A Q-band Line Survey toward Orion KL Using the Tianma Radio Telescope

The emission of RRLs was modeled through adding up the spectra of all of the RRL transitions of H, He, and C within the frequency range of this survey. For species X (denoting H and He), its RRL spectrum can be calculated using (Gordon & Sorochenko 2002)

$$\begin{eqnarray}&&{\tau }_{{n}_{1},{n}_{2}}=3.867\times {10}^{-12}\displaystyle \frac{{b}_{{n}_{2}}}{{\rm{\Delta }}\nu }\displaystyle \frac{{\rm{\Delta }}n}{{n}_{1}}{f}_{{n}_{1},{n}_{2}}\left(1-\displaystyle \frac{3{\rm{\Delta }}n}{2{n}_{1}}\right)\displaystyle \frac{\mathrm{EM}}{{T}_{e}^{5/2}},\end{eqnarray} \tag{ 4 }$$

and

$$\begin{eqnarray}&&{T}_{{\rm{X}}}=\sum _{{n}_{1},{n}_{2}}{T}_{e}{\tau }_{{n}_{1},{n}_{2}}\displaystyle \frac{{n}_{X}}{{n}_{e}}\exp \left(-\displaystyle \frac{{\left[\nu -{\nu }_{\mathrm{rest}}^{\mathrm{RRL}}({n}_{1},{n}_{2})(1-\tfrac{\upsilon }{c})\right]}^{2}}{2{\rm{\Delta }}{\nu }^{2}/(8\mathrm{ln}(2))}\right).\end{eqnarray} \tag{ 5 }$$

Here, the parameters are adopted as the values in cgs units, EM is the emission measure, n_X is the volume density of X, υ is the velocity, ${b}_{{n}_{2}}$ is the upper-level departure coefficient, and T is the excitation temperature. The optically thin limit (${\tau }_{{n}_{1},{n}_{2}}\ll 1$) is satisfied for all of the detected RRLs. The term of $\exp ({E}_{{n}_{2}}/{kT})$ is omitted in Equation (4) since ${E}_{{n}_{2}}\ll {kT}$ for the detected RRLs. The oscillator strength ${f}_{{n}_{1},{n}_{2}}$ can be approximated as (Menzel 1968)

$$\begin{eqnarray}&&{f}_{{n}_{1},{n}_{2}}\sim {n}_{1}{M}_{{\rm{\Delta }}n}\left(1+1.5\displaystyle \frac{{\rm{\Delta }}n}{{n}_{1}}\right)\end{eqnarray} \tag{ 6 }$$

for high-level transitions of hydrogen-like atoms. We empirically fitted the values of M_Δn tabulated in Menzel (1968) as

$$\begin{eqnarray}&&{M}_{{\rm{\Delta }}n}\sim 0.1905{\left(1/{\rm{\Delta }}n\right)}^{2.887}.\end{eqnarray} \tag{ 7 }$$

The oscillator strength can be calculated more accurately from the method described in Brocklehurst (1971) and Regemorter & Prudhomme (1979). The accurate values have been used in recent models of Prozesky & Smits (2018) and Zhu et al. (2019, 2022). The modeled spectrum in this work will not change obviously when the more accurate values of ${f}_{{n}_{1},{n}_{2}}$ are adopted.

The electron temperature (T_e ) of H ii regions are assumed as 8000 K (e.g., Wilson et al. 1997; Zuckerman et al. 1967). If the RRL emitters are in LTE state, the departure coefficients should be unity for all levels. For n_e = 10⁴ cm⁻³ and T = 8000 K, we calculated the population of hydrogen using the method by Zhu et al. (2019, 2022). The departure coefficients b_n increase from 0.85–0.99 as n increases from 50–100. It means that the departure coefficients in Equation (6) cannot be omitted to precisely reproducing the observed RRLs and calculating the EM. The RRLs of H and He from Hii regions under strong ultraviolet fields originate from similar emission regions. The states of helium atoms considered in this work are approximately hydrogenic in nature with n ≫ 1. Storey & Hummer (1995) suggested that the departure coefficients for hydrogen atoms can also be applied to these states of helium atoms.

At large n, the carbon atom is physically similar to H and He; thus, the above equations for H and He are also valid for deriving the emission of carbon RRLs (Salgado et al. 2017). For carbon in PDRs, an electron temperature of 2000 K is adopted considering $({E}_{\mathrm{ionize}}^{{\rm{H}}}-{E}_{\mathrm{ionize}}^{{\rm{C}}})/k\lt 3000$ K. We also try to fit the carbon lines using a T_e of 300 K (Section 4.1.2). We assume that the carbon atoms are in an LTE state with the departure coefficients of carbon adopted as unity, although this is somewhat uncertain.

To determine the values of line widths and velocities of the modeled RRLs, we have referred to the Gaussian fitting results of the detected RRLs (see Table 3 and Section 5.1) and the values of previous studies (Goddi et al. 2009; Gong et al. 2015; Rizzo et al. 2017). The Δν is the FWHM line width of the modeled spectrum in frequency scale, which can be calculated through

$$\begin{eqnarray}&&{\rm{\Delta }}\nu ={f}_{{n}_{1},{n}_{2}}^{\mathrm{RRL}}\displaystyle \frac{{\rm{\Delta }}\upsilon }{c}.\end{eqnarray} \tag{ 8 }$$

Here, the Δυ is fixed as 20 km s⁻¹ for H/He RRLs and 5 km s⁻¹ for C RRLs. The velocity (in V_LSR) of H/He and C is fixed as −3 km s⁻¹ and 9 km s⁻¹, respectively (see Section 5.1.2 for further discussion).

The beam filling factor of H/He RRLs is adopted as unity, since they mainly originate from the extended M42 Hii region. The EM is fitted as 3.9 × 10⁶ cm⁻⁶ pc. The abundance ratio between He and H (He/H) is fitted to 8.5%. We note that this value is derived with an assumption that the H and He have identical ionized regions. We will further discuss the He/H ratio in Section 5.1.2. The fitted intensities of unblended H/He RRLs are all consistent with the observed values within 20%. The fitted emission measure of Hii region (EM^HII) is compatible with the value of 7.5 × 10⁶ cm⁻⁶ pc derived by Dicker et al. (2009) using the 21.5 GHz data of the GBT with a beam size of 335.

The carbon RRLs mainly originate from PDR between the M42 and molecular clouds. Since the main body of the Orion bar is larger than the beam of this survey, a beam filling factor of unity is also adopted for carbon RRLs. If the ionized carbons contribute to most of the electrons within the PDR region, the EM of the PDR can be fitted to EM^PDR = 3.9 × 10³ cm⁻⁶ pc. If an electron temperature of 300 K (instead of 2000 K) is adopted, the fitting result gives EM^PDR = 1.8 × 10² cm⁻⁶ pc.

The molecular emission was modeled by adding up a set of spectral components contributed by different species. A species (denoted as X) may contribute multiple spectral components with different velocities. A spectral component may consist of more than one line feature contributed by different transitions of its corresponding species. Each spectral component is assumed to be in an LTE state with an identical emission source size (D) and the same excitation temperature (T_ex) for all transitions of the corresponding species. For each spectral component (denoted as s), its spectrum (T_s in main beam temperature scale) could be modeled through adding up a set of emission lines of the transitions of the corresponding species. Those lines have Gaussian-shaped optical depths with the same line width and T_ex. Specifically, a spectral component can be calculated following (e.g., Mangum & Shirley 2015)

$$\begin{eqnarray}&&{\tau }_{{ij}}^{s}=\displaystyle \frac{{A}_{{ij}}{c}^{2}}{8\pi {\left({\nu }_{{ij}}\right)}^{2}}\left[\exp \left(\displaystyle \frac{h{\nu }_{{ij}}}{{{kT}}_{\mathrm{ex}}}\right)-1\right]{N}_{\mathrm{tot}}^{s}\displaystyle \frac{{g}_{u}}{Q}\exp \left(\displaystyle \frac{-{E}_{u}}{{{kT}}_{{ex}}}\right)\displaystyle \frac{1.06}{{\rm{\Delta }}{\nu }_{s}}\end{eqnarray} \tag{ 9 }$$

and

$$\begin{eqnarray}\begin{array}{ll}{T}_{{\rm{s}}} & =\,\displaystyle \sum _{i,j}{{fT}}_{{ex}}\left\{1-\exp \left[-{\tau }_{i,j}^{s}\exp \left(-\displaystyle \frac{{\left[\nu -{\nu }_{i,j}(1-\tfrac{{\upsilon }_{s}}{c})\right]}^{2}}{2{\rm{\Delta }}{\nu }_{s}^{2}/(8\mathrm{ln}(2))}\right)\right]\right\}.\end{array}\end{eqnarray} \tag{ 10 }$$

Here, ${\tau }_{{ij}}^{s}$ is the peak optical depth, ${N}_{\mathrm{tot}}^{s}$ is the column density of the corresponding species, Δv_s is the line width in frequency scale (Δν = ν_i,j ΔV_s /c), f is the beam filling factor ${(D/{{\rm{\Theta }}}_{{\nu }_{i,j}})}^{2}$, ${{\rm{\Theta }}}_{{\nu }_{i,j}}$ is the beam size at ν_i,j , E_u is the upper-level energy, g_u is the degree of degeneracy of the upper level, and Q is the partition function. The Q is a function of T_ex, and it could be interpolated from the tabulated values quoted from online databases (Section 3.3). The modeled spectrum of a species is

$$\begin{eqnarray}&&{T}_{X}=\sum _{s}{T}_{s}.\end{eqnarray} \tag{ 11 }$$

For each species, we fit T_X to the observed spectrum. Since we only tried to build a radiative transfer model with the goal to reproduce our observed spectrum, source size and T_ex for each spectral component are fixed. The values of line widths (ΔV_s ), velocity (υ_s), and column density (${N}_{\mathrm{tot}}^{s}$) are fitted.

The parameters for different species were basically independent. However, to reproduce the profiles of emission lines of isotopologues or species with similar chemical characteristics, we tend to adopt similar fixed values for their D and T_ex, and use a similar initial guess for their fitted parameters. For most species whose transitions have similar upper-level energies, the fitting result is not sensitive to T_ex. For those species, T_ex is manually adopted as 50, 100, 150, 200, or 300 K. We have also referred to the literature (Rizzo et al. 2017; Tercero et al. 2010) to guess the initial values of the fitted parameters. We try to model the spectrum with as few spectral components as possible. After each iteration of model fitting, we manually checked the fitting result to see whether there are line peaks, skewed line shoulders, or line wings that are obviously not well fitted. If so, we added a spectral component to fit them.

The spectra of different species are fitted separately. The species with strong lines are first fitted. For each of the rest species, the fitting procedure is conducted on the residual spectrum (the difference between the original spectrum and the modeled spectrum of all other fitted species). This procedure is iterated until all of the species are well fitted.

All of the identified species except SiO and its isotopologues were modeled. The transition parameters of C₂H₅CN υ13/υ21 are not yet publicly available from the databases of JPL and CDMS, and we obtained those parameters from Endres et al. (2021) through private communication. We will discuss the fitting of C₂H₅CN υ13/υ21 in Section 5.2.4.

The spectrum of E-CH₃OH υ_t = 1 cannot be well fitted with a T_ex of 100 K. Adopting T_ex as 110 K, the temperature of the compact ridge (Tercero et al. 2010), the spectrum of E-CH₃OH υ_t = 1 can be better reproduced. It is consistent with the result of its rotational diagram (Figure 7), which yields a T_ex of 115 ± 5 K for E-CH₃OH υ_t = 1. We adopted T_ex as 110 K to fit the corresponding spectral components of E-CH₃OH υ_t = 1, A-CH₃OH υ_t = 1, CH₃OH, and ¹³CH₃OH. The spectra of CH₃OCHO, HNCO, C₂H₃CN, and C₂H₅CN can be well fitted with the spectral components quoted from Rizzo et al. (2017) with T_ex unaltered. For CH₃OCHO and HNCO, the deviation between the fitted column densities of this survey and those of Rizzo et al. (2017) are smaller than 50%. However, for C₂H₃CN and C₂H₅CN, the fitted column densities of this survey are 3–10 times lower than the values of Rizzo et al. (2017). Only one line of CH₃CN is covered, and it is strong with a highly non-Gaussian shape. The fitting results of CH₃CN should be treated as rough estimations with large uncertainties.

Zoom In Zoom Out Reset image size

The fitting results of the modeled parameters are listed in Table 2. For each species, we extracted the spectral segments that may contain emission of that species. Those spectral segments were then spliced together (referred to as the "spliced spectrum"). The spliced spectra for all fitted species are shown in Figures 8 and 9. For the transitions of most species, the deviations between the modeled intensities and the observed values are smaller than 10%. The spliced spectrum of NH₂D is an exception. NH₂D has two detected transitions (43,042 and 49,962 MHz), and those two lines cannot be simultaneously well reproduced.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The spectral components are divided into three groups in the velocity and line width (V_LSR–ΔV) space using the K-means algorithm (Lloyd 1982) implemented in SciPy, ¹⁶ as shown in Figure 10. The three groups are associated with three gas components of Orion KL: the plateau, the hot core, and the extended/compact ridge (Tercero et al. 2010; Rizzo et al. 2017). The median values of V_LSR of the three groups are 7 km s⁻¹, 6 km s⁻¹, and 8.5 km s⁻¹, respectively. The median values of ΔV of the three groups are 25 km s⁻¹, 8 km s⁻¹, and 3 km s⁻¹, respectively. Roughly, the line widths of spectral components in groups of the extended/compact ridge, hot core, and plateau are in the range of 1–5 km s⁻¹, 5–15 km s⁻¹, and 15–30 km s⁻¹, respectively. The kinetic temperature of the compact ridge, ∼100 K, is higher than the value of the extended ridge, ∼60 K (Tercero et al. 2010). However, most of the T_ex listed in Table 2 have quite large uncertainties, and thus it is difficult to explicitly assign a spectral component with ΔV < 5 km s⁻¹ to the compact ridge or the extended ridge.

Zoom In Zoom Out Reset image size

For oxygen-bearing species, most of their spectral components (29/49) have ΔV < =5 km s⁻¹ (Figure 10). It is expected since most of the oxygen-bearing species have spectral components originating from the extended/compact ridge. Thirty-eight percent (10/28) of the spectral components of sulfur-bearing species have ΔV > 15 km s⁻¹. In contrast, for oxygen-bearing species, only 24% (12/49) of the spectral components have ΔV > 15 km s⁻¹. Aldehyde-containing species and species with both sulfur and oxygen elements tend to have spectral components originating from the plateau (Table 2), and this supports a possible enhancement of these species by shocks (e.g., Feng et al. 2015; Lefloch et al. 2017; Liu et al. 2021).

Among the unblended hydrogen RRLs detected in this survey, H135π has the maximum Δn of 16. The maximum Δn of this survey is larger than the values of the Q-band survey of Rizzo et al. (2017) and the radio K-band (1.3 cm) survey of Gong et al. (2015). The maximum Δn detected by Rizzo et al. (2017) and Gong et al. (2015) was 11 for both. For spectrally resolved helium RRLs in this survey, the maximum Δn is 7, and it is also higher than the values of 4 detected by Rizzo et al. (2017) and Gong et al. (2015). The carbon RRLs detected by Gong et al. (2015) are all carbon α lines highly blended with helium RRLs. Only two carbon RRLs (C52α and C53α) were firmly detected and resolved by Rizzo et al. (2017). On the contrary, 10 RRLs of carbon are resolved in this survey, including the C81γ with a Δn of 3 (Figure 12). Thanks to the higher sensitivity and wider frequency coverage of this survey, this work has more than doubled the number of detected RRLs, particularly those with larger Δn, in the Q band compared with Rizzo et al. (2017).

Zoom In Zoom Out Reset image size

Benefiting from a large number of RRLs detected in this survey, we can estimate the abundance ratio of He/H via the intensity ratio between their RRLs with the same n and Δn:

$$\begin{eqnarray}&&y=\displaystyle \frac{N(\mathrm{He})}{N({\rm{H}})}\sim \displaystyle \frac{1}{{R}^{1/3}}\times \displaystyle \frac{\int {T}_{\nu }(\mathrm{He})d\nu }{\int {T}_{\nu }({\rm{H}})d\nu }.\end{eqnarray} \tag{ 12 }$$

Here, ∫T_ν (X)d ν is the integrated intensity of the RRL of X, and the R is the volume ratio between the ionized regions of He and H. If the R value is adopted as unity considering the similarity between the line widths of hydrogen and helium RRLs (Table 3), the inferred mean value of y is 8.4% with a standard deviation of 0.2% (Figure 13).

Zoom In Zoom Out Reset image size

The y-value derived here is consistent with the result of global fitting described in Section 4.1.2. It is also consistent with the values of (8.7 ± 0.7)% derived by Gong et al. (2015), (8.3 ± 1.2)% by Rizzo et al. (2017), and (8.8 ± 0.6)% by Baldwin et al. (1991) derived based on optical observations. The value of y seems to be compatible with the value of 8.3% due to Big Bang nucleosynthesis (Olive & Skillman 2004) and lower than the solar value, 9.8% (Wilson & Rood 1994). However, if an R value of 0.6 is adopted (Copetti & Bica 1983), the derived value of y is 10%, much closer to the solar value.

The velocities of unblended hydrogen, helium, and carbon RRLs are shown in the lower panel of Figure 13. The velocities of hydrogen and helium RRLs at different frequencies are consistent with a value of −3 ± 1.5 km s⁻¹. The velocities of carbon RRLs are ∼+9 ± 1 km s⁻¹. The computed velocities are compatible with previous results that the hydrogen and helium RRLs arise from the M42, while the carbon RRLs originate from the PDR between M42 and the associated molecular clouds (the Orion Bar; see Cuadrado et al. 2015; Gong et al. 2015; Rizzo et al. 2017).

Sulfur-bearing molecules (e.g., H₂S, SO, SO₂, CS, OCS, and H₂CS) have been detected in various star-forming environments (e.g., Turner et al. 1973; Luo et al. 2019; Wu et al. 2019b). Thirteen kinds of isotopologues of sulfur-bearing molecules are detected in this survey. Among them, five species were not confirmed by Rizzo et al. (2017).

The comparison between the HCS⁺ spectra of this survey and that of Rizzo et al. (2017) is shown in Figure 14. The narrow emission line of HCS⁺ is overlapped with the broad emission of C₂H₅CN. Although Rizzo et al. (2017) marked this line as a blended line of HCS⁺ and C₂H₅CN, the broad emission of C₂H₅CN cannot be seen in their spectrum. This further reduces the degree of confidence of the possible HCS⁺ emission feature with a limited signal-to-noise ratio (S/N). In the spectrum of this survey, the emission of HCS⁺ and C₂H₅CN can be spectrally resolved. Our survey has higher sensitivity and smaller beam size, and this may make our survey more sensitive to emission of compact sources such as the emission regions of COMs.

Zoom In Zoom Out Reset image size

Two transitions of H₂CS were detected in this survey. The line width of H₂CS is consistent with the value of HCS⁺ (see Figure 9 and Table 2). This implies that H₂CS and HCS⁺ may originate from similar emission regions and have tight chemical correlation, through reactions such as (McElroy et al. 2013)

$$\begin{eqnarray}&&\begin{array}{l}\begin{array}{l}{\mathrm{CH}}_{2}^{+}+\mathrm{OCS}\to {{\rm{H}}}_{2}{\mathrm{CS}}^{+}+\mathrm{CO}\\ {\mathrm{CH}}_{2}^{+}+\mathrm{OCS}\to {\mathrm{HCS}}^{+}+\mathrm{HCO}\end{array}.\end{array}\end{eqnarray} \tag{ 13 }$$

We note that the emission feature identified as the J = 8–7 of C₃S by Rizzo et al. (2017) cannot be seen in the spectrum of this survey (Figure 14). There is also no emission feature at C₃S J = 7–6 (40,465 MHz). Since J = 8–7 is the only line of C₃S detected by Rizzo et al. (2017) with limited S/N, the contribution of noise cannot be ignored. The column density of C₃S may be much lower than the value estimated by Rizzo et al. (2017). However, since our observation has a smaller beam size, it cannot be excluded that the C₃S emission detected by Rizzo et al. (2017) originates from regions uncovered by the beam of this survey. The CCS emission identified by Rizzo et al. (2017) is also not detected or only marginally detected in this survey (Figure 9).

One transition of O¹³CS is observed by Rizzo et al. (2017) and two by this survey. The emission of O¹³CS is marginally detected in this survey and Rizzo et al. (2017). For our survey, since there are weak emission features at both of its two transitions, O¹³CS is marked as a detected species.

In total, 15 kinds of oxygen-bearing organic species are detected in this survey. We use three velocity components to fit the emission of CH₃OH, including two narrow components with ΔV = 4 km s⁻¹ and a broad component with ΔV = 25 km s⁻¹ (Table 2). The wide wings of ¹³CH₃OH have not been obviously detected; thus, the broad component (ΔV = 30 km s⁻¹) has not been reused from CH₃OH to fit the spectrum of ¹³CH₃OH. The spectrum of ¹³CH₃OH can be well modeled when adopting the column densities of ¹³CH₃OH as 0.01 times the values of the two narrow components of CH₃OH (see Figure 8). If the CH₃OH and ¹³CH₃OH have identical emission regions, it implies that the abundance ratio between ¹³CH₃OH and CH₃OH can be estimated as 0.01. For E-CH₃OH υ_t = 1 and A-CH₃OH υ_t = 1, the model parameter of CH₃OH can also be applied to them without any modification to reproduce their spectra, assuming that the A-type and E-type of CH₃OH have equal abundances. C₂H₅OH is also detected in this survey. The model fitting gives an abundance ratio between C₂H₅OH and CH₃OH of 0.003.

Apart from H₂CO and CH₃OCHO, which have been detected by Rizzo et al. (2017), the molecules containing an aldehyde group detected in this survey include ${{\rm{H}}}_{2}^{13}$ CO, H₂CCO, and CH₃CHO. The abundance ratio between ${{\rm{H}}}_{2}^{13}$ CO and H₂CO is estimated as 0.02, and the value between CH₃CHO and H₂CO is 0.06. The first vibrationally excited state of CH₃OCHO (υ_t = 1) was detected by Rizzo et al. (2017) through stacking faint lines; most of them are near the detection limit. In our survey, more than 10 lines of CH₃OCHO υ_t = 1 have been detected with significant S/Ns (Figures 15 and 9).

Zoom In Zoom Out Reset image size

Three lines of HCOOH (formic acid) are detected in this survey, with a velocity of ∼7.5 km s⁻¹. Liu et al. (2002) observed the HCOOH lines at 225 GHz and 262 GHz using the Berkeley–Illinois–Maryland Association array, and the line emission was found to peak in the velocity range of 6.9–8.4 km s⁻¹. Near the compact ridge, the HCOOH emission was spatially resolved by Liu et al. (2002), showing a partial shell morphology. If an emission size of 10'' and an excitation temperature of 60 K is adopted, the fitting of Q-band lines of this survey gives an HCOOH column density of 3.0 × 10¹⁴ cm⁻² (Table 2). This is consistent with the value of 4.8 × 10¹⁴ cm⁻² derived by Liu et al. (2002) adopting an excitation temperature of 62 K.

Interstellar acetone (CH₃COCH₃) is a COM that has been previously detected in several hot cores and low-mass protostars (e.g., Combes et al. 1987; Jørgensen et al. 2011; Friedel & Widicus Weaver 2012; Zou & Widicus Weaver2017). Toward the Orion KL, the transitions of CH₃COCH₃ were not detected or only marginally detected by Rizzo et al. (2017). Six CH₃COCH₃ lines were identified by Goddi et al. (2009). In this survey, more than 10 weak emission features were assigned to transitions of CH₃COCH₃ (see Table 3). However, most of them are marginally detected. We stacked the lines of CH₃COCH₃ to justify the detection of CH₃COCH₃ in Section 5.2.6.

Cyanopolyynes (HC_2n+1N) are chemically young species, and they are usually detected in cold cores, cold/warm ambient gas of star-forming regions, and the envelopes of late-type carbon stars (e.g., Dickens et al. 2001; Esplugues et al. 2013a; Agúndez et al. 2017; Liu et al. 2021). The vibrational levels of HC₃N are mainly excited by mid-IR radiation (de Vicente et al.2000), and it also makes HC₃N a good tracer of hot cores (Turner 1971; Liu et al. 2020).

For the isotopologues of HC₃N (including H¹³CCCN, HC¹³CCN, and HCC¹³CN), the observations of Rizzo et al. (2017) only covered their J = 5–4 transitions. This survey covers the transitions of both J = 4–3 and J = 5–4. Thus, we can better constrain the emission model (Section 4.2) of HC₃N and its isotopologues. The modeled emission of H¹³CCCN, HC¹³CCN, and HCC¹³CN with identical column densities and excitations approximately reproduces the observations for all of their detected transitions (Figure 9). We speculate that the ion-molecule reactions are preferred to produce HC₃N in Orion KL, since they may not lead to significant divergence of abundance among different isotopologues of HC₃N as compared with the neutral-neutral reactions (Taniguchi et al. 2016).

Only two vibrationally excited HC₃N transitions were firmly detected in Rizzo et al. (2017). In this survey, at least two unblended transitions are detected for each of the three vibrational levels (υ₆ = 1, υ₇ = 1, and υ₇ = 2) of HC₃N. Five transitions of HC₅N are also detected in this survey.

Ten lines of C₂H₅CN υ13/υ21 are detected in this survey. Among those lines, nine are spectrally resolvable and they are marked by red strips in Figures 5 and 4. Those lines are all undetected or only marginally detected (with S/N < 3) by Rizzo et al. (2017). To identify those lines, we privately communicated with the authors of Endres et al. (2021) to obtain the transition parameters (rest frequency, E_u, A_i,j , and g_u ) of transitions of C₂H₅CN υ13/υ21 within 34.8–50 GHz. Emission lines corresponding to those transitions are clearly detected in our survey (Figure 15). We modeled them using the method described in Section 4.2. The modeled spectrum is consistent with that observed in this survey, as shown in the upper panel of Figure 15. A T_ex of 50 K is assumed to model the spectrum of C₂H₅CN υ13/υ21, although T_ex cannot be well constrained since all of the detected transitions have similar E_u. The line width is fitted as 6.5 km s⁻¹. The column density of C₂H₅CN υ13/υ21 is undecided since we do not know the partition function.

The in-plane bending vibration υ13 = 1 and torsional state υ21 = 1 of C₂H₅CN (Daly et al. 2013) have been detected toward Sgr B2(N-LMH) by Mehringer et al. (2004) with rest frequencies ranging from 100–270 GHz. C₂H₅CN in vibrationally excited states has also been identified in Orion KL by Daly et al. (2013) from the data of an IRAM millimeter survey (Tercero et al. 2011). Rizzo et al. (2017) tried to search for signals of those transitions in the Q-band spectrum of their survey, but they failed to find signals with intensities higher than 3σ. The vibrationally excited states (υ13/υ21) of C₂H₅CN in the Q band are for the first time firmly detected in this survey.

HCN in the vibrational state (υ₂ = 1) is detected in their direct l-type lines in this survey (Figure 9). HCN υ₂ = 1 is a vibrationally bending mode with a ground state energy of 1024.4 K (Radford & Kurtz 1970; Rolffs et al. 2011; Bruderer et al. 2015). If HCN in the vibrational state is thermally excited, an HCN column density of 1.8 × 10¹⁷ cm⁻² is derived with an assumed excitation temperature of 200 K and an emission size of 10''. It is consistent with near-infrared observations toward the Orion KL using the Short Wavelength Spectrometer of Infrared Space Observatory (Boonman et al. 2003), where an HCN column density of ∼5 × 10¹⁶ cm⁻² was derived from the vibrational emission band of HCN at 14.05 μm, adopting a Doppler line width of 5 km s⁻¹ and an excitation temperature of 275 K.

Three transitions of NH₃ are detected, including the inversion lines of (14,14), (15,15), and (16,16). The (14,14) is highly blended with the hydrogen RRLs and the emission of CH₃OCH₃. The (15,15) is lightly blended with the emission of E-CH₃OCHO, which has a much narrower line width and contributes no more than 10% intensities. The (16,16) is close to a line feature of HC₅N, and it is easy to separate them through Gaussian fitting. The emission of NH₃ lines is modeled by two components with line widths of 8 and 30 km s⁻¹. The emission of NH₃ is necessary to reproduce the observed spectrum at the frequencies of the three transitions. Thus, we consider NH₃ as a detected species. The narrower component has a line width consistent with the value for NH₂D (8 km s⁻¹). It is also consistent with the result of Wilson et al. (1993), who detected inversion lines of NH₃ up to (14,14) with a line width of ∼10 km s⁻¹. The detection of NH₃ (15,15) and (16,16) in this survey pushes the upper-level energy of NH₃ emission lines detected toward Orion KL upward to >2000 K.

We stacked the emission of CH₃COCH₃ to improve the S/N and to consolidate the detection of CH₃COCH₃ of this survey. To avoid bias, all of the unblended transitions of CH₃COCH₃ (including both the detected ones and undetected ones) were stacked. We first calculate the relative intensity (r_i ) of the i_th transition through Equation (10) assuming a T_ex of 100 K. The r_i is then normalized to have a maximum value of 1. Then, the spectrum of each segment was rescaled through multiplying a factor of 1/r_i to make them have equal expected intensities. Obviously, the noise has also been amplified by a factor of 1/r_i , simultaneously. The rescaled spectra are then averaged to get the stacked spectrum, with a weight of r_i ² for each. These procedures are equivalent to directly stacking up the original spectra weighted by ${r}_{i}/{\sum }_{1}^{N}{r}_{i}^{2}$. Here, N is the number of stacked transitions. We use ${\sum }_{i}^{N}{r}_{i}^{2}$ to represent the effective number of stacked transitions (N_eff). If N_eff is small, bias arising from possible noise spikes at a few transitions may be unignorable. The N_eff is 11 and 9.5 for CH₃COCH₃ of this survey and of Rizzo et al. (2017), respectively. Thus, the stacking procedure is valid for CH₃COCH₃. The stacked line of CH₃COCH₃ of our survey shows an obvious emission feature (Figure 16). In contrast, the stacked line of CH₃COCH₃ of Rizzo et al. (2017) only has a weak emission feature lower than 3σ. We also stack a group of spectral segments, with each of the N segments having a rest frequency randomly chosen within the unblended frequency ranges. The randomly stacked spectrum shows no emission feature as expected. Thus, the weak emission features we assign to CH₃COCH₃ should be real, and CH₃COCH₃ is considered a detected species.

Zoom In Zoom Out Reset image size

The upper panel of Figure 17 shows the J = 1–0 spectra of SiO and its isotopologues (SiO υ = 0, SiO υ = 1, SiO υ = 2, ²⁹SiO υ = 0, and ³⁰SiO υ = 0) detected in this survey. Overall, the line shapes agree with the single-dish results presented by Goddi et al. (2009) and Rizzo et al. (2017). The υ = 0 lines of ²⁹SiO and ³⁰SiO are broad and smooth, while the spectrum of SiO υ = 0 shows narrow features (Figure 17). This is consistent with previous results that the emission of SiO υ = 0 is part maser and part thermal (Chandler & de Pree 1995; Goddi et al. 2009).

Zoom In Zoom Out Reset image size

SiO υ = 1, 2 transitions are known to have been inverted since their discovery (Buhl et al. 1974; Thaddeus et al. 1974). The SiO υ = 1 of this survey and Goddi et al. (2009) have a nearly identical shape. This implies that the shape of SiO υ = 1 may have not obviously changed during the past 10 yr under the single-dish view. However, the bright and very narrow feature in the υ = 2 line at ∼−1.4 km s⁻¹ reported by Rizzo et al. (2017) cannot be seen in the spectra of Goddi et al. (2009) and this survey. The velocity component of SiO υ = 2 with a V_LSR of ∼22 km s⁻¹ has a much lower intensity compared with that of Goddi et al. (2009). The SiO υ = 2 seems to have experienced a more significant change than SiO υ = 1 during the past 10 yr.

There is a very narrow (spectrally not well resolved) and weak (∼0.1 K) emission feature at V_LSR ∼ −22 km s⁻¹ in the SiO υ = 1 spectrum (Figure 17). We have conducted observations covering this line on three different days adopting different frequencies of the LO, and this narrow and weak feature was detected in all sets. In the spectrum of SiO υ = 2, a broader feature was detected at a similar velocity (Figure 17). Thus, we identify the narrow and weak emission feature in the SiO υ = 1 spectrum as a narrow velocity component of SiO υ = 1.