{c}^{\prime }$	v'	J''	$K_{a}^{\prime \prime }$	$K_{c}^{\prime \prime }$	v''	Laboratory	Lab.−Pred.	Conf.	Eu	Sij	Sky Obs. Freq.	Tmb	Blend
								Freq. (MHz)	(MHz)		(K)		(MHz)	(K)
13	5	8	1	12	5	7	1	131926.7036	−0.0205	G	72.8	11.1	131928.9	0.08
13	5	9	1	12	5	8	1	131926.7036	0.0443	G	72.8	11.1	†
13	5	8	0	12	5	7	0	131928.2800	−0.0229	G	72.7	11.1	†
13	5	9	0	12	5	8	0	131928.2800	0.0420	G	72.7	11.1	†
13	4	10	1	12	4	9	1	131978.8083	−0.0052	G	62.6	11.8	131977.9	0.07
13	4	10	0	12	4	9	0	131980.4297	−0.0062	G	62.5	11.8	†
13	4	9	1	12	4	8	1	131983.4352	−0.0123	G
13	4	9	0	12	4	8	0	131985.0626	−0.0126	G
13	3	11	1	12	3	10	1	132035.3178	−0.0038	G	54.6	12.3	132033.3	0.03	CH3COOCH3

Notes. This table gives the laboratory measurements and emission lines of gauche- and trans-CH₃CH₂SH present in the Orion KL survey. All the transition frequencies included in the least-squares analysis were weighted inversely proportionally to the second power of their estimated uncertainties. The unresolved pairs of the asymmetry splitting were always included as two overlapped transitions and fitted to their intensity weighted averages. Columns 1–8 indicate the upper and lower state quantum numbers. Column 9 gives the measured frequency at the laboratory. Column 10 gives the difference between frequency measurements and predictions. Column 11 gives the corresponding conformer. Column 12 gives the upper level energy. Column 13 gives the line strength. Column 14 gives the observed frequency in the line survey of Orion assuming a v_LSR of 5 km s⁻¹. Column 15 gives the main beam temperature. Column 16 indicates blended transitions with other molecules. †Blended with previous line.

Only a portion of this table is shown here to demonstrate its form and content. A machine-readable version of the full table is available.

Download table as:  DataTypeset image

The rotational spectrum of trans-ethyl mercaptan is dominated by pure rotational a- and b-type transitions (∣μ_a∣ = 1.06(3) D, ∣μ_b∣ = 1.17(3) D; Quade et al. 1975). An example of measured a-type R-branch transitions is illustrated in Figure 1(c). Many assigned trans transitions with higher values of $K_{a}^{\prime \prime }$ quantum numbers have been found to be affected by perturbations which may originate from interactions with low-lying first excited states of the –SH and –CH₃ torsional modes. Only the unperturbed transitions were included in the final fit. More than 600 distinct transitions with J'' and $K_{a}^{\prime \prime }$ quantum numbers up to 64 and 13, respectively, were analyzed using the standard single-state semi-rigid A-reduced Hamiltonian (Watson 1977) with octic centrifugal distortion constants. The resulting spectroscopic constants are given in Table 1. The measured transitions are given in the full Table 2 (see the online journal).

Five observing sessions (from 2004 September to 2007 January) were required for completing a molecular line survey in all frequencies available with the A, B, C, and D receivers (80–115.5, 130–178, 197–281 GHz) of the IRAM 30 m telescope toward Orion KL (IRc2 source at α(J2000) = 5^h35^m145, δ(J2000) = −5°22'300). System temperatures, image side band rejections, and half power beam widths were in the ranges 100–800 K, 27–13 dB, and 29''–9'', respectively, from lower to higher frequencies. The intensity scale was calibrated using the Atmospheric Transmission Model package (Cernicharo 1985; Pardo et al. 2001). The observations were performed in the balanced wobbler-switching mode. As backends we connected two filter banks (512 MHz of bandwidth for each one and 1 MHz of spectral resolution) and a correlator (2 × 512 MHz of bandwidth and 1.25 MHz of spectral resolution). Pointing and focus were checked every 1–2 hr on nearby quasars. The data were processed with the GILDAS software.⁴ The data reduction consisted of removing lines from the image side band and fitting and removing baselines. Figures are shown in units of the main beam temperature, T_MB = $T^{\star }_A$/η_MB, where η_MB is the main beam efficiency. More detailed description of the observations can be found in Tercero et al. (2010).

At least four cloud components could be identified in the line profiles of our low resolution spectral lines, characterized by different radial velocities and line widths (Blake et al. 1987; Schilke et al. 2001; Persson et al. 2007; Tercero et al. 2010, 2011; Neill et al. 2013). Each component corresponds to a specific region of the cloud that overlaps in our telescope beam: the extended ridge or ambient cloud (T_K ≃ 60 K); the compact ridge, a dense clump characterized by the emission of organic saturated O-rich molecules at a T_K ≃ 150 K; the plateau, or outflow from the newborn stars (T_K ≃ 150 K); and the hot core, a dense and warm region rich in organic saturated N-bearing molecules (T_K ≃ 250 K).

Direct laboratory measurements and derived spectroscopic constants given in Table 1 have allowed us to detect gauche-ethyl mercaptan and to tentatively detect the trans conformer in the molecular line survey of Orion KL by means of a large number of spectral lines free of blending with other species. Table 2 provides (together with the transitions measured in the laboratory) the observational parameters of the detected lines that are not strongly blended with other molecules. Owing to the weakness of these features, the main beam temperature has been obtained from the peak channel of the spectra, so errors in the baselines and contribution from other species could affect this parameter. Therefore, T_MB has to be considered as the total intensity of the detected feature and an upper limit to the intensity of ethyl mercaptan in this study. A total of 34 transitions free of blending (in the peak channel of the feature) and 43 slightly blended lines are reported in Table 2 for gauche-ethyl mercaptan. No missing lines have been found for this conformer. Figure 2 shows some of the lines of gauche-ethyl mercaptan reported here. Even though the gauche conformer is more stable, the number of potential lines to be detected for

Zoom In Zoom Out Reset image size

the trans conformer could be larger due to the higher value of its b-dipole moment component (see above). For trans-ethyl mercaptan, 72 lines free of blending and 52 slightly blended lines could be present in the survey. They are listed in Table 2 with the estimate of their parameters. However, the weakness of the transitions of trans-CH₃CH₂SH and the high overlap with other molecules make the assignment of isolated lines of this conformer difficult as they appear at the confusion limit of the line survey. Hence, we consider only a tentative detection for the trans conformer of ethyl mercaptan (see below). We also searched for CH₃CH₂SH in the PRIMOS survey⁵ with the Green Bank Telescope in the frequency range between 300 MHz and 50 GHz and in the >13σ U lines of the 3 mm IRAM 30 m survey reported by Belloche et al. (2013), both surveys toward SgrB2, finding a negative detection in both sets of data.

In order to compare the emission of ethyl mercaptan with that of methyl mercaptan, the latter species was searched for in Orion KL. The laboratory frequencies and dipole moment values used for the methyl mercaptan are those predicted by Bettens et al. (1999) and have been implemented in the MADEX code (Cernicharo 2012). Figures 2 and 3 show selected lines of gauche-CH₃CH₂SH and A/E-CH₃SH, respectively. Both figures show our best model for these species (see below) and the total model for the species already studied in this survey (see Tercero et al. 2013, and references therein). The figures show lines free of blending, or moderately blended with lines from other species. No missing lines have been found for methyl mercaptan in the frequency range covered by our line survey.

Zoom In Zoom Out Reset image size

To model the emission of A/E-CH₃SH and gauche/trans-CH₃CH₂SH we have considered that both molecules come from the same region of Orion KL (hot core at v_LSR = 5 km s⁻¹ and Δv = 7 km s⁻¹). Lines from methyl mercaptan are typically 10 times stronger than those of ethyl mercaptan. We used the MADEX code in LTE conditions due to the lack of collisional rates for these species. Beam dilution and the position of the source with respect to the pointing position were taken into account in our models. We assumed d_sou = 10''and offset = 3''. In Tercero et al. (2010) we estimated the uncertainties of the column density results in this survey to be between 20%–30% considering different sources of uncertainty such us the spatial overlap of the different cloud components, the modest angular resolution of any single-dish line survey or pointing errors. Nevertheless, the uncertainty due to high overlap problems has to be considered for results obtained by means of weak lines such as those of ethyl mercaptan (raising the uncertainty up to 50%). The physical and chemical parameters derived by the model are a common kinetic temperature of 200 ± 50 K for ethyl and methyl mercaptan and a column density of (5.0 ± 2.0) × 10¹⁵ cm⁻² for each A and E state of CH₃SH and of (2.0 ± 1.0) × 10¹⁵ cm⁻² for gauche-ethyl mercaptan. For trans-CH₃CH₂SH we can report only a tentative detection with an upper limit to its column density of ⩽(2.0 ± 1.0) × 10¹⁵ cm⁻². Therefore, we found that methyl mercaptan is ≃5 times more abundant than ethyl mercaptan in the hot core of Orion KL. The population of different conformers of CH₃CH₂SH follows a Boltzmann distribution so the abundance ratio between both conformers of ethyl mercaptan is given by N(trans)/N(gauche) = exp(−E_rel/T_k), where E_rel is the difference in the energy between the conformers (230 K; M. L. Senent 2013, private communication) and T_k is the kinetic temperature of the medium. At 200 K this formula yields a N(trans)/N(gauche) value of 0.32 whereas at 300 K this number rises up to 0.46. These abundance ratios have to be corrected by the vibrational partition function of both conformers. For T_K = 200 K the vibrational partition function for gauche is 1.47 and for trans it is 1.54 (M. L. Senent 2013, private communication). Applying these corrections, our abundance ratio between trans and gauche (both states 0⁺ and 0⁻) is ⩽1.

In order to compare these results with structurally similar molecules, we modeled the emission of ethanol (trans/gauche-CH₃CH₂OH) and methanol (A/E-¹³CH₃OH) in our line survey using the MADEX code and LTE conditions (A. López et al., in preparation). To fit the line profiles in this large spectral range, five cloud components are required: the four components described in Section 4 and a hotter (300 K) and smaller (d_sou = 7'') compact ridge. For the four former components, the assumed physical conditions are those derived in Tercero et al. (2010). We obtained column densities of (6 ± 2) × 10¹⁷ cm⁻², (2.2 ± 0.6) × 10¹⁶ cm⁻², and (1.7 ± 0.4) × 10¹⁶ cm⁻² for each state of CH₃OH (assuming a ¹²C/¹³C ratio of 45; see Tercero et al. 2010), for the trans conformer of ethanol, and for gauche-CH₃CH₂OH, respectively. Therefore, methanol is 30 times more abundant than ethanol in Orion KL. This difference between X(CH₃OH/CH₃CH₂OH) and X(CH₃SH/CH₃CH₂SH) could be due to the methanol and methyl mercaptan stay time on the dust grains before evaporation. If methyl mercaptan stays longer than methanol on the dust grains, then, further chemical processing could occur, changing the abundance ratios between methyl and ethyl species. Taking into account that the emission of the −OH species comes mainly from the compact ridge whereas methyl and ethyl mercaptan emit from the hot core, the differences of relative abundances between the methyl and ethyl species could be also due to the chemical differentiation between these two regions inside Orion KL. More sensitive observations, with interferometers such as ALMA, are needed to derive accurate column densities of the two conformers of ethyl mercaptan and to study their spatial distribution in Orion.

The authors thank the Spanish MINECO for support under grants AYA 2009-07304, AYA2011-29375, CTQ 2006-05981/BQU, CTQ 2010-19008, and the CONSOLIDER program "ASTROMOL" CSD 2009-00038 and Junta de Castilla y Leon (Grant VA070A08). B.P.G. and S.T.S. acknowledge support from the National Science Foundation Division of Chemistry under grant No. 1111101 (co-funded by MPS/CHE and the Division of Astronomical Sciences) and B.P.G. also thanks New College of Florida for a Student Research and Travel Grant award.