COMPLEX ORGANIC MOLECULES DURING LOW-MASS STAR FORMATION: PILOT SURVEY RESULTS^*

Figure 1 shows the complete 16 GHz spectra toward all six sources, ordered in terms of line richness. Two of the YSOs, SVS 4-5 and B1-a, stand out as particularly line rich. Both sources are in the vicinity of two other YSOs, SMM4 and B1-b, that are known hosts of complex molecules. Of the remaining four sources, B5 IRS1 and IRAS 03235+2004 are more line dense compared to the two Taurus sources L1489 IRS and IRAS 04108+2803.

These differences in line density seems correlated with the strength of the CH₃OH 2–1 ladder as shown in Figure 2. SVS 4-3 and B1-a both display strong CH₃OH lines, B5 IRS1 moderate ones, and IRAS 03235, IRAS 04108, and L1489 IRS very weak lines (peak intensities <0.2 K). In addition to the 2–1 lines, there are a handful of other CH₃OH lines throughout the observed frequency range with excitation energies of 6–83 K (as well as higher energy lines which are not detected toward any source). Table 2 lists the integrated line intensities or 3σ upper limits of all CH₃OH lines detected toward at least one source. The integrated intensities were determined using IDL and MPFIT to fit Gaussians to the expected line positions. The 1σ integrated line intensity uncertainties were extracted from the fit procedure and are often larger than the 1σ rms because it includes the fit uncertainty. 3σ upper limits were determined using the rms in each 4 GHz chunk and the average CH₃OH line FWHM (Table 3) for each source. Table 2 also presents the integrated line intensities of two ¹³CH₃OH lines detected toward B1-a and SVS 4-5.

Zoom In Zoom Out Reset image size

Table 3. Line FWHM with Standard Deviations

Source	FWHM$_{\rm CH_3OH}$
	[km s−1]
B1-a	1.6[0.4]
SVS 4-5	3.7[0.7]
B5 IRS1	0.8[0.1]
IRAS 03235	0.9[0.5]
IRAS 04108	1.2[0.6]
L1489 IRS	1.6[1.0]

Download table as:  ASCIITypeset image

Figure 3 shows that in the case of SVS 4-5 (and B1-a) the multitude of weak lines hinted at in Figure 1 are to a large degree associated with the COMs HNCO, H₂CCO, CH₃CHO, CH₃OCH₃, and CH₃CN. Line identifications were made using the Splatalogue web tool drawing upon the CDMS and JPL spectral databases (Pickett et al. 1998; Müller et al. 2001). Other lines are identified with simple molecules, CH₃OH and carbon chains (carbon chain abundances will be the topic of a future publication).

Zoom In Zoom Out Reset image size

Figure 4 shows blow-ups of the spectral regions with the strongest lines of the six complex organics detected toward at least one source. HNCO is detected toward all sources in the sample. CH₃CHO and CH₃CN are clearly detected toward SVS 4-5 and B1-a, and marginally toward B5 IRS1. CH₃OCH₃ is only detected toward B1-a. H₂CCO is detected toward B1 a, SVS 4-5, IRAS 03235, and IRAS 04108. HCOOCH₃ is marginally detected toward SVS 4-5 and B1-a (the reality of these 3σ detections are supported by several more marginal detections throughout the spectral range). In general, we claim marginal detections for line intensities that exceed the 3σ upper limit for that source and sideband. To count as a clear detection we furthermore require that the Gaussian fit is sufficiently well-defined to result in an integrated intensity estimate of 30% or less. Table 4 lists the detected COM line intensities and upper limits, calculated using Gaussian fits as described for CH₃OH lines above.

Zoom In Zoom Out Reset image size

The CH₃OH excitation temperatures and column densities were determined using the rotational diagram method (Goldsmith & Langer 1999), assuming optically thin lines and LTE at a single temperature—the validity of these assumptions and the constraints on the kinetic temperatures provided by the excitation temperatures are explored further in Section 4.3. Figure 5 shows the results for CH₃OH. A single fit to all lines result in excitation temperatures of 15–20 K for the sources with any high energy lines detected.

Zoom In Zoom Out Reset image size

The observed CH₃OH emission does not generally seem to be well described by a single excitation temperature, however, but rather seems to trace a warm and cold component. Too few high-energy lines are detected to quantify the warm component, but the cold component excitation temperatures and column densities can be estimated by focusing on the low-energy lines for the fit. This result in excitation temperatures of 6–8 K. The derived column densities should be representative of the protostellar envelopes outside of the core region where thermal evaporation is possible, i.e., at T <100 K—because CH₃OH is readily sub-thermally excited it is not possible to a priori constrain the emission region any further. For IRAS 03235, too few low energy lines were detected to determine an envelope temperature, and an excitation temperature of 8 K was assumed to constrain the envelope column density.

To estimate the amount of warm CH₃OH column (beam-averaged) in each line of sight, we use an excitation temperature of 26 K, based on the CH₃CN analysis below, and the one high-energy CH₃OH line excluded in the cold component fit. Table 5 reports the excitation temperatures and column densities calculated for the three fits characterizing the total, the cold and the warm protostellar CH₃OH column. The cold column densities range between 0.5–10 × 10¹³ cm⁻², the warm column densities range between <0.12–9 × 10¹³ cm⁻² and the total column densities range between 0.5–23 × 10¹³ cm⁻². The column density uncertainties includes a 10% calibration uncertainty in addition to the fit uncertainty. The listed uncertainties do not incorporate the fact that the cold component will contribute slightly to the high-energy line intensity and vice versa, resulting in systematic overestimates of the cold and warm component column densities, and indeed the sum of the cold+warm component column densities is 10%–50% higher compared to the column densities derived from the single component fit. The component column densities are thus at best accurate within a factor of two.

Table 5. CH₃OH Excitation Temperatures and Column Densities

Source	Trot Ave	N$_{\rm CH_3OH}$ Ave	Trot Cold	N$_{\rm CH_3OH}$ Cold	Trot Warma	N$_{\rm CH_3OH}$ Warmb
	(K)	(cm−2)	(K)	(cm−2)	(K)	(cm−2)
B1-a	14.9[0.6]	1.3[0.2] × 1014	5.9[0.3]	8.4[1.9] × 1013	(26)	9[3] × 1013
SVS 4-5	20.2[0.9]	2.3[0.4] × 1014	5.9[0.3]	10[2] × 1013	(26)	2.2[0.5] × 1014
B5 IRS1	17[2]	4.3[0.9] × 1013	8[1]	2.1[0.7] × 1013	(26)	2.3[1.2] × 1013
IRAS 03235	18[4]	2.8[0.9] × 1013	(8)c	1.1[0.5] × 1013	(26)	1.2[1.1] × 1013
IRAS 04108	9[3]	1.2[0.5] × 1013	9[3]	1.2[0.5] × 1013	(26)	<8 × 1012
L1489 IRS	8[4]	5[5] × 1012	8[4]	5[5] × 1012	(26)	<1.3 × 1012

Notes. The listed uncertainties are the formal errors; the systematic column density uncertainties are estimated to ∼50% based on LVG analysis of B1-a. ^a26 K is the assumed excitation temperature for the warm CH₃OH component. ^bBased on the intensity or upper limit of the CH₃OH 95.169 GHz line. ^cNo excitation temperature could be independently derived and the sample average of 8 K was assumed.

Download table as:  ASCIITypeset image

Figure 6 shows that CH₃CHO and CH₃CN line intensities can be fit by a single excitation temperature. Based on this analysis, CH₃CHO is characterized by a low excitation temperature (8–9 K) consistent with the CH₃OH cold component. This does not rule out the existence of a warm CH₃CHO component, since upper limits of higher energy lines are inconclusive, but the detected CH₃CHO does not originate in a hot core. In contrast the CH₃CN excitation temperature of 26 K is suggestive of a warm, possibly a hot core, origin for CH₃CN. The HCOOCH₃ detections have low signal-to-noise ratio and also a small spread in energy levels and the derived excitation temperatures of 7–8 K are therefore highly uncertain. For CH₃OH as well as these COMs, the derived excitation temperatures depend on the kinetic temperatures. The two temperatures are rarely identical, however, and the relationship between them depend on the details of the excitation conditions and molecular excitation properties. Excitation temperatures cannot, thus, be used to determine specific emission locations if different molecules are unconstrained. Because CH₃CHO, CH₃CN, and CH₃OH have comparable dipole moments, however, their relative excitation temperatures constrain their relative excitation conditions. That is, based on the excitation temperatures, CH₃CN is more centrally peaked compared to CH₃CHO (see Section 4.3).

Zoom In Zoom Out Reset image size

Table 6 lists the resulting excitation temperatures and beam-averaged column densities, i.e., the reported column densities do not take into account potential beam dilution, which may be substantial for some or all molecules. For the detected COMs without independent constraints on their excitation, we used 8, 18, or 26 K. The choice for each molecule was based on recent observations of spatial and excitation temperature correlations of different molecules around high-mass protostars (Bisschop et al. 2007; Öberg et al. 2013; Fayolle et al. 2013). In these studies H₂CCO and CH₃CHO are cold and we therefore used the CH₃CHO excitation temperature of 8 K to determine the H₂CCO column. CH₃OCH₃ and CH₃CN emission were generally associated with only warm material and we used the CH₃CN excitation temperature of 26 K to determine the CH₃OCH₃ column. HNCO, like CH₃OH, has been observed to trace both cold and warm material and we therefore used an excitation temperature of 18 K, typical for the CH₃OH when fitting a single LTE component to the protostellar data. It is important to note that the excitation temperature choices are based on a small set of high-mass protostellar data and may need to be revised once more spatially resolved observations of complex molecules toward low-mass protostars exist.

We used the derived complex molecule column densities together with the derived CH₃OH column densities to calculate the abundances of organic molecules with respect to CH₃OH in our source sample. We tested the sensitivity of the derived abundances on the selected CH₃OH components for different molecules, and generally found that the choice of excitation temperature and CH₃OH reference column density (beam-averaged, cold or warm components) affected the derived abundances with respect to CH₃OH by less than a factor of two. Table 7 shows that the CH₃CHO abundances are between 7% and 9% and the upper limits between 5% and 20% with respect to CH₃OH. CH₃CN and CH₃OCH₃ abundances or upper limits could only be derived toward the first four sources since no lines associated with warm CH₃OH were detected toward two of the sources. The CH₃CN abundances are 0.8%–1.7% and the CH₃OCH₃ abundance toward SVS 4-5 is 10%. The HCOOCH₃ abundances are ∼10%, but abundances up to 40% cannot be excluded. The HNCO abundances vary between 1.0%–6.4%, and the H₂CCO abundances vary between 1.1%–9%. Derived upper limits are typically similar or higher compared to detections.

Table 7. Complex Molecule Abundances with Respect to CH₃OH in %

Source	x$_{\rm CH_3CHO}$	x$_{\rm CH_3CN}$	xHNCO	x$_{\rm H_2CCO}$	x$_{\rm CH_3OCH_3}$	x$_{\rm HCOOCH_3}$
B1-a	7[3]	1.3[0.8]	2.0[0.4]	1.4[0.3]	<7	11[19]
SVS 4-5	7[2]	0.8[0.4]	1.0[0.3]	1.1[0.2]	10[5]	10[30]
B5 IRS1	9[4]	1.7[1.3]	2.8[1.1]	<0.8	<35	<21
IRAS 03235	<6	<2.5	2.1[0.9]	2.0[1.5]	<42	<15
IRAS 04108	<5	–a	3.3[2.2]	2.4[1.4]	–a	<12
L1489 IRS	<20	–a	6.4[6.7]	9[9]	–a	<46

Note. ^aNo upper limit could be derived because of lack of detection of warm CH₃OH.

Download table as:  ASCIITypeset image

The analysis in the previous section assumes that all observed complex organic emission is optically thin. This is a tenuous assumption for CH₃OH, since CH₃OH lines are known to sometimes be optically thick toward protostars (e.g., Bisschop et al. 2007). Fortunately, the ¹³CH₃OH 2–1 ladder falls within the observed frequency range, enabling a direct test of this assumption. Two ¹³CH₃OH lines are detected toward B1-a and SVS 4-5, but not toward any other sources. The line intensity ratios between corresponding CH₃OH and ¹³CH₃OH lines are 57–60[15] and 71–76[19] toward B1-a and SVS 4–5, respectively, consistent with the solar ¹²C/¹³C ratio of 77. CH₃OH, and by inference all other COM, emission lines thus appear to be optically thin.

A second question is how the excitation temperatures, derived using rotational diagrams, relate to the kinetic temperatures in the emission regions. To explore this, we carried out a large velocity gradient (LVG) radiative transfer analysis using RADEX (van der Tak et al. 2007) for the molecules with both sufficient number of lines to constrain the excitation conditions and known collision cross sections, i.e., CH₃OH and CH₃CN, toward B1-a. B1-a is here taken to be representative of the sample because of the similar emission patterns of CH₃OH and CH₃CN (when detected) across the sample (Figures 5 and 6).

We ran a grid of RADEX models with n_H = [10⁵, 10⁶] cm⁻³, T_kin = [10, 25, 50, 100] K, and CH₃OH and CH₃CN column densities ranging between 0.5× and 2× those listed in Tables 2–4 for B1-a. The models with a factor of two lower and higher column densities compared to the rotational diagram values cannot be excluded for all temperatures and densities, but generally fit the data less well. This indicates that the rotational diagram method provides a good estimate of the beam-averaged column densities, but that the real uncertainty in the derived numbers are at least a factor of two rather than the formal uncertainties of 20%–50%.

Figure 7 shows the model results for the best-fit column densities (i.e., the ones derived using the rotational diagram method), and a density of 10⁶ cm⁻³. The best-fit temperature to the CH₃OH data, assuming a single component, is between 10 and 25 K, consistent with the rotational diagram excitation temperature. If a lower density of 10⁵ cm⁻² is assumed, kinetic temperatures up to 50 K fit the data. However, such a low average density is unlikely based on Jørgensen et al. (2002), where a detailed envelope modeling shows that the average density toward deeply embedded protostars on 1000 AU scales is ∼10⁶ cm⁻³. In either case, it can be excluded that most of the CH₃OH emission originates in a hot core. As in the rotational diagram analysis, no single temperature component fits all the CH₃OH lines. The two components derived in Table 5 reproduces the relative line intensities very well, but overpredicts the line intensities by about 50% as discussed above. In the case of CH₃CN, only kinetic temperatures of 25 K and higher are consistent with the data, confirming the rotational diagram results that CH₃CN has of a more centrally peaked origin compared to CH₃OH.

Zoom In Zoom Out Reset image size

In summary, the analyzed emission lines are optically thin and the rotational diagram method provides an accurate (within a factor of two) derivation of the molecular column densities. The excitation temperatures and the kinetic temperatures are not identical, but the observed difference in excitation temperatures for CH₃CN and CH₃OH corresponds to a real difference in kinetic temperatures for reasonable density assumptions.

In addition to the low-mass YSOs presented here, CH₃CHO, HCOOCH₃, CH₃CN, and/or CH₃OCH₃ have been previously quantified toward B1-b in Perseus (Öberg et al. 2009, 2010a), SMM1, SMM4, and SMM4-W in Serpens (Öberg et al. 2011b), L1157 (Arce et al. 2008), NGC 1333 IRAS 2A, 4A, and 4B in Perseus (Bottinelli et al. 2004a, 2007), and toward IRAS 16293−2422, both including the binary in one beam, and toward the A and B cores separately (Cazaux et al. 2003; Bisschop et al. 2008). The abundance frequency histograms with respect to CH₃OH are shown in Figure 8 for the combined low-mass sample of literature sources and our six YSOs. Complex molecule abundances above 1% are clearly common. The median detected abundances for the four species HCOOCH₃, CH₃CHO, CH₃OCH₃, and CH₃CN varies between 1% (CH₃CN) and 5% (HCOOCH₃ and CH₃OCH₃) with respect to CH₃OH. It is important to note that most upper limits seems to be consistent with these values and thus that the lack of complex molecules in half of our sources is most likely due to an overall low CH₃OH abundance or a low overall ice desorption rate rather than a significantly different chemistry with respect to CH₃OH.

Zoom In Zoom Out Reset image size

Figure 8 also shows that the COM abundances vary significantly between sources. Disregarding upper limits CH₃CN/CH₃OH = 0.07%–1.7%, HCOOCH₃/CH₃OH = 0.6%–56%, CH₃OCH₃/CH₃OH = 0.8%–20% and CH₃CHO/CH₃OH = 0.1%–9%. For all molecules there is thus a one or two order of magnitude variation across the sample.