A SURVEY OF MOLECULAR LINES TOWARD MASSIVE CLUMPS IN EARLY EVOLUTIONARY STAGES OF HIGH-MASS STAR FORMATION

Figure 1 shows the spectra of the observed MSX dark sources. In Figure 1, the top trace of each panel shows the CH₃OH spectrum, where three K components of the J = 2–1 transition can be seen. The three components correspond to the J_K = 2₀–1₀ E, 2₀–1₀ A⁺, and 2₋₁–1₋₁ E from left to right. Hereafter we call these three lines as J = 2–1 triplet. In this figure, we also present the CH₃OH J_K = 2₁–1₁ A⁻ line, whose upper state energy (22 K) is slightly higher than those of the J_K = 2₀–1₀ E, 2₀–1₀ A⁺, and 2₋₁–1₋₁ E lines (5–9 K).

Zoom In Zoom Out Reset image size

The SiO line, which is a good shock tracer, was detected toward all the observed MSX dark sources. The detection rate is higher than that reported by Beuther & Sridharan (2007, ∼40 %). This may originate from the difference in the source selection criteria. We select the MSX dark sources toward which the CH₃OH J = 7–6 line is detected, while Beuther & Sridharan observed almost all the MSX dark sources found by the bolometer array observation (Beuther et al. 2002a). Our observation therefore seems to be biased on active MSX dark sources. In fact, the Spitzer 24 μm sources are associated with all our observed sources.

The SiO lines look broader than the other molecular lines. The broad linewidth would reflect shocks due to molecular outflows. According to Chambers et al. (2009), the "green fuzzy" feature, which is a 4.5 μm excess due to the shock-excited lines, such as H₂ 0–0 S(9) and CO ν = 1–0, is seen toward most of our observed MSX dark sources. This supports that shocks are occurring in the observed sources.

The C³⁴S peak intensities are generally weaker than the H¹³CO⁺ peak intensities, except toward G23.60+00.00 MM2. Toward G34.43+00.24 MM3, the C³⁴S line is clearly broader than the H¹³CO⁺ line. This indicates that the C³⁴S line comes from relatively active regions.

The H¹³CO⁺, HN¹³C, and CCH lines are relatively narrow, showing a Gaussian shape. The H¹³CO⁺ and HN¹³C line shapes are very similar to each other. This suggests that they come from relatively quiescent regions. The OCS and SO emissions are not detected toward all the MSX dark sources; very weak OCS emission can be seen toward a few sources.

Figure 2 shows the spectra observed toward the MSX sources. We have detected strong CH₃OH emission toward five sources; G34.43+00.24 MM2, I18089-1732 MM1, I18182-1433 MM1, I18264-1152 MM1, and I18337-0743 MM1. It is reported that hot molecular cores are associated with I18089-1732 MM1 (Beuther et al. 2004, 2005b) and I18182-1433 MM1 (Beuther et al. 2006). In the five sources, the other observed molecular line emissions are also strong, although there is a variation in the intensities among them. This variation suggests that the chemical compositions are slightly different from source to source.

Zoom In Zoom Out Reset image size

The CH₃OH emission is relatively weak toward four sources; I18151-1208 MM1, I18223-1243 MM1, I18272-1217 MM1, and I18385-0512 MM1. However, strong H¹³CO⁺ and CCH emissions are detected toward I18151-1208 MM1 and I18223-1243 MM1, indicating that a large amount of dense gas exists there. On the other hand, all the emissions are weak toward I18272-1217 MM1 and I18385-0512 MM1, although their CCH intensities are comparable to those of the MSX dark sources. For these two sources, most of the dense gas has been dissipated by various star formation activities. Hence, they are more evolved than the others, as discussed by Beuther et al. (2002a) on the basis of their density profiles. In fact, the dust temperatures of I18272-1217 MM1 and I18385-0512 MM1 are 52 K and 49 K, respectively, being higher than those of the others. Furthermore, the centimeter-wave continuum emission is relatively strong toward these two objects (Sridharan et al. 2002).

Figure 3 shows the spectra toward the other clumps. Strong CH₃OH lines were detected toward all the four sources. All the molecular lines are strong toward G34.43+00.24 MM1, where even the OCS and SO lines are clearly detected. It is known that there is a hot core containing a high-mass protostar (Rathborne et al. 2008), and hence, the strong emissions toward G34.43+00.24 MM1 seem to be related to the hot core activity.

Zoom In Zoom Out Reset image size

The SiO line was detected toward all the sources in this category. All the SiO lines show a broad feature, except for G34.43+00.24 MM4. Even for G34.43+00.24 MM4, a weak broad emission could be seen. The C³⁴S line shapes are different from one another among the four sources. G34.43+00.24 MM1 and G24.60+00.08 MM1 show a broad feature of the C³⁴S line. On the other hand, the C³⁴S lines observed toward G34.43+00.24 MM4 and I18102-1800 MM1 have a narrow line shape that is similar to the H¹³CO⁺ line shape. These differences may reflect the difference in evolutionary stages, as discussed in a later section.

The SiO molecule is known to be absent in cold quiescent clouds like TMC-1, because Si is almost depleted onto dust grains (e.g., Ziurys et al. 1989). When a shock occurs in a cloud, SiO is released into the gas phase by disruption of silicate grains through sputtering and grain–grain collisions (e.g., Seab & Shull 1983; Schilke et al. 1997). Alternatively, the Si atom is ejected into the gas phase, which reacts with O₂ or OH to produce SiO (e.g., Caselli et al. 1997; Schilke et al. 1997). Thus, SiO is generally abundant in a shocked region caused by an interaction between an outflow and a surrounding medium (Mikami et al. 1992; Bachiller & Pérez Gutiérrez 1997).

In Figure 8, it is remarkable that the N(SiO)/N(H¹³CO⁺) ratios of the MSX dark sources are higher than those of the MSX sources. This indicates that the fraction of shocked gas is lower in the MSX sources than in the MSX dark sources, although the observed MSX sources have powerful outflows (Beuther et al. 2002b). Since the H¹³CO⁺ integrated intensity tends to be higher in the MSX sources (Figure 6), the lower N(SiO)/N(H¹³CO⁺) ratios in the MSX sources could partly be due to a large amount of dense gas. However, it seems more likely that the SiO abundance decreases in the MSX sources, because the SiO integrated intensities of some MSX sources are actually much lower than those of the MSX dark sources (Figure 6). Motte et al. (2007) also indicate that the SiO integrated intensities of the infrared-quiet cores are higher than those of the luminous infrared sources (see Figure 8 in Motte et al. 2007).

As mentioned above, a strong shock is necessary for the SiO production in the gas phase. In MSX dark sources, which are in early stages of star formation, a dense gas would still surround the protostars. Hence a strong shock can occur by an impact of an outflow on the ambient dense gas. On the other hand, the outflow cavity may have been relatively large in evolved sources, like MSX sources, so that a strong shock does not occur frequently in the vicinity of the protostars. Thus, SiO cannot be provided efficiently into the gas phase in the evolved sources. Furthermore, it has been suggested that the SiO abundance in the gas phase would decrease with time by depletion onto dust grains or the gas-phase destruction with OH, where the timescale of destruction is estimated to be about 10⁴ yr (Pineau des Forets et al. 1997). Since the lifetime of SiO is shorter than the dynamical timescale of the outflows of the MSX sources (several times 10⁴ yr; Beuther et al. 2002b), SiO tends to be less abundant in the MSX sources. On the basis of the SiO observations of 15 high-mass star forming cores, Miettinen et al. (2006) suggest that the SiO abundance decreases with rising temperature. This can also be interpreted as the SiO abundance decreasing with time.

Codella et al. (1999) pointed out that the SiO line profile reflects evolutionary stages; the SiO line has a large velocity width due to shock in the early stage, while it becomes narrower in the late stage as the shock slows down. In our data, the SiO line width tends to be narrower toward the MSX sources than toward the MSX dark sources. Thus, the SiO emission from the MSX sources would trace relatively old shocks, whereas it would mainly trace newly formed shocks in the MSX dark sources.

In Figure 8, the N(CH₃OH)/N(H¹³CO⁺) ratios of the MSX dark sources tend to be higher than those of the MSX sources, and we can see a correlation with N(SiO)/N(H¹³CO⁺), where the correlation coefficient is 0.77. The high abundance of CH₃OH in the MSX dark sources is consistent with the results by Leurini et al. (2007b). In contrast, Beuther & Sridharan (2007) suggest that the CH₃OH abundance of IRDCs is as low as that of low-mass star forming cores. This difference may be due to the difference in the source selection criteria; we observed rather active MSX dark sources, as mentioned earlier.

The correlation between the CH₃OH and SiO abundances means that the origin of CH₃OH in the MSX dark sources is related to shocks, as in the case of SiO. This is further supported by the result that the CH₃OH velocity widths of the MSX dark sources are larger than those of the MSX sources. In Figure 5(a), it can be seen that the CH₃OH linewidth is narrower than the SiO linewidth. This is because CH₃OH can be evaporated from icy grain mantles by moderate shocks (Bachiller & Pérez Gutiérrez 1997), whereas the SiO production requires stronger shocks to disrupt silicate grains (Mikami et al. 1992).

On the other hand, the CH₃OH velocity widths are correlated with the H¹³CO⁺ velocity widths in the MSX sources, as mentioned in Section 3.2. This may be because CH₃OH traces old shocks as in the case of SiO in the MSX sources. In addition, CH₃OH in the MSX sources could be evaporated from grain mantles by radiation from protostars, as pointed out in Paper I. In fact, Beuther et al. (2007) and Leurini et al. (2007a) suggest that the CH₃OH lines trace a rotating disk toward several HMPOs. However, the CH₃OH emission from such hot regions is estimated to have a compact distribution of a few arcseconds (Leurini et al. 2007b) and is diluted by the telescope beam. Thus, the N(CH₃OH)/N(H¹³CO⁺) ratio of the MSX sources is lower than that of the MSX dark sources.

The N(C³⁴S)/N(H¹³CO⁺) ratio is also roughly correlated with the N(CH₃OH)/N(H¹³CO⁺) ratio with a correlation coefficient of 0.58, as shown in Figure 8. This, along with the rough correlations between the C³⁴S and CH₃OH velocity widths (Figure 5), suggests that the origin of CS is similar to that of CH₃OH. The origin of CS would be shock evaporation from grain mantles in the MSX dark sources, while it could be related to radiation from protostars in the MSX sources. It should be noted that the enhancement of the CS abundance as well as the large velocity width of its spectral lines is reported for the shocked region of L1157 (Mikami et al. 1992; Bachiller & Pérez Gutiérrez 1997).

However, the N(C³⁴S)/N(H¹³CO⁺) ratio tends to be slightly higher toward the MSX sources than toward the MSX dark sources at a given value of N(CH₃OH)/N(H¹³CO⁺). Marseille et al. (2008) also suggest that CS is more abundant in an MSX source than in an MSX dark source in the I18151-1208 region. This may be due to the difference in evaporation temperature between CS and CH₃OH. Since the evaporation temperature of CS (∼30 K) is lower than that of CH₃OH (∼100 K; Aikawa et al. 1997), CS can be abundant in a relatively large area by radiation heating as compared with CH₃OH. Therefore, the beam filling factor of the C³⁴S emission could be larger than that of CH₃OH in the MSX sources. Although Beuther et al. (2008, 2009) suggest from higher resolution observations that the CS abundance decreases near protostars in HMPOs, such area is as small as a few arcsec. Hence, this effect can be ignored in our single-dish observation.

As shown in Figure 8, the N(HN¹³C)/N(H¹³CO⁺) and N(CCH)/N(H¹³CO⁺) ratios are not well correlated with the N(CH₃OH)/N(H¹³CO⁺) ratio, where the correlation coefficient is 0.40 and 0.29, respectively. In addition, no systematic difference can be seen between the MSX dark sources and the MSX sources. Since the HN¹³C and CCH velocity widths are as narrow as the H¹³CO⁺ velocity width, the HN¹³C and CCH emissions do not trace shocked regions, but come from quiescent dense gas. The absence of correlations with the CH₃OH abundance implies that HN¹³C and CCH abundances in the clumps are not enhanced significantly by star formation activities.

Beuther et al. (2008) present that the distribution of CCH emission shows a small hole of a few arcsec around a hot core, and they suggest that CCH would decrease in the hot core phase. Although the CCH emission mainly traces dense gas as mentioned above, CCH may preferentially exist in the outer part of the clump, as Beuther et al. suggest. Thus, the CCH abundance in the envelope would not change significantly, even if the CCH abundance near protostars would decrease.

In addition, it is known that CCH can be abundant in a photodissociation region including relatively diffuse interclump medium irradiated by UV radiation (Fuente et al. 1993). Toward I18272-1217 MM1 and I18385-0512 MM1, the CCH emissions are relatively strong, as compared with the other emissions. Since the centimeter continuum emission is relatively strong toward I18272-1217 MM1 and I18385-0512 MM1, the CCH emission may trace such diffuse interclump medium in these sources.

As shown in Figure 6, the OCS and SO emissions are weak toward all the MSX dark sources. The weak emissions may partly be due to the relatively high upper state energies of the OCS and SO lines (21.0 K and 19.3 K, respectively). Since the size of the hot regions would be as small as a few arcsec, as mentioned above, the beam dilution effect could be serious for these lines. Since OCS is thought to be a main reservoir of S atoms on the dust grains (van der Tak et al. 2003), OCS could be evaporated directly from dust grains by shocks and radiation heating. Thus, the origin of OCS might be similar to CH₃OH.

The SO intensities of several MSX sources are clearly higher than those of the MSX dark sources. This may indicate that SO could efficiently form in a hot gas phase. van der Tak et al. (2003) also suggest by observations toward high-mass star-forming regions that the SO emission mainly comes from hot cores rather than from outflows. Their result is consistent with ours.