© ESO, 2013

1. Introduction

The Orion Molecular Cloud 1 (OMC-1) is a unique environment, especially in the surroundings of the Orion BN/KL region (Becklin & Neugebauer 1967; Kleinmann & Low 1967) and one of the closest massive star-forming regions (413 ± 10 pc, Menten et al. 2007; Hirota et al. 2007; Sandstrom et al. 2007) for the study of interstellar chemistry. In this region, evidence of active star formation is overwhelming from detections of multiple strong infrared emission sources or embedded compact radio sources and strong or moderate molecular outflows. For example, a recent unusual explosive event (see, e.g., Allen & Burton 1993; Zapata et al. 2009) causes the finger-like shock structures seen in vibrationally excited H₂ and high-velocity (30−100 km s^-1) CO emission. Although several sources in BN/KL have been considered as the origin of the explosive outflow, e.g., the continuum source SMA-1¹ proposed by Beuther & Nissen (2008), this explosion is very likely related to the disintegration of a stellar system, which involves the Orion BN object, Source I, and maybe Source n (see, e.g., Gómez et al. 2005; Zapata et al. 2009; Goddi et al. 2011a). In addition, another outflow with a lower velocity (about 18 km s^-1) was detected through the proper motions of H₂O masers (e.g., Genzel et al. 1981; Gaume et al. 1998) accompanied by SiO masers near the infrared source IRc2 (e.g., Greenhill et al. 1998; Doeleman et al. 1999). Evidence indicates that this low-velocity bipolar outflow along a NE-SW axis is driven by source I (see e.g., Plambeck et al. 2009).

As one of the prime targets of interstellar chemistry study, Orion BN/KL clearly shows different molecular distributions between large nitrogen- (e.g., C₂H₃CN and C₂H₅CN) and oxygen-bearing (HCOOCH₃ and CH₃OCH₃) molecules (see, e.g., Blake et al. 1987; Liu et al. 2002; Friedel & Snyder 2008, and the references therein). Since the critical densities are fairly similar for N- and O-bearing species, Friedel & Snyder (2008) have proposed the age effect, the time that the molecular gas has had to evolve from an initial composition after being released from ice mantles, may be the most critical factor in the Orion BN/KL N-O differentiation. In addition, they find that the distribution of acetone (CH₃)₂CO is clearly different from those of typical N- or O-bearing molecules.

The differences between acetone and other large O-bearing molecules make acetone an excellent tool for testing different chemical models of complex molecules productions, i.e., in the gas phase by pure cosmic ray induced ion-molecule chemistry, on the surface of grains, or both. The higher temperatures in the vicinity of star-forming regions are certainly an important influence, because these allow some endothermic reactions to proceed. Another influence may be the recent shock event. If so, there would be a dependence on the time at which molecules were released to the gas phase and started to be processed by the gas-phase ion-molecule reactions.

Acetone was first detected in the interstellar medium (ISM) by Combes et al. (1987) and later confirmed by Snyder et al. (2002) toward the hot molecular core Sagittarius B2(N-LHM). They derived an acetone column density of 2−9 × 10¹⁶ cm^-2 with a rotational temperature of 100−270 K in a ~13″ × 7″ beam toward Sgr B2(N). The first acetone detection toward Orion BN/KL was reported by Friedel et al. (2005) using the BIMA, and the follow-up observations at 1 mm with the CARMA by Friedel & Snyder (2008) show that the acetone emission is compact in the Orion BN/KL region, and high angular resolution observations have the advantage of detecting weak signals from these complex organic species with compact spatial distributions. The derived acetone column density is 1−7 × 10¹⁷ cm^-2 assuming a rotational temperature of 150−500 K toward the acetone emission peaks in BN/KL.

The formation process of acetone in space is still unclear. Herbst et al. (1990) have shown that the ion-molecule radiative association reaction $\mathrm{C}{\mathrm{H}}_{\mathrm{3}}^{\mathrm{+}}\mathrm{+}\mathrm{C}{\mathrm{H}}_{\mathrm{3}}\mathrm{CHO}\mathrm{\to }\mathrm{(}\mathrm{C}{\mathrm{H}}_{\mathrm{3}}{\mathrm{)}}_{\mathrm{2}}\mathrm{CH}{\mathrm{O}}^{\mathrm{+}}$ proposed by Combes et al. (1987) is very inefficient even at low temperatures, where the rate coefficient for radiative association shows an inverse temperature dependence (≈T^-1). Their results suggest that other reactions may in work, e.g., grain surface chemistry, because the radiative association reaction cannot account for the observed acetone abundance in Sgr B2. Therefore, the strategy for determining the most likely formation route of interstellar acetone involves a high angular and velocity comparison of the acetone spectrum with other large O-bearing molecules, and an attempt to correlate the distribution of the acetone emission where shocks or heating sources are likely to release molecules from icy grain mantles.

The observations and acetone spectroscopy are presented in Sect. 2, and the high-quality images of the acetone emission distribution at <3′′ resolution in Orion BN/KL are presented in Sect. 3. The spatial and spectral comparisons between acetone and other molecules are discussed in Sect. 4. Furthermore, the acetone formation and the cloud structure of BN/KL are discussed in Sect. 5. Moreover, we have compared our Plateau de Bure Interferometer (PdBI) data with the science verification (SV) data of the Atacama Large Millimeter Array (ALMA) at 223 GHz in Appendix B. The three-dimensional visualization for different molecules in BN/KL is demonstrated in Appendix C.

2. Observations and spectroscopy

2.1. IRAM observations and line blending

The data used in this study are part of the large 1−3 mm data sets obtained from 1999 to 2007 using the PdBI² toward the Orion BN/KL region (see Favre et al. 2011a, for more observational details). Six data sets (Table 1) were used in this study, and the observations were carried out with five antennas equipped with two SIS receivers. The quasars 0458−020, 0528+134, 0605−085, and 0607−157, and the BL Lac source 0420−014 were observed for the phase and amplitude calibrations. The six units of the correlator allowed us to achieve different bandwidths and spectral resolutions. Using the IRAM 30 m single-dish data (J. Cernicharo, priv. comm.), the missing flux for the (CH₃)₂CO lines near 223 GHz is estimated as between 20% and 50%. The large uncertainty in this estimate is due to line confusion and the difficulty of determining spectral baselines in the 30 m data. The missing flux in our (CH₃)₂CO line interferometric observations near 101 GHz is estimated to be 25% to 35%.

The PdBI data were reduced with the GILDAS³ package, and the continuum emission was subtracted by selecting line-free channels. Our continuum emission images were presented in Favre et al. (2011a) where the H₂ column densities of selected clumps were also estimated. What we call continuum includes both true continuum emission (from the dust thermal and free-free emission) and a pseudo-continuum made of overlapping faint lines, which cannot be separated from the former true continuum. The data cube was then cleaned channel-by-channel with the Clark algorithm (Clark 1980) implemented in the GILDAS package.

Possible line blending with acetone was investigated by using the JPL database⁴, the Cologne database for molecular spectroscopy (CDMS⁵), and Splatalogue⁶. It is known that ¹³C-substituted ethyl cyanide (¹³CH₃CH₂CN, CH$\begin{array}{c}\mathrm{13}\\ \mathrm{3}\end{array}$CH₂CN, and CH₃CH$\begin{array}{c}\mathrm{13}\\ \mathrm{2}\end{array}$CN) has intense transitions in mm and submm regimes (Demyk et al. 2007; Fortman et al. 2012), and may blend with acetone lines in Orion BN/KL. We have examined the possible acetone line blending with ¹³C-substituted ethyl cyanide and found: (1) Their spatial distributions are concentrated at the hot core (HC, the strongest mm and submm continuum peak) region judging from the rather isolated CH$\begin{array}{c}\mathrm{13}\\ \mathrm{3}\end{array}$CH₂CN 25_5,20 − 24_5,19 line at 223 426.5 MHz (not shown). (2) The synthesized spectrum of ¹³C-substituted ethyl cyanide indicates that their contributions to the acetone intensities are likely to be minor (<10%) in our data (see Fig. 5) toward the main acetone emission peaks. The upper limit of ¹³C-substituted ethyl cyanide column density (beam averaged, $\mathrm{1}\stackrel{\mathrm{″}}{.}\mathrm{8}\mathrm{\times }\mathrm{0}\stackrel{\mathrm{″}}{.}\mathrm{8}$) at Ace-1 is about 2 × 10¹⁴ cm^-2 with a rotational temperature of 160 K (see also the synthesized spectra in Fig. 5). Additionally, a ¹³C-substituted ethyl cyanide column density of about 2 − 4 × 10¹⁴ cm^-2 was estimated toward IRc2 (with a diameter of 7′′) at 300 K, which is lower than the value (1.6 × 10¹⁵ cm^-2 with 300 K) given by Demyk et al. (2007) assuming the same source size toward IRc2. This difference may be because the ¹³C-substituted ethyl cyanide in Orion BN/KL is extended and was largely resolved out in our PdBI observations. Therefore, we conclude that the line blending due to ¹³C-substituted ethyl cyanide is not severe in our case.

2.2. (CH₃)₂CO spectroscopy

Acetone is a double methyl-substituted version of the ubiquitous species formaldehyde H₂CO. Owing to its two internal rotators (C_2V symmetry), acetone has four torsional substates AA, EE, EA, and AE with slightly different rotational energy levels and different spin statistical weights (see Groner et al. 2002, for more details). The designations AE, EA, EE, and AA refer to the different symmetry states of the molecule. Under normal interstellar conditions, these different symmetry states will not interchange. Because of the different degenerate nature of the torsional substates, the spin weighting has to be taken into account in the column density calculation. The acetone spin weights for the AA:EE:EA:AE symmetry states are 6:16:4:2 for K_aK_c = ee,oo and 10:16:4:6 for K_aK_c = eo,oe. The b-type asymmetric top selection rules (ΔJ = 0, ± 1; ΔK_a = ± 1,3,...; ΔK_c = ± 1,3,...) apply to the transitions without severe torsional-rotational interactions (see Groner et al. 2002).

The effective rotational partition function q is approximated by q = 261.67 T^1.5 for the temperature above 100 K (Groner et al. 2002). The rotational-vibrational partition function used by Friedel et al. (2005) is ${\mathit{Q}}_{\mathrm{rv}}\mathrm{\approx }{\sum }_{\mathit{i}\mathrm{=}\mathrm{0}}^{\mathrm{2}}{\mathit{e}}^{\mathrm{-}{\mathit{E}}_{\mathit{i}}\mathit{/}\mathit{T}}\mathit{q}$, where E_i are the first three vibrational states at 0, 115, and 180 K (Friedel et al. 2005). However, we found that the acetone partition function given in the JPL database starts to deviate from the one used by Friedel et al. (2005) as T ≳ 120 K (see Fig. A.1). It is likely that the approximation used by Friedel et al. (2005) does not take higher rotational and/or vibrational states into account which start to be essential at higher temperatures. Therefore, whereas the difference is limited for the temperature range (≲160 K) derived here for acetone, one should use the more complete acetone partition function given in the JPL database for higher temperatures. The acetone molecular parameters in this paper were taken from the JPL database.

2.3. (CH₃)₂CO critical density

The critical densities n_c of acetone can be estimated according to ${\mathit{n}}_{\mathrm{c}}\mathrm{=}\frac{{\sum }_{\mathit{l}\mathit{<}\mathit{u}}{\mathit{A}}_{\mathit{ul}}}{{\sum }_{\mathit{l}\ne \mathit{u}}{\mathit{C}}_{\mathit{ul}}}\mathit{,}$(1)where A_ul is the radiative de-excitation rate (Einstein A-coefficient) and C_ul the collisional de-excitation rate (Kwok 2007). Assuming a two-level system, the collisional de-excitation rate can be estimated as (see Friedel & Snyder 2008; Tielens 2005; Flower 2007) ${\mathit{C}}_{\mathit{ul}}\mathrm{=}{\left(\frac{\mathrm{8}\mathit{kT}}{{\mathit{M}}_{\mathit{\mu }}\mathit{\pi }}\right)}^{\frac{\mathrm{1}}{\mathrm{2}}}{\mathit{\sigma }}_{\mathit{ul}}\mathit{,}$(2)where M_μ is the reduced mass of acetone and its main collisional partner H₂, and σ_ul is the collisional cross section of acetone. Here we assume that the collisional cross section of acetone is approximated by its geometrical cross section (~8 × 10^-16 cm^-2). We find C_ul ≈ 1.0 × 10^-10 cm³ s^-1 for acetone at 150 K. The critical densities for the detected acetone transitions at 3 and 1.3 mm are estimated to be 4 × 10⁵ − 2 × 10⁸ cm^-3. Therefore, the acetone emission generally probes a dense region with a density higher than a few times 10⁵ cm^-3 at these frequencies, which is similar to HCOOCH₃, C₂H₅CN, and (CH₃)₂O (Favre et al. 2011a; Friedel & Snyder 2008).

3. Results

Our data include 22 acetone lines corresponding to 40 transitions (see Table 2) with an upper-state energy below 300 K. Nine of these lines appear free of contamination by other molecular emission. Three main acetone emission peaks (Ace-1, 2, and 3) were identified in the lower angular resolution images around 101.4 GHz. The acetone emission toward Ace-3 has been detected for the first time. The acetone transition parameters are summarized in Table 2, and the line measurements toward selected positions are listed in Tables A.1 to A.4. The mean local standard of rest (LSR) velocities and FWHM line widths toward selected positions are listed in Table 3.

The LSR velocities of the selected positions are consistent with those reported by Favre et al. (2011a) for methyl formate and Chandler & Wood (1997) for CS J = 1 − 0. The average FWHM line width of acetone is about 3.5 km s^-1, similar to those of HCOOCH₃ and (CH₃)₂O lines (Favre et al. 2011a; Brouillet et al. 2013). It is interesting to note that the acetone line widths of Ace-1 are narrower than those of Ace-2, Ace-3, and HC.

3.1. Overall acetone distribution

The acetone emission in BN/KL appears mainly at three peaks (Ace-1, Ace-2, and Ace-3) in our lower resolution images ($\mathrm{3}\stackrel{\mathrm{″}}{.}\mathrm{8}\mathrm{\times }\mathrm{2}\stackrel{\mathrm{″}}{.}\mathrm{0}$, Fig. 1), where the acetone emission around Ace-1 seems more extended than those around Ace-2 and Ace-3. The positions of the Ace-1 and Ace-2 peaks were determined at the central velocity channels of our highest resolution images ($\mathrm{1}\stackrel{\mathrm{″}}{.}\mathrm{8}\hspace{0.17em}\mathrm{\times }\hspace{0.17em}\mathrm{0}\stackrel{\mathrm{″}}{.}\mathrm{8}$), and the Ace-3 position was determined on the lower resolution images due to the interferometric filtering effects and lower sensitivities of high-resolution images toward this position. The positions of the three acetone peaks are also consistent with the ALMA-SV data (Fig. B.2). The arc-like structure in the north of HC is revealed for the first time in our highest resolution image (Fig. 2). The reason Friedel & Snyder (2008) did not detect the arc-like extended emission and the Ace-3 peak is likely the different uv plane coverage and lower sensitivities.

Ace-1 corresponds to the strongest methanol and deuterated methanol emission peak (dM-1, Peng et al. 2012a) and the second strongest methyl formate emission peak (MF-2, Favre et al. 2011a). Ace-2 is close to the infrared source IRc7 (see, e.g., Shuping et al. 2004; Gezari et al. 1998) but located about 1′′ southwest of its emission peak. In addition, Ace-3 is located at the same position of the MF-5 (Favre et al. 2011a) and methanol (including deuterated methanol) peak KL-W (e.g., Peng et al. 2012a), and close to the infrared source IRc6 (see, e.g., Shuping et al. 2004; Gezari et al. 1998). The Ace-3 peak shows weak emission in our highest resolution images near 223 GHz because the extended and weaker acetone emission at Ace-3 is likely to be filtered out by the PdBI. We have also investigated the ALMA-SV data at the same frequency with slightly lower resolution and found clear detections of acetone toward Ace-3 (see Appendix B for more details). It indicates that the two velocity components seen in the acetone and methyl formate emission are most likely caused by the similar filtering effects. Alternatively, there is a possibility of two large-scale structures associated with Ace-3, which would be too extended to be detected by the PdBI (see also Chandler & Wood 1997). These large structures may extend to the outer part of the BN/KL region, which may also explain the two velocity components at about 6 and 11 km s^-1 of the Herschel/HIFI O₂ detection (Goldsmith et al. 2011) toward H₂ Peak 1, about 26′′ to the northwest of Ace-3.

Acetone shows weak emission in the southern part of Orion BN/KL close to MF-1 and MF-3. However, the emission is extended (e.g., in the low resolution maps Fig. 1) and also suffers from filtering effects in our PdBI data (see Fig. B.5).

3.2. Excitation temperature and column density

We use population diagrams to estimate the column density of (CH₃)₂CO (e.g., Goldsmith & Langer 1999; Turner 1991) $\ln \frac{{\mathit{N}}_{\mathrm{up}}}{{\mathit{g}}_{\mathrm{up}}}\mathrm{=}\ln \frac{\mathrm{3}\mathit{k}\mathrm{\int }{\mathit{T}}_{\mathrm{B}}\mathrm{d}\mathit{V}}{\mathrm{8}{\mathit{\pi }}^{\mathrm{3}}\mathit{\nu S}{\mathit{\mu }}^{\mathrm{2}}{\mathit{g}}_{\mathrm{s}}\mathit{f}}\mathrm{=}\ln \frac{\mathit{N}}{\mathit{Q}}\mathrm{-}\frac{{\mathit{E}}_{\mathrm{up}}}{\mathit{k}{\mathit{T}}_{\mathrm{rot}}}\mathit{,}$(3)where N is the total column density, T_rot the rotational temperature, S the line strength, μ the dipole moment (μ_b = 2.93 D, Peter & Dreizler 1965), g_s the spin statistical weight, and f the beam filling factor. The mean optical depth of the acetone lines is estimated to be 0.02–0.03 toward the selected clumps near 101 GHz and less than 0.06 near 223 GHz with excitation temperatures of 100–150 K, taking the dust continuum emission into account and assuming a 90% continuum flux loss due to filtering. The (CH₃)₂CO beam-averaged column densities and rotational temperatures are estimated from a least-squares fit in the population diagrams. The uncertainties shown in these diagrams reflect the statistical fitting error in data where thermal noise and 10% calibration uncertainty dominate. There is no attempt to estimate an absolute error, and we note that estimating the beam filling factor is uncertain. In addition, the column density uncertainties owing to the missing flux are likely to be less than a factor of two (see Appendix B for the comparison with the ALMA-SV data). The estimated acetone column densities, rotational temperatures, and fractional abundances toward the selected positions are summarized in Table 3.

Ace-1

Toward Ace-1 (Fig. 3), an acetone column density is estimated to be 3.4 ± 0.1 × 10¹⁶ cm^-2 with an excitation temperature of 159 ± 13 K, assuming a source size of 2′′. An H₂ column density of 2.4 ± 0.2 × 10²⁴ cm^-2 is derived toward the same position by using the 1 mm continuum emission (Favre et al. 2011a), assuming a dust temperature of 150 K and an absorption coefficient κ_ν = 5 × 10^-4 cm² g^-1. Therefore, the acetone abundance toward Ace-1 is 1.4 ± 0.4 × 10^-8 and agrees with the result (0.34 − 1.84 × 10^-8) of Friedel et al. (2005).

Ace-2

We derived an acetone rotational temperature of 145 ± 16 K and a column density of 2.1 ± 0.1 × 10¹⁶ cm^-2 toward Ace-2 (Fig. 3), similar to those of Ace-1. From the 1.3 mm continuum emission of Favre et al. (2011a), we derived an H₂ column density of 3.8 ± 2.0 × 10²³ cm^-2 toward Ace-2. Owing to its weaker dust continuum emission, the acetone fractional abundance (5.5 ± 4.0 × 10^-8) of Ace-2 seems to be higher than that of Ace-1 but with a larger uncertainty.

Ace-3

The acetone rotational temperature of Ace-3 is estimated to be ~100 K, and the acetone column density to be 6.3 ± 1.9 × 10¹⁵ cm^-2 (Fig. 4). From the 1.3 mm continuum emission of Favre et al. (2011a), we derived an H₂ column density of 1.3 ± 0.3 × 10²⁴ cm^-2 toward Ace-3 with a dust temperature of 100 K. The acetone abundance toward Ace-3 is 4.8 ± 1.8 × 10^-9.

HC

The acetone rotational temperature of HC is estimated to be about 141 K and the column density is estimated to be 1.2 ± 0.2 × 10¹⁶ cm^-2 (Fig. 4). From the 1.3 mm continuum emission of Favre et al. (2011a), we have derived an H₂ column density of 1.3 ± 0.3 × 10²⁴ cm^-2 toward HC with a dust temperature of 140 K. The acetone abundance toward HC is 2.1 ± 0.4 × 10^-9.

3.3. Synthesized spectra of acetone

The synthesized spectra of acetone (Fig. 5) were produced with Weeds (Maret et al. 2011) in the GILDAS package, assuming that the source sizes equal the PdBI synthesized beams at 101 and 223 GHz. Although the synthesized spectra do not fully reproduce the observed acetone intensities, the uncertainty of 10%−20% is likely contributed by line blending, different beam sizes, the accuracy of the acetone column density, excitation temperature, and possible non-LTE effects.

4. Comparison with other molecules

4.1. Spatial comparison

The detection of two molecules with strong and not blended lines, C₂H₅CN and HCOOCH₃, within the same data sets as acetone allows us to directly compare their distributions with the same observational parameters. In Fig. 7, the integrated intensity maps of C₂H₅CN, (CH₃)₂CO, and HCOOCH₃ were overlaid on the same velocity interval. The emission of the large N-bearing molecule C₂H₅CN is mainly located in the northern part of the Orion BN/KL region, i.e., close to HC and IRc7, similar to the highly excited NH₃ (Goddi et al. 2011b). The emission of the typical O-bearing molecule HCOOCH₃ appears mostly in the south, i.e., the Compact Ridge region with a typical V_LSR of about 8 km s^-1, whereas the acetone emission appears in the regions where the C₂H₅CN emission is also present. It is interesting to note that the HCOOCH₃ emission does not appear in the regions associated with hot (≈400 K) and dense ammonia (Goddi et al. 2011b). Since the emissions of C₂H₅CN and highly excited NH₃ (i.e., inversion transitions from J,K ≥ 9,9) share a similar distribution, it is likely that C₂H₅CN traces somewhat higher temperatures than HCOOCH₃ does, and (CH₃)₂CO may be excited in the intermediate temperature range (see also Friedel & Snyder 2008). For example, the excitation temperature of C₂H₅CN toward the HC region is about 150 K, and is about 100 K for HCOOCH₃ toward the Compact Ridge region (Favre et al. 2011a; Blake et al. 1987). It is probably related to the acetone formation route in the interstellar medium (see Sect. 5). The high similarity of the acetone and C₂H₅CN distributions is also seen in the 3D images in Appendix C, where both molecules show a similar arc-like structure that may be associated with the explosive event.

Chandler & Wood (1997) obtained a high-resolution (${\mathit{\theta }}_{\mathrm{syn}}\mathrm{=}\mathrm{2}\stackrel{\mathrm{″}}{.}\mathrm{1}\mathrm{\times }\mathrm{1}\stackrel{\mathrm{″}}{.}\mathrm{7}$) CS J = 1 − 0 image with VLA, and they identified six main CS peaks in Orion BN/KL. It is essential to note that all acetone peaks have counterparts in the CS J = 1 − 0 emission. The CS peaks CS-3, 5, and 6 correspond to Ace-1, 2, and 3, respectively, and CS-1 is located at the HC position (see Table 3 for the counterpart summary). In addition, the CS J = 1 − 0 emission is strong in the south of BN/KL (i.e., CS-4) close to MF-1 and 3, where most of the large O-bearing molecules are detected. The overall CS distribution covers the region where large N- and O-bearing molecules are detected. The sulfur-bearing molecules (e.g., CS, SO, and SO₂) mainly form via warm gas-phase chemistry or shock chemistry (e.g., Charnley 1997; Pineau des Forêts et al. 1993), and they coincide with both N- and O-bearing molecules in BN/KL (Schreyer et al. 1999; Chandler & Wood 1997). This suggests that the formation routes via warm gas phase or shock chemistry may lead to the formation of acetone in BN/KL.

Furthermore, Chandler & Wood (1997) noticed that the integrated CS emission does not coincide with any identified clumps. This may be because the velocity dispersion in the arc-like structure is higher, and the integrated flux is dominated by the line width rather than the peak intensity. Although the acetone integrated intensity shown in Figs. 2 and 7 peaks at SMA-1 (Tang et al. 2010; Beuther et al. 2004), it is likely to be caused by increased velocity width, similar to the CS emission peak mentioned above.

4.2. Spectral comparison

In Fig. 8, we compare the line profiles of (CH₃)₂CO with C₂H₅CN and HCOOCH₃, together with formamide (HCONH₂), which is an interesting molecule because it contains one oxygen and one nitrogen atom and was detected in our data set near 223 GHz. It is interesting to note that the acetone spectra toward Ace-1 and Ace-2 show strong blueshifted line wings at about 4 km s^-1. The HCONH₂ 11_1,11 − 10_1,10 spectrum toward Ace-2 also shows a blueshifted line wing. These blue-wing features may result from shocks. On the other hand, the HCOOCH₃ spectra do not have clear line wings.

Toward Ace-3, the (CH₃)₂CO and HCOOCH₃ spectra reveal that at least two velocity components are present. However, a velocity shift in V_LSR of the two components is observed, where the velocity difference between the two components is larger in acetone (~4 km s^-1) than in methyl formate (~2 − 3 km s^-1). In addition, the ALMA-SV acetone spectrum toward Ace-3 shows redshifted line wings (see Fig. B.4), which is not clear in our PdBI data. The differences observed between the ALMA-SV and PdBI data are most probably due to different uv coverages and sensitivities (see Table B.1). Because of its extended spatial nature, the acetone emission from Ace-3 is less likely to be recorded in our PdBI data.

Similarly, blueshifted wings can be seen in CS J = 1−0 toward CS-3 (Ace-1) and CS-5 (Ace-2) (Chandler & Wood 1997). Additionally, the CS J = 1−0 line widths are broader in CS-6 (Ace-3) than in CS-3 and CS-5. The two velocity components seen in CS-6 are also due mainly to the filtering effects similar to acetone and methyl formate, compared with the ALMA-SV data. CS-3 (Ace-1) has a rather narrow line width of 3.3 km s^-1, similar to acetone. Therefore, most of the molecular lines in some regions of BN/KL (i.e., HC/CS-1, Ace-2/CS-5, and Ace-3/CS-6) show broader line widths than those in the other regions (i.e., Ace-1/CS-3, MF-3/dM-2/CS-4, and dM-3/MF-1), which indicate that some regions may be more directly affected by shocks than others. These shocks could be driven by local outflow phenomena and/or the Orion explosion event.

5. Discussion

5.1. Acetone formation

Some studies have suggested there is formation of carbonyl species, such as H₂CO, CH₃CHO, and (CH₃)₂CO, on or within icy grain mantles (e.g., Snyder et al. 2002; Bennett et al. 2005). In the model of Garrod et al. (2008), acetone is said to form by addition of CH₃ to CH₃CO and behaves similarly to HCOOCH₃ and (CH₃)₂O. Recent results from Favre et al. (2011b) and Brouillet et al. (2013) show good agreement in the distributions between HCOOCH₃ and (CH₃)₂O in BN/KL. However, acetone clearly has a different distribution from those of HCOOCH₃ and (CH₃)₂O, and more similar ones to that of C₂H₅CN. It is not clear whether the acetone binding energy (it evaporates from the ice mantle at about 65 K) used in the model of Garrod et al. (2008) is a critical factor or not. The acetone excitation temperature toward Ace-1 is estimated to be about 150 K, and it agrees with that of HCOOCH₃ (140 ± 14 K, Favre et al. 2011a). Therefore, (CH₃)₂CO and HCOOCH₃ seem to trace the same gas toward Ace-1. On the other hand, not detecting HCOOCH₃ toward the Ace-2/IRc7 position indicates that different formation processes from (CH₃)₂CO and HCOOCH₃ are involved. It is possible that some forming pathways of acetone have not been taken into account in the previous modelings. For example, Chen & Woon (2011) carried out a theoretical investigation of possible reactions between carbonyl species and ammonia and show that acetone may react with water and ammonia via nucleophilic addition with low barriers. Although the final products of these reactions on icy grain mantles are not clear, further research is important. Because of its similar distribution to the N-bearing molecules, the formation of acetone may involve ammonia or N-bearing molecules or need similar conditions, e.g., shocks.

It has been proposed by Remijan et al. (2004) that the formation of acetic acid (CH₃COOH) favors the hot molecular cores with well-mixed N- and O-bearing molecules. This may be another example that the formation of a large O-bearing molecule involves N-bearing molecules or needs similar conditions, although acetic acid has not been detected so far toward BN/KL (Remijan et al. 2003). However, acetone emission is detected toward the positions (i.e., IRc7 and arc-like structure in the north) where there is no strong emission of large O-bearing species like HCOOCH₃ and (CH₃)₂O (Favre et al. 2011a; Brouillet et al. 2013). Furthermore, it is interesting to search for the molecules with distributions similar to acetone in Orion BN/KL in the future. Investigations of molecules related to acetone will be important for understanding the origins of interstellar acetone.

5.2. Overall cloud structure of BN/KL

A clearer picture has emerged from the investigations of different molecules’ distribution and the recent understanding of the stellar feedbacks inside the Orion BN/KL region. It has been shown that the explosive outflow that happened about 500 years ago may affect the BN/KL region in many ways. For example, Zapata et al. (2011) propose that HC, the strongest mm and submm continuum peak, is probably heated by the explosive outflow owing to its lack of embedded heating sources. In addition, Peng et al. (2012b) suggest that the high-velocity bullets from the explosive outflow may be absorbed by the BN/KL cloud and induce the methanol masers seen in the south. The absorbed kinetic energy can then be transferred to thermal energy that affects the chemical processes in the cloud. As expected for the decelerated bullets, the thermal energy (or velocity) deposit is higher close to the explosion center than close to the stop point. This is clearly seen in the north-south differences in excitation temperatures and line widths of many molecules (see Table 4).

Figure 9 shows a cartoon overview of different molecular distributions in BN/KL. As discussed above, because the spectral features of CS are consistent with other molecules toward the same clump (e.g., smaller line widths in the south compared to the north of the cloud), the N-O differentiation seen in BN/KL is very likely caused by physical conditions (i.e., warmer and shocked gas) instead of their chemical properties (e.g., ice mantle composition). On the other hand, the age effect cannot be ruled out since different outflows in the northern region (e.g., the bipolar outflow of source I) may contribute, too, and the reason the northern part is more turbulent may be its more evolved nature. Nevertheless, the bipolar outflow of source I alone cannot explain the blueshifted wings seen in most of the acetone peaks, because the outflow lies along a NW-SE axis (see, e.g., Plambeck et al. 2009) and its orientation tends to be constant when moving from the explosion center to the present position (Bally et al. 2011). In addition, the acetone peaks are located in the direction of the source I southern outflow lobe, which appears to be more redshifted than the northern one, but the line wings associated with Ace-1 and Ace-2 are blueshifted.

6. Conclusions

We have explored the acetone distribution in Orion BN/KL at angular resolutions between $\mathrm{1}\stackrel{\mathrm{″}}{.}\mathrm{8}\mathit{,}\mathrm{\times }\hspace{0.17em}\mathrm{0}\stackrel{\mathrm{″}}{.}\mathrm{8}$ and $\mathrm{6}\stackrel{\mathrm{″}}{.}\mathrm{0}\hspace{0.17em}\mathrm{\times }\hspace{0.17em}\mathrm{2}\stackrel{\mathrm{″}}{.}\mathrm{3}$ (about 300−2500 AU) from the 3 mm and 1.3 mm PdBI data sets. The main findings follow.

• 1.

  Our sample consists of 22 lines of acetone with an upper-state energy below 300 K, and nine of these lines appear free of contamination by other molecular emission. The FWHM line width of these acetone lines is about 3.5 km s^-1, similar to those of HCOOCH₃ and (CH₃)₂O lines.

• 2.

  The acetone distribution exhibits three main peaks (Ace-1, 2, and 3) and the extended emission toward the Hot Core and the northern region, which appears to be arc-like. Ace-1 is also the strongest peak of the deuterated methanol emission (dM-1, Peng et al. 2012a) and second strongest peak of HCOOCH₃ (MF-2 Favre et al. 2011a). Ace-2 is located close to the peak of the IR source IRc7 with a separation of about 1′′. The new weaker source Ace-3, which was not detected by Friedel & Snyder (2008), coincides within 2′′ with the KL-W/MF-5 position (Peng et al. 2012a; Favre et al. 2011a) and is close to the IR source IRc6.

• 3.

  The overall distribution of acetone in BN/KL is more extended than the one found by Friedel & Snyder (2008) and is similar to that of the large N-bearing species (i.e., C₂H₅CN). This is likely due to the higher sensitivity and different uv coverage of our PdBI data compared with the CARMA data of Friedel & Snyder (2008).

• 4.

  Rotational temperatures of 159 ± 13 are determined toward Ace-1 and 145 ± 16 K toward Ace-2. The acetone column density toward these two peaks is about 2−4 × 10¹⁶ cm^-2 with a relative abundance of 1−6 × 10^-8, within the range given by Friedel et al. (2005). In addition, Acetone is a few times less abundant toward the hot core and Ace-3 peaks (2−7 × 10^-9) compared with Ace-1 and Ace-2.

Current chemical models require a formation of acetone on grain surfaces as for (CH₃)₂O, and they cannot explain the difference seen between the two species. Including reactions with N-bearing species either to form or to destroy acetone might be necessary to explain its different distribution (e.g., Chen & Woon 2011). Additional laboratory and theoretical chemical studies are needed to progress in the understanding of the acetone abundance. High-resolution maps with more sensitive ALMA data will hopefully lead to the discovery of other species with the same spatial distribution, and confirm any association with shocks.

Online material

Appendix A: Complementary figures and tables

Appendix B: Comparison with the ALMA science verification data

Basic observational parameters

The calibrated ALMA-SV data were retrieved from the ESO archive. The observations were carried out in 2012 January with 16 12-m antennas (see e.g., Zapata et al. 2012; Niederhofer et al. 2012, for more observational details). The primary beam at 223 GHz is about 28′′. Some observational parameters of the ALMA-SV and our PdBI data are listed in Table B.1. The ALMA-SV data were analyzed with the GILDAS package, and the acetone emission images were cleaned with the same Clark algorithm as for the PdBI data to minimize the effects contributed by different cleaning algorithms. The ALMA and PdBI channel maps of the acetone 17_7,11 − 16_6,10 EE line are shown in Fig. B.2, and the spectrum comparison around 223 GHz toward different positions is shown in Figs. B.3–B.5. The missing flux of the ALMA-SV data is estimated to be 10−20% compared with the IRAM 30 m single-dish data (J. Cernicharo, priv. comm.) around 223 GHz, less than our PdBI data (20−50% missing flux).

Spatial comparison

Figure B.1 shows the uv coverages of the PdBI and ALMA-SV data. They are comparable, but the ALMA-SV data show a denser coverage for shorter spacings. The PdBI data have longer baselines, which result in a higher angular resolution. However, due to the few uv tracks in the −50 to 50 m domain, the PdBI observations are likely to filter out larger structures in BN/KL. This is clearly seen in Fig. B.2 in which most of the emission toward Ace-3 and in the southern part of BN/KL close to MF-1 and MF-3 is filtered out in the PdBI data. Likewise, it is important to note that the acetone emission associated with the arc-like structure in the north of HC is not resolved out by the PdBI.

Spectral comparison

Figure B.3 shows the ALMA-SV and PdBI spectra toward Ace-1 and 2. The differences between the two data sets are small for the HCOOCH₃ and (CH₃)₂CO weak lines, but greater for the C₂H₅CN and SO₂ strong or broad lines. It is because these lines tend to be spatially extended and are easily resolved out by the PdBI. Filtering effects are more obvious toward Ace-3 and MF-3 (Figs. B.4 and B.5) where most of the line emission, including SO₂, C₂H₅CN, and (CH₃)₂CO, is filtered out.

Summary

Our PdBI spatial distribution and individual spectra are very consistent with the new ALMA-SV data toward Orion BN/KL. However, due to different uv coverages and the smaller PdBI primary beam, some extended emission is filtered out in our data, especially toward Ace-3 and in the southern part (close to MF-3) of BN/KL. It is not clear why the arc-like acetone emission close to HC has not been resolved out. One possible reason is that the acetone emission in the southern part is weaker and suffers more easily from filtering effects, as observed toward the Ace-3 region. The fully functional ALMA in the near future with a denser uv coverage and the short spacings of the Atacama Compact Array will surely help solve these issues and greatly improve the molecular line images.

Appendix C: Three-dimensional view of molecular structures in BN/KL

The three-dimensional visualization enables us to investigate the distributions and possible structures of different molecules in great detail. Therefore, we have prepared the three-dimensional (RA, Dec, and V_LSR) images of different molecules in Orion BN/KL using the 3D Slicer program⁷ and the complemented procedures provided by the Astronomical Medicine Project⁸. Figures C.1−C.4 show the different viewing angles of the C₂H₅CN, (CH₃)₂CO, and HCOOCH₃ structures in BN/KL. The superposed images of these molecules were also plotted. In addition, a complementary 3D movie will be available on the A&A website.

C ₂ H ₅ CN

The C₂H₅CN line has a larger line width compared with those of (CH₃)₂CO and HCOOCH₃ and is clearly seen in Figs. C.3 and C.4. The arc-like structure close to HC is clearer in the 3D images (e.g., Figs. C.1 and C.2). In addition, the slow outflow associated with IRc7 in the NH₃ emission proposed by Goddi et al. (2011b) may actually be the two different clumps in the

SE and NW of IRc7 (see Figs. C.3 and C.4) because of the multi-component structure in the blueshifted clump revealed in the image. This complicated blueshifted clump was detected in the methyl cyanide emission (Wang et al. 2010) toward southeast of IRc7. These components may be due to the explosive outflow. However, we cannot rule out the possibility of a multi-outflow that comes from IRc7.

(CH ₃ ) ₂ CO

The arc-like structure is clearly seen in the acetone 3D images. It is interesting to note that the filtered-out emission toward Ace-3 is mostly centered at 8 km s^-1, which creates an artificial shell-like structure in the 3D images.

HCOOCH ₃

Part of the arc-like structure seen in C₂H₅CN and (CH₃)₂CO is also seen in HCOOCH₃ where the HCOOCH₃ emission extends from MF-3 to Ace-1 and even to the northwest of HC (Figs. C.1 and C.2). In addition, some structures close to MF-1 are seen and may be related to outflows/shocks.

 

Movie of Figures C.1−C.4 (MOV 15.8 MB) Access here

 

---

Acknowledgments

We thank the referee for the comments and helpful suggestions. T.-C. Peng acknowledges the support from the ALMA grant 2009-18 at Université de Bordeaux1/LAB. This paper makes use of the following ALMA data: ADS/JAO.ALMA#2011.0.00009.SV. ALMA is a partnership of ESO (representing its member states), NSF (USA) and NINS (Japan), together with NRC (Canada) and NSC and ASIAA (Taiwan), in cooperation with the Republic of Chile. The Joint ALMA Observatory is operated by ESO, AUI/NRAO and NAOJ.

References

All Tables

All Figures