An Ammonia Spectral Map of the L1495-B218 Filaments in the Taurus Molecular Cloud. II. CCS and HC₇N Chemistry and Three Modes of Star Formation in the Filaments

We probe the kinematics of dense core L1495A-N using spectral maps of CCS 2₁−1₀ and NH₃ (1,1) and investigate the motions of filaments and dense cores traced by the CCS 2₁−1₀ emission. Figure 7 presents channel maps of the CCS 2₁−1₀ emission along with the integrated intensity of NH₃ (1,1) in the B7 region. The LSR velocities of NH₃ dense cores are 7.2 and 7.25 km s⁻¹ at their centers. The first panel in Figure 7 shows the blueshifted CCS emission relative to the NH₃ (1,1) emission in L1495A-N, whereas the two right panels present the redshifted CCS 2₁−1₀ emission relative to the NH₃ (1,1) emission in L1495A-N. In the first panel, the blueshifted CCS 2₁−1₀ emission is bright in the northeastern part of L1495A-N and elongated similar to the NH₃ core. With increasing LSR velocity, we find that CCS 2₁−1₀ emission moves toward the southeast direction, which indicates that L1495A-N has shear flows, or rotation, or converging flows.

Zoom In Zoom Out Reset image size

We investigate the flow of gas using the LSR velocity maps of CCS 2₁−1₀ and NH₃ (1,1) (Figure 8). We find a steep transition in the LSR velocities of both CCS 2₁−1₀ and NH₃ (1,1) emission within L1495A-N (pointed with black arrow). The transition shows a velocity shift of 0.3 km s⁻¹ within the size of beam (0.022 pc) and the steepest velocity gradient of ∼14 km s⁻¹ pc⁻¹ in both NH₃ (1,1) and CCS 2₁−1₀, which is at least seven times larger than the typical velocity gradient measured for rotation (∼2 km s⁻¹ pc⁻¹, Goodman et al. 1993). Also, the transition is parallel to the long axis of L1495A-N rather than its short axis, which is a rare case. Only L183 is reported to have the rotation axis parallel to its long axis (Kirk et al. 2009). However, the velocity gradient of L1495A-N is considerably larger than that of L183 (8.9 km s⁻¹ pc⁻¹). Thus, an unusually large gradient of the LSR velocity in L1495A-N may be due to a converging flow or a shear flow, but not all of it is due to rotation.

Zoom In Zoom Out Reset image size

To check whether the flow in L1495A-N is a converging flow, we observed L1495A-N in HCN J = 1−0 and HCO⁺ J = 1−0 using the 12 m ARO and 100 m GBT. The beam of the 12 m ARO covers most of L1495A-N; thus, the HCN 1−0 and HCO⁺ 1−0 observations using the 12 m ARO show the overall internal dynamics of L1495A-N. We found that L1495A-N has strong blue asymmetric profiles in both HCO⁺ and HCN J = 1−0 (Figures 9 and 10), whereas optically thin CCS 2₁−1₀ and NH₃ (1,1) lines are well-fitted by a single Gaussian profile. To see whether the blue asymmetric profiles are due to self-absorption of a converging motion, we estimated the optical depths of HCN and HCO⁺ 1−0 lines through our chemical model followed by LTE-radiative-transfer model (Majumdar et al. 2017a) and found their optical depths to be larger than 2 in L1495A-N. Also, ¹³CO 1−0 has two peaks at 6 and 7.2 km s⁻¹, which do not coincide the two intensity peaks of HCN and HCO⁺ 1−0. These results indicate that the blue asymmetric profiles of HCN and HCO⁺ 1−0 are not due to two different velocity components along the line of sight but are the results of self-absorption, which is often observed in collapsing dense cores (e.g., Lee et al. 1999; Sohn et al. 2007).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

All hyperfine components of HCN J = 1−0 observed using the ARO show blue asymmetric line profiles. The converging-flow velocities estimated using a simple two-layer model (Equation (9) in Myers et al. 1996) are 0.41 km s⁻¹ for F = 2−1, 0.17 km s⁻¹ for F = 1−1, and 0.10 km s⁻¹ for F = 0−1. The converging-flow speed measured in HCO⁺ 1−0 is 0.83 km s⁻¹, which is roughly four times larger than the sound speed. This converging-flow speed is higher than the maximum infall speed in a similarity solution of a gravitationally collapsing core (Larson 1969; Penston 1969) and is similar to that formed in the collapse of dense core triggered by a supersonic converging flow (Gong & Ostriker 2009).

Sparsely mapped HCN and HCO⁺ using Argus on the GBT shows the dynamics of L1495A-N at higher angular resolution (Figure 10). Although the 12 m ARO data show overall dynamics of L1495A-N, the GBT data show spatially resolved dynamics within L1495A-N and its surroundings because the beam size of GBT is much smaller than that of L1495A-N. Dense cores mapped previously in HCN 1−0 and HCO⁺ 1−0, such as L1544, L694-2, and L1197, all have strongest emission and largest infall speed (converging-flow speed due to gravitational collapse) at or near the core center and show a roughly azimuthally symmetric distribution of emission. On the contrary, L1495A-N has highly asymmetric distributions of HCN and HCO⁺ intensities and converging-flow speeds. The brightest emission from HCN and HCO⁺ is found on the northeast outskirts of the core rather than its center. Moreover, the converging-flow speeds are also the largest at northeast outskirts (0.9 km s⁻¹ or Mach 4.5), whereas the west side of L1495A-N shows almost no converging flow (<0.05 km s⁻¹). The quasi-static collapse models of a gas sphere and filaments typically have the maximum infall velocity less than Mach 4, and the fastest infall occurs very close to the core center or the axis of the filament (<0.01 pc) at the final stage of the collapse (Foster & Chevalier 1993; Myers 2005; Gong & Ostriker 2009; Seo et al. 2013; Burge et al. 2016). The high converging-flow speed only at the outskirts and the large degree of asymmetry seen in L1495A-N do not agree with the quasi-static collapse models and suggest that large-scale gas flows are likely driving the dynamics of L1495A-N.

To see whether the converging flow observed in L1495A-N in HCN 1−0 and HCO⁺ 1−0 is connected with a large-scale flow, we examine ¹³CO 1−0 integrated intensity and channel maps along with the integrated intensity map of NH₃ (1,1) in Figure 11. The ¹³CO 1−0 is observed using the 14 m FCRAO with a spatial resolution of 47'' and a velocity resolution of 0.256 km s⁻¹ (Goldsmith et al. 2008). We found that there are roughly three different groups of ¹³CO structures radiating from L1495 in the integrated intensity map and found in the channel maps that the hub region may be a place where the three groups of ¹³CO structures are colliding or converging (upper panel of Figure 11). In Figure 11, we show the three groups of CO structures with the velocity ranges chosen to emphasize each group of ¹³CO structures. One group of ¹³CO structures (Group I at 6.28–6.54 km s⁻¹) stretches from the hub to B10, B213, and B216 and forms the L1495-B218 filaments. Another group of ¹³CO structures (Group II at 6.79–7.05 km s⁻¹) is distributed from the hub and extends to a southwest direction. The last group of ¹³CO structures (Group III at 7.55–7.81 km s⁻¹) stretches from the hub to the west direction. We also present channel maps in a full velocity range from 6.28 to 7.81 km s⁻¹ in Figure 12 to probe the hub and the three groups in the Position–Position–Velocity (PPV) space. The ¹³CO 1−0 emission from 7.05 to 7.55 km s⁻¹ is the CO emission from the hub. Group III connects to the hub from redshifted velocity and has the velocity gradient of roughly −0.8 km s⁻¹ pc⁻¹ from the Group I to the hub. On the other hand, Groups I and II are at blueshifted velocities and bifurcate from the hub with the velocity gradients of 0.6 km s⁻¹ pc⁻¹ and 2 km s⁻¹ pc⁻¹, respectively. The velocity gradients are the median values of the velocity gradients measured following 1 pc distance along the three groups of ¹³CO structures from L1495A-N (Figure 13). Groups II and III seem to be converging toward the same location in the hub, whereas Group I connects to a position slightly to the south. The fact that the velocity gradients of the three groups are toward the hub with signatures of converging motions in the HCN 1−0 and HCO⁺ 1−0 observations and that the hub is the most massive and gravitational unstable part of the region on the basis of a virial calculation suggest that the velocity gradients of the three groups are likely due to accretion of gas on the hub along the filaments (Seo et al. 2015).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

We found that L1495A-N is embedded within the hub in PPV space and exactly located where Groups II and III interface with each other in a zoomed view of L1495A-N with ¹³CO channel maps (bottom panel of Figure 11). The LSR velocity of ¹³CO 1−0 emission also connects seamlessly with the LSR velocity of NH₃ and CCS emission, and the difference in the LSR velocities between the Groups II and III is consistent with the speed of converging gas flows measured in HCO⁺, which suggests that L1495A-N and the large-scale ¹³CO gas are connected in the PPV space. Having a coherent gas flow from parsec to subparsec scale suggests that the formation and evolution of the L1495A-N core are not determined solely by local dynamics but may be driven by large-scale gas flows.

We probe the kinematics of the B213E region containing dense cores L1521D, No.30, No.31, and No.32 using channel maps of CCS and NH₃ (Figure 14). The NH₃ LSR velocity of L1521D is 6.75 km s⁻¹ at its center. The blueshifted CCS emission is bright and covers the outskirts of L1521D except for the south side of the core (upper three panels). The redshifted CCS emission is bright at the center and the west side of L1521D (the bottom three panels). The last two panels in Figure 14 show CCS emission near the west side of B213E where there are three NH₃ dense cores (Seo et al. 2015). It shows that the CCS intensity peak coincides with the NH₃ intensity peak of the NH₃ core No.32 in the channel map.

Zoom In Zoom Out Reset image size

We further investigate the kinematics using the LSR velocity map of B213E in Figure 15. In L1521D, the velocity gradient within L1521D is 3 km s⁻¹ pc⁻¹ in both NH₃ and CCS, which is similar to the average velocity gradient of fast rotating dense cores (Goodman et al. 1993). In B213E, the LSR velocity changes steeply from the dense core No.31 to the dense core No.32 with 13 km s⁻¹ pc⁻¹, which is a considerably larger gradient than a fast rotating core. The transition may be due to either the two cores being at different LSR velocities that overlap at their outskirts or a recent collision with each other.

Zoom In Zoom Out Reset image size

To see whether L1521D and dense cores No.30, No.31, and No.32 have converging motions, we observed them in HCN 1−0 and HCO⁺ 1−0 using the 12 m ARO and the 100 m GBT with Argus. We found that only L1521D shows blue asymmetric profiles in HCO⁺ 1−0 and HCN 1−0 (Figures 9, 16). Using a radiative transfer calculation, we evaluated the optical depths of the observed HCN and HCO⁺ 1−0 lines and found that they are ≥2, which is similar to the optical depths of other dense cores in HCN and HCO⁺ 1−0 (De Vries & Myers 2005; Williams et al. 2006; Campbell et al. 2016). This confirms that the blue asymmetric profiles of HCN 1−0 and HCO⁺ 1−0 in L1521D are the self-absorbed features due to a converging motion. The HCN J = 1−0 emission of L1521D observed using the 12 m ARO is relatively weak compared to HCO⁺ J = 1−0 emission (Figure 9). On the other hand, HCO⁺ J = 1−0 from the 12 m ARO has bright emission (>1 K) with a blue asymmetric, self-absorbed profile. The estimated converging-motion speed using HCO⁺ 1−0 is 0.08 km s⁻¹ using the two-layer model (Myers et al. 1996), which is a bit less than half of the sound speed and substantially slower than the speeds measured in L1495A-N.

Zoom In Zoom Out Reset image size

Observations toward L1521D using the GBT show internal dynamics similar to those found using the 12 m ARO. The position of the observation is chosen to cover the outskirts and the inner region of L1521D. L1521D shows converging-motion signatures on both east and west sides with converging-motion speeds in the range 0.05–0.1 km s⁻¹. HCN and HCO⁺ 1−0 intensities are stronger on the east side of the core, similar to CCS emission, which may be due to an asymmetry in the chemistry or the excitation. The converging-motion speeds measured in HCN and HCO⁺ 1−0 are similar on both sides of the core, and no significant asymmetry in the internal dynamics is found. As we shall discuss in Section 4, in terms of the kinematics, L1521D is similar to a slowly contracting core because of self-gravity. As for dense core No.32, mapping using the GBT shows complex line profiles and shows that there are at least three velocity components, but there is no suggestion of a significant converging motion.

To see whether the gas motions in L1521D are connected with large-scale flows, we investigate ¹³CO and ¹²CO channel maps together with NH₃ and dust structures in B213E in Figure 17. The velocity ranges of ¹³CO and ¹²CO channel maps are chosen to highlight the CO flows, so the redshifted and blueshifted components are not isolated structures but continuous flows in the PPV space. In the region close to L1521D, the ¹³CO 1−0 emission is redshifted in the south of L1521D and is blueshifted in the north of L1521D. On the other hand, on a large scale, the ¹²CO 1−0 emission is redshifted in the northwest of L1521D and is blueshifted in the south of L1521D. The CCS and NH₃ flows in L1521D do not have the same direction as the ¹²CO flows but are seamlessly connected with ¹³CO in the PPV space. Hence, the gas flow in L1521D is possibly a local flow restricted by the filaments but not connected to the large-scale flows that follow the striations seen in dust continuum emission (Palmeirim et al. 2013). To the west side of B213E, where three NH₃ cores (No.30, No.31, and No.32) are located, the LSR velocity of CCS changes from the southeast to the northwest direction, a shift that is aligned with the LSR velocities of both ¹²CO and ¹³CO.

Zoom In Zoom Out Reset image size

We investigate the kinematics in the B216 region using the channel maps of CCS compared to the integrated NH₃ (1,1) and dust continuum at 500 μm (Figure 18). The CCS emission in B216 spans 6.10–6.87 km s⁻¹. The CCS intensity peak does not coincide with NH₃ or dust continuum intensity peaks but is centered 0.08 pc to the southwest from the NH₃ peaks and dust continuum peaks. The two NH₃ dense cores, No.35 and No.36, have LSR velocities of 6.66 and 6.76 km s⁻¹. CCS emission only at 6.6–6.8 km s⁻¹ is spatially associated with NH₃ cores and dust continuum peak. The CCS peak (also known as L1521B) was previously observed by Hirota et al. (2004) as an example of a CCS-bright young core. Our observations indicate that CCS does not trace the dense core in this region.

Zoom In Zoom Out Reset image size

The CCS-bright "core" predicted by Suzuki et al. (1992) is rarely found. In surveys of the Taurus molecular cloud (Suzuki et al. 1992; Hirota et al. 2009), CCS is detected in 18 of 29 cores, but only dense core L1521E is confirmed to have a CCS emission peak coinciding with the dust continuum peak (Hirota et al. 2002; Tafalla & Santiago 2004), which indicates that the lifetime of the CCS peak is, at most, 0.04 Myr (the lifetime of a dense core is roughly 1 Myr at the core mean density of ${\bar{n}}_{{{\rm{H}}}_{2}}={10}^{4}\,{\mathrm{cm}}^{-3}$; André et al. 2014). The frequency of finding CCS-bright cores would be higher if Suzuki's chemical model were true for all cores, because CCS is abundant up to 1 Myr in Suzuki's model.

We find that CCS is rarely bright at the center of the cores traced by NH₃ or dust continuum, but rather peaks on the core outskirts or outside the cores. Only one core (No.32) of 39 NH₃ cores along the L1495-B218 filaments has a good agreement among the dust, NH₃, and CCS intensity peaks. Thus, there are only two dense cores (No.32 and L1521E) in the Taurus molecular cloud known to have coincident CCS and dust continuum intensity peaks. In particular, CCS emission in B216, which contains only young dense cores with no star formation activity, clearly shows that CCS intensity peaks do not coincide with NH₃ or dust continuum peaks (L1521B), even at an early evolutionary stage. This suggests that CCS is not a good tracer with which to trace core density structure.

We found that CCS emission is detected in both less-evolved regions (B216 and B211) and the more-evolved regions with star formation activity (L1495A/B7N and B213E), whereas we did not detect any significant CCS emission in B213W and B10, which seem to be at evolutionary stages slightly younger or similar to those of L1495A/B7N and B213E (Seo et al. 2015). Comparing the column density among the starless cores, we find that two more-condensed starless cores (e.g., L1521D and L1495A-N) have the brightest CCS and HC₇N emission, whereas some other starless cores at the similar column density (e.g., starless cores in B213W) have no significant CCS and HC₇N detection at our noise level. We also find that the less-condensed starless cores and filaments in our survey (e.g., starless cores in B216 and filaments in B211) have CCS emission but no HC₇N emission. In the moderately condensed starless cores in B10, we could not detect significant CCS or HC₇N emission. In summary, our CCS survey toward the L1495-B218 filaments shows that there is significant CCS emission in the less-condensed starless cores and filaments but also occasionally in more-condensed starless cores. This is not consistent with the prediction of the Suzuki model, which predicts that most of the starless cores in Taurus should have significant CCS emission.

Aikawa et al. (2001) made a chemical model of a dense core considering grain-surface chemistry and collapse dynamics. They demonstrated that the CCS abundance is more sensitive to the absolute age of the cores than to the gas density, suggesting that a more-condensed core can have a high CCS abundance if it is young. For example, L1544 and L1689B have similar densities and show signatures of converging-motion, but L1689B has considerably brighter CCS emission than does L1544. This may be because L1689B may be younger or has accreted surrounding gas more recently than has L1544 (Lee et al. 2003; Sohn et al. 2007; Seo et al. 2013). Although Aikawa's model suggests a possibility of having high CCS abundance in more-condensed starless cores, it still predicts considerably more CCS-bright cores than our observations and is relatively similar to Suzuki's model. This may be because the physical and chemical conditions of L1495-B218 filaments are different from Aikawa's model and because their chemical network including both gas and grain-surface chemistry is relatively simple. In this study, we evaluate chemical evolution in the L1495-B218 filaments using a state-of-the-art astrochemistry code using the observed physical conditions in Taurus.

We modeled the chemistry in dense cores in Taurus using the state-of-the-art chemical code NAUTILUS (Ruaud et al. 2016). It simulates a three-phase chemistry, including gas-phase, grain-surface, and grain-mantle chemistry, where both surface and the mantle are chemically active with possible exchanges between the different phases. Adsorption of gas-phase species onto grain surfaces, the thermal and nonthermal desorption of species from the grain surface into the gas phase, and species from the surface to mantle and mantle to the surface are the primary exchange processes. This chemical code considers the latest updates on various physico-chemical processes from the KIDA database¹¹ which is more advanced than previous studies reported in Suzuki et al. (1992) and Aikawa et al. (2001). NAUTILUS has been tested, benchmarked, and applied to multiple cases, including dense cores (Majumdar et al. 2017b; Vidal et al. 2017), low-mass protostellar envelopes (Majumdar et al. 2017a, 2018a), hot cores (Vidal & Wakelam 2018), and protoplanetary disks (Wakelam et al. 2016). The gas-phase chemistry of NAUTILUS is based on the public chemical network kida.uva.2014 (Wakelam et al. 2015) with updates for sulfur chemistry (Vidal et al. 2017) and the chemistry of complex organics (Majumdar et al. 2018b) from the KIDA database, whereas surface reactions are based on Garrod (2008) with additional updates from Ruaud et al. (2016), Vidal et al. (2017), and Majumdar et al. (2018b). Using NAUTILUS, we explored a wide range of parameters, including density, temperature, visual extinction, and initial elemental abundances to model the chemistry in different environments. We limited excessive complexity of three-phase physio-chemical modeling by assuming that there is no inflow to the cores.

We use the following model parameters to represent the chemical evolution of CCS, NH₃, and HC₇N in the majority of starless cores in the L1495-B218 filaments.

• 1.  
  High-density core: T = 8 K; ${n}_{{{\rm{H}}}_{2}}=5\times {10}^{5}\,{\mathrm{cm}}^{-3}$; A_V = 20 mag
• 2.  
  Mid-density core: T = 10 K; ${n}_{{{\rm{H}}}_{2}}=5\times {10}^{4}\,{\mathrm{cm}}^{-3}$; A_V = 15 mag
• 3.  
  Low-density core: T = 10 K; ${n}_{{{\rm{H}}}_{2}}=2\times {10}^{4}\,{\mathrm{cm}}^{-3}$; A_V = 8 mag
• 4.  
  Core outskirts: T = 15 K; ${n}_{{{\rm{H}}}_{2}}=4\times {10}^{3}\,{\mathrm{cm}}^{-3}$; A_V = 3 mag.

The densities are chosen from the average density along the lines of sight to the starless cores in the L1495-B218 filaments. The extinction values are directly estimated from the Schmalzl et al. (2010) observations. The temperature values are adopted from the kinetic temperature of the starless cores estimated using NH₃ (Seo et al. 2015).

The selection of the set of initial gas-phase elemental abundances in the chemical models is not a trivial task because the abundances of the elements strongly influence the chemical state of the gas (Agúndez & Wakelam 2013). A prime example is the relative amount of carbon and oxygen, which is known to strongly affect the chemistry (Le Bourlot et al. 1995; van Dishoeck & Blake 1998). To reflect the large uncertainties in the gas-phase elemental abundances of carbon and oxygen, the range of carbon-to-oxygen gas-phase elemental abundance ratio (noted C/O hereafter) is often varied between 0.3 and 1.5 (Cartledge et al. 2004; Le Gal et al. 2014). The elemental gas-phase abundance of sulfur is also very poorly constrained and we have yet to locate the main reservoir of sulfur in the interstellar matter (ISM) (Majumdar et al. 2017a; Vidal & Wakelam 2018). The range of initial sulfur abundances can vary by almost three orders of magnitude if we consider the range determined by the so-called low metal and high metal abundances defined by Graedel et al. (1982).

Here, we first show chemical models with an initial C/O elemental ratio of 0.35, which is close to the solar value (Le Gal et al. 2014). The initial sulfur abundance in the models is 5 × 10⁻⁷ with respect to hydrogen nuclei, whereas the solar value is 1.5 × 10⁻⁵. Thus, we have applied a depletion factor (defined by the ratio of solar to adopted value in the model) of 30 from the gas phase, which is smaller than the best-fit depletion factor of 200 adopted to reproduce the observations of different sulfur-bearing species in L1544 prestellar core by Vastel et al. (2018). They also discussed the variation of sulfur depletion for different sources. We find that the initial sulfur abundance of 5 × 10⁻⁷ results in the best fits to the observed fractional abundances in the L1495-B218 filaments.

Figure 19 shows the fractional abundance profiles of NH₃, CCS, and HC₇N as a function of time. By running a grid of chemical models, we find that the chemical models with C/O = 0.35 provide the best fit to our observations, except for L1495A-N and L1521D, because the models and observations have good agreement within a factor of five. We consider a factor of five as the reasonable uncertainty between the models and the observations because the uncertainties of the observed fractional abundances are typically a factor of three. The uncertainties of the NH₃ fractional abundance are mostly in the H₂ column density because of uncertainty in the dust properties, whereas the column densities of NH₃ have uncertainties typically lower than 30% as a result of a high S/N ratio. The column densities of HC₇N and CCS depend on the excitation temperature. If the excitation temperature is 6 K, the column density will be 30% higher than a case when T_k = T_ex, whereas it can be three times higher if T_ex = 4 K. Because we have not observed other transitions of CCS and HC₇N, we could not measure the excitation temperature. A typical excitation temperature of CCS 2₁–1₀ measured in Taurus is around 6 K (Suzuki et al. 1992; Wolkovitch et al. 1997; Vastel et al. 2018). The excitation temperature of HC₇N 21−20 is not measured in Taurus.

Zoom In Zoom Out Reset image size

In this subsection, we will first discuss our chemical models with C/O = 0.35 and compare the models to our observations, except for the two dense cores, L1495A-N and L1521D, observed to have converging-motion signatures. We will elaborate on these two dense cores in the following subsection because they show significantly different chemical characteristics compared to the other starless cores in the L1495-B218 filaments.

The general trend of the fractional abundances of NH₃, CCS, and HC₇N in the models with C/O = 0.35 show evolution similar to that in Aikawa's models with small differences in details. The main chemical channels for forming and destroying NH₃, CCS, and HC₇N are given in Tables 1 and 2 for the core outskirts and mid-density core models. The fractional abundance of NH₃ increases almost monotonically as a function of time, except for the high-density core model (see Figure 19). In the high-density model, the NH₃ fractional abundance decreases after 3 × 10³ yr because of depletion onto dust grains. The peak NH₃ abundance is lower when the gas density is higher, because the depletion of NH₃ onto the dust grain surface becomes more efficient at the high-density core model, which is also observed in Aikawa's model. The fractional abundance of CCS shows a more-complicated evolution compared to NH₃ with three different trends as a function of time: a rapid increase of the fractional abundance at early times (<a few 1000 yr), a decrease of the fractional abundance from ∼10⁻⁹ to 10⁻¹¹ from ∼1000 yr to ∼0.1 Myr, and a slow increase of the abundance again after ∼0.1 Myr. The decrease of the CCS abundance is mainly due to its destruction by oxygen in gas-phase and is not due to depletion on dust grains (Table 1 and Table 2), unlike other carbon-bearing species (e.g., Tafalla & Santiago 2004; Pagani et al. 2005). The slow increase in CCS after ∼0.1 Myr is because its destruction rate by oxygen is significantly reduced as a result of the depletion of oxygen on dust grains. At this stage, the formation rate of CCS is still relatively low, so the CCS abundance does not rapidly increase as it does at earlier times. HC₇N also shows chemical trends similar to CCS except at low density, where the HC₇N formation via ${\rm{N}}\,+\,{{\rm{C}}}_{8}{\rm{H}}\to {\rm{C}}+{\mathrm{HC}}_{7}{\rm{N}}$ is much slower than CCS.

Table 1.  The Main Production and Destruction Reactions with Their Relative Contributions for NH₃, CCS, and HC₇N at Different Ages for Mid-density Core Outskirts (4 × 10³ cm⁻³) Models with Initial C/O = 0.35

Density	Age	Species	Formation	Destruction
cm−3	yr
		NH3	${\mathrm{NH}}_{4}^{+}+{{\rm{e}}}^{-}\to {\rm{H}}+{\mathrm{NH}}_{3}(69.2 \% )$	${\mathrm{NH}}_{3}+{{\rm{C}}}^{+}\to {\rm{H}}+{\mathrm{HCNH}}^{+}(62.9 \% )$
			H-ice + NH2-ice → NH3 (30.2%)	${\mathrm{NH}}_{3}+{{\rm{C}}}^{+}\to {\rm{C}}+{\mathrm{NH}}_{3}^{+}(31.5 \% )$
	102	CCS	HC2S+ + e− → H + CCS (89.2%)	CCS + C+ → S+ + C3 (68.7%)
			HC3S+ + e− → CH + CCS (8.4%)	CCS + C+ → C + C2S+ (19.6%)
		HC7N	N + C7H− → HC7N + e− (41%)	HC7N + C+ → H + C8N+ (49.7%)
			CN + C6H2 → H+ HC7N (35.6%)	HC7N + C+ → CN + C7H+ (49.7%)
		NH3	H-ice + NH2-ice → NH3 (69.9%)	${\mathrm{NH}}_{3}+{{\rm{S}}}^{+}\to {\rm{S}}+{\mathrm{NH}}_{3}^{+}\,(34 \% )$
			${\mathrm{NH}}_{4}^{+}+{{\rm{e}}}^{-}\to {\rm{H}}+{\mathrm{NH}}_{3}\,(29.4 \% )$	NH3 + C+ → H + HCNH+ (20.5%)
4 × 103	105	CCS	HC2S+ + e− → H + CCS (50.4%)	CCS + O → CO + CS (61.6%)
			C + HCS → CCS + H (28.4%)	CCS + C → C2 + CS (30.3%)
		HC7N	N + C8H → HC7N + C (85.4%)	HC7N + C → H + C8N (47.3%)
			C7H2N+ + e− → H+ HC7N (8%)	HC7N + H+ → H + HC7N+ (16.1%)
		NH3	${\mathrm{NH}}_{4}^{+}+{{\rm{e}}}^{-}\to {\rm{H}}+{\mathrm{NH}}_{3}\,(96.5 \% )$	${\mathrm{NH}}_{3}+{{\rm{H}}}^{+}\to {\rm{H}}+{\mathrm{NH}}_{3}^{+}\,(27.2 \% )$
			H-ice + NH2-ice → NH3 (3.1%)	NH3 + hν → H + NH2 (21.5%)
	106	CCS	HC2S+ + e− → H + CCS (58%)	CCS + H+ → H + C2S+ (61%)
			S + CCH → H + CCS (37.1%)	CCS + H3+ → H2 + HC2S+ (10%)
		HC7N	CN + C6H2 → H+ HC7N (47.6%)	HC7N + H+ → H + HC7N+ (41.5%)
			C7H2N+ + e− → H+ HC7N (40.4%)	HC7N + H3+ → H2 + C7H2N+ (49.7%)

Download table as:  ASCIITypeset image

Table 2.  The Main Production and Destruction Reactions with Their Relative Contributions for NH₃, CCS and HC₇N at Different Ages for Mid-density Core (5 × 10⁴ cm⁻³) Models with Initial C/O = 0.35

Density	Age	Species	Formation	Destruction
cm−3	yr
		NH3	H-ice + NH2-ice → NH3 (92%)	NH3 + C+ → H + HCNH+ (61.9%)
			${\mathrm{NH}}_{4}^{+}+{{\rm{e}}}^{-}\to {\rm{H}}+{\mathrm{NH}}_{3}(7.9 \% )$	${\mathrm{NH}}_{3}+{{\rm{C}}}^{+}\to {\rm{C}}+{\mathrm{NH}}_{3}^{+}\,(31 \% )$
	102	CCS	HC2S+ + e− → H + CCS (78.2%)	CCS + C+ → S+ + C3 (65.7%)
			HC3S+ + e− → CH + CCS (17.1%)	CCS + C+ → C + C2S+ (18.8%)
		HC7N	N + C7H− → HC7N + e− (56.8%)	HC7N + C+ → H + C8N+ (48.4%)
			CN + C6H2 → H+ HC7N (20.6%)	HC7N + C+ → CN + C7H+ (48.4%)
		NH3	${\mathrm{NH}}_{4}^{+}+{{\rm{e}}}^{-}\to {\rm{H}}+{\mathrm{NH}}_{3}\,(95.7 \% )$	${\mathrm{NH}}_{3}+{{\rm{H}}}_{3}^{+}\to {{\rm{H}}}_{2}+{\mathrm{NH}}_{4}^{+}\,(25.2 \% )$
			H-ice + NH2-ice → NH3 (3.3%)	${\mathrm{NH}}_{3}+{{\rm{H}}}_{3}{{\rm{O}}}^{+}\to {{\rm{H}}}_{2}{\rm{O}}+{\mathrm{NH}}_{4}^{+}\,(24.3 \% )$
5 × 104	105	CCS	HC2S+ + e− → H + CCS (62.2%)	CCS + O → CO + CS (90.2%)
			HC3S+ + e− → CH + CCS (15.3%)	CCS + N → CN + CS (2.3%)
		HC7N	N + C8H → HC7N + C (45%)	${\mathrm{HC}}_{7}{\rm{N}}\,+\,{{\rm{H}}}_{3}^{+}\to {{\rm{H}}}_{2}+{{\rm{C}}}_{7}{{\rm{H}}}_{2}{{\rm{N}}}^{+}\,(49.4 \% )$
			C7H2N+ + e− → H+ HC7N (44.5%)	HC7N + HCO+ → CO + C7H2N+ (24.8%)
		NH3	${\mathrm{NH}}_{4}^{+}+{{\rm{e}}}^{-}\to {\rm{H}}+{\mathrm{NH}}_{3}\,(99.5 \% )$	${\mathrm{NH}}_{3}+{{\rm{H}}}_{3}^{+}\to {{\rm{H}}}_{2}+{\mathrm{NH}}_{4}^{+}\,(76.2 \% )$
				${\mathrm{NH}}_{3}+{{\rm{H}}}^{+}\to {\rm{H}}+{\mathrm{NH}}_{3}^{+}\,(10.6 \% )$
	106	CCS	HC2S+ + e− → H + CCS (73%)	$\mathrm{CCS}+{{\rm{H}}}_{3}^{+}\to {{\rm{H}}}_{2}+{\mathrm{HC}}_{2}{{\rm{S}}}^{+}\,(79.9 \% )$
			S + CCH → H + CCS (14.8%)	CCS + H+ → H + C2S+ (11.5%)
		HC7N	C7H2N+ + e− → H+ HC7N (53.1%)	${\mathrm{HC}}_{7}{\rm{N}}+{{\rm{H}}}_{3}^{+}\to {{\rm{H}}}_{2}+{{\rm{C}}}_{7}{{\rm{H}}}_{2}{{\rm{N}}}^{+}\,(81.7 \% )$
			CN + C6H2 → H+ HC7N (45.5%)	HC7N + H+ → H + HC7N+(11.9%)

Download table as:  ASCIITypeset image

We compared our observations to the chemical models with the initial gas-phase C/O = 0.35 in Figure 19. We found that most regions, including B10, B213W, B211, B216, and B218, show good agreement with the models. For example, young and relatively less-condensed starless cores in B211 and B216 have NH₃ fractional abundance ∼1 × 10⁻⁹ with respect to H₂ at the core centers and have CCS fractional abundances of (0.4–1) × 10⁻⁹ with respect to H₂ at CCS peak-intensity positions. Considering the uncertainties in the fractional abundances, we found that B211 and B216 have better agreement with the core outskirts model. From this model, fractional abundances of NH₃ and CCS are 2.2 × 10⁻⁹ and 3 × 10⁻¹⁰, respectively, with respect to H₂ at a chemical age of ∼0.1 Myr, which is reasonable for young starless cores. The HC₇N abundance at this chemical age is 1 × 10⁻¹² with respect to H₂ which explains our nondetection as well. The NH₃ and CCS abundances in B211 are well-reproduced by our models. For B216, NH₃ is in good agreements with the model at 0.1 Myr, but CCS is a factor of five higher compared to the models at the same age.

Starless cores in B10 and B213W, which are more condensed than the ones in B211 and B216, have fractional abundances of NH₃ of 3 × 10⁻⁹ and 5 × 10⁻⁹, respectively, at their centers and 1 × 10⁻⁹ and 2 × 10⁻⁹, respectively, at their outskirts. Considering the uncertainties in the fractional abundances, we think that the these two regions may correspond to a chemical age around 0.3 Myr or older, where the CCS fractional abundance steeply decreases in core outskirts models. At this age, the NH₃ abundance is (1–20) × 10⁻⁹ depending on density, whereas the CCS abundance is (0.1–2) × 10⁻¹⁰ with respect to H₂, which mostly falls below the lowest detected fractional abundance of CCS of 4 × 10⁻¹⁰. We did not detect HC₇N in B10 and B213W, and the fractional abundance of HC₇N in the model is ∼10⁻¹² or lower, which is below the lowest observed fractional abundance of 5 × 10⁻¹¹ in our entire HC₇N survey.

Starless cores in B218 are more-condensed cores similar to those in B10 and B213W. These starless cores show weak CCS rings, whereas we did not detect any HC₇N emission at our noise level. The NH₃ fractional abundance ranges are (3–4) × 10⁻⁹ at the core centers. The NH₃ and CCS fractional abundances are (0.5–1) × 10⁻⁹ and (3–6) × 10⁻¹⁰, respectively, at the outskirts. Comparing these abundances to the mid-density and outskirts models with the initial C/O = 0.35, we found that these starless cores may be at a chemical age between 0.1 Myr and 0.2 Myr. At this age, the mid-density and outskirts models with the initial C/O = 0.35 show the NH₃ and CCS fractional abundance ranges of (0.8–10) × 10⁻⁹ and (0.5–1) × 10⁻¹⁰, respectively, and are in good agreement with the observed abundances.

We found that most of the regions in the L1495-B218 filaments agree with the chemical models with C/O = 0.35. However, we found that two starless cores, L1495A-N and L1521D, have significantly different characteristics compared to other starless cores in Taurus. In the following section, we elaborate chemical evolution of L1495A-N and L1521D and the implications related to their dynamics.

The physical conditions of L1495A-N and L1521D, including density structures, extinction at the core surface, kinetic temperature, and turbulence, are similar to those of the starless cores in B213W and B10, but the fractional abundances of CCS and HC₇N in L1495A-N and L1521D are much higher than in the starless cores in B213W and B10.

The fractional abundances of NH₃ and CCS with respect to H₂ at the core centers (NH₃ peak-intensity position) are 4 × 10⁻⁹ and 6 × 10⁻¹⁰, respectively, for L1495A-N, and 5 × 10⁻⁹ and 3 × 10⁻¹⁰, respectively, for L1521D. We did not detect any HC₇N at the core centers. At the CCS peak-intensity position at the outskirts of the cores, the fractional abundances of NH₃ and CCS with respect to H₂ are 3 × 10⁻¹⁰ and 3 × 10⁻⁹, respectively, for L1495A-N, and 5 × 10⁻¹⁰ and 2 × 10⁻⁹, respectively, for L1521D. We did not detect HC₇N emission at these positions. On the other hand, at the HC₇N peak-intensity position, the fractional abundances of NH₃, CCS, and HC₇N with respect to H₂ are 6 × 10⁻¹⁰, 1 × 10⁻⁹, and 2 × 10⁻¹⁰, respectively, for L1495A-N and 2 × 10⁻⁹, 1 × 10⁻⁹, and 3 × 10⁻¹⁰, respectively, for L1521D.

For L1495A-N and L1521D, we found that the models with ${n}_{{{\rm{H}}}_{2}}=4\times {10}^{3}\,{\mathrm{cm}}^{-3}$ and 5 × 10⁴ cm⁻³ correspond to the average densities in the core outskirts and the core center, respectively, evaluated by column density/size. The actual volume densities of the core centers may be several times higher than 5 × 10⁴ cm⁻³ but still less than 5 × 10⁵ cm⁻³ if their density profiles are similar to the Bonnor–Ebert sphere (Ebert 1955; Bonnor 1956). Comparing the observed NH₃ and CCS abundances to the models with the initial C/O = 0.35, we found that the observed CCS fractional abundances of L1495A-N and L1521D are significantly higher (at least 20 times more abundant at core center and at the CCS peak-intensity position) than the models with the initial C/O = 0.35 if we determine the age of the cores using the observed NH₃ abundance to be ∼0.1 Myr. Also, the peak CCS abundances in the models are still more than five times lower compared to the observed CCS fractional abundance at the CCS peak-intensity positions of L1495A-N and L1521D. We found that the observed HC₇N fractional abundances at the HC₇N peak-intensity position are at least 100 times higher than the fractional abundance predicted by the model with the initial C/O = 0.35. These discrepancies between the observed fractional abundances and the models with the initial C/O = 0.35 strongly suggest that the two cores are going through a different chemical evolutionary path compared to the other starless cores in the L1495-B218 filaments.

Tables 1 and 2 clearly show the importance of the ${\rm{O}}+\mathrm{CCS}\to \mathrm{CO}+\mathrm{CS}$ reaction that controls the chemistry of CCS in the majority of starless cores in the L1495-B218 filaments described in an earlier subsection. By running a grid of chemical models, we found that the reaction ${\rm{O}}+\mathrm{CCS}\to \mathrm{CO}+\mathrm{CS}$ is highly sensitive to the initial gas-phase C/O ratio because the abundance of the gas-phase atomic oxygen controls the reaction flux of this channel (i.e., the higher the C/O ratio, the lower the rate of destruction of CCS). To explain the observed high abundances of CCS together with HC₇N within a reasonable uncertainty, we have found the best agreement to be with the chemical models having C/O = 1.

The right column of Figure 19 shows the evolution of the fractional abundances as a function of time. The general trend of the NH₃ fractional abundance over time in the C/O = 1 models is similar to that of the C/O = 0.35 models. On the other hand, the evolution of the CCS and HC₇N fractional abundances in the C/O = 1 models is very different from the C/O = 0.35 models. The CCS and HC₇N fractional abundances in the C/O = 1 models are significant even at an age of >1 Myr, having peak abundances >10⁻⁹, whereas the CCS and HC₇N fractional abundances in the C/O = 0.35 models tend to decrease and maintain an abundance of <10⁻¹⁰ over a long period of time after a few times 0.01 Myr. This is mainly because of less oxygen being available to destroy CCS and more carbon to form CCS and HC₇N when the initial gas-phase C/O ratio is high.

The observed fractional abundances of NH₃ and CCS at the center of L1495A-N and L1521D correspond to chemical ages between 0.2 and 0.4 Myr using the mid-density model, which gives the modeled abundances of 4 × 10⁻⁹ for NH₃ and 2 × 10⁻¹⁰ for CCS. Table 3 shows the major reaction channels for production and destruction of NH₃, CCS, and HC₇N at the age between 0.2 and 0.4 Myr for core outskirts and mid-density core model. If we use the high-density model, we have a chemical age matching the observed NH₃ abundance around 0.01 Myr. The high-density model results in younger ages because of the low abundance of CCS after 0.03 Myr. Considering that we have average densities of (0.5–2) × 10⁵ cm⁻³ toward the core centers of L1495A-N and L1521D, we do not think that the two cores have central densities near to 5 × 10⁵ cm⁻³. Also, we do not observe any central depletion of NH₃ in L1495A-N and very little depletion of NH₃ in L1521D (only 20% decrease of the NH₃ fractional abundance), whereas the high-density model shows an order of magnitude decrease of the NH₃ fractional abundance. The density is also unlikely to reach the central density of 5 × 10⁵ cm⁻³ within <0.01 Myr because a dense core typically takes much longer than 0.01 Myr to reach central density 5 × 10⁵ cm⁻³ (e.g., Gong & Ostriker 2009, 2011). Thus, chemical ages of L1495A-N and L1521D using the mid-density core model, 0.2–0.4 Myr, fit better to the observed quantities and are similar to the ages of other dense cores (e.g., Ward-Thompson et al. 2007; Brünken et al. 2014).

Table 3.  The Main Production and Destruction Reactions with Their Relative Contributions for NH₃, CCS, and HC₇N at the Age between 0.2 and 0.4 Myr for Core Outskirts and Mid-density Core Models with Initial C/O = 1

Density	Species	Formation	Destruction
cm−3
	NH3	${\rm{H}} \mbox{-} \mathrm{ice}+{\mathrm{NH}}_{2} \mbox{-} \mathrm{ice}\to {\mathrm{NH}}_{3}(\,90.4 \% )$	${\mathrm{NH}}_{3}+{{\rm{C}}}^{+}\to {\rm{H}}+{\mathrm{HCNH}}^{+}(40.2 \% )$
		${\mathrm{NH}}_{4}^{+}+{{\rm{e}}}^{-}\to {\rm{H}}+{\mathrm{NH}}_{3}\,(9.5 \% )$	${\mathrm{NH}}_{3}+{{\rm{C}}}^{+}\to {\rm{C}}+{\mathrm{NH}}_{3}^{+}\,(20.1 \% )$
4 × 103	CCS	${\mathrm{HC}}_{2}{{\rm{S}}}^{+}+{{\rm{e}}}^{-}\to {\rm{H}}+\mathrm{CCS}(68.9 \% )$	$\mathrm{CCS}+{{\rm{C}}}^{+}\to {{\rm{S}}}^{+}+{{\rm{C}}}_{3}\,(41.4 \% )$
		${\mathrm{HC}}_{3}{{\rm{S}}}^{+}+{{\rm{e}}}^{-}\to \mathrm{CH}+\mathrm{CCS}(27.1 \% )$	$\mathrm{CCS}+{\rm{C}}\to {{\rm{C}}}_{2}+\mathrm{CS}(25.8 \% )$
	HC7N	${\rm{N}}\,+\,{{\rm{C}}}_{8}{\rm{H}}\to {\rm{C}}+{\mathrm{HC}}_{7}{\rm{N}}(68.6 \% )$	${\mathrm{HC}}_{7}{\rm{N}}\,+\,{{\rm{C}}}^{+}\to {\rm{H}}\,+\,{{\rm{C}}}_{8}{{\rm{N}}}^{+}\,(32.9 \% )$
		$\mathrm{CN}+{{\rm{C}}}_{6}{{\rm{H}}}_{2}\to {\rm{H}}+{\mathrm{HC}}_{7}{\rm{N}}(22 \% )$	${\mathrm{HC}}_{7}{\rm{N}}\,+\,{{\rm{C}}}^{+}\to \mathrm{CN}+{{\rm{C}}}_{7}{{\rm{H}}}^{+}\,(32.9 \% )$
	NH3	${\mathrm{NH}}_{4}^{+}+{{\rm{e}}}^{-}\to {\rm{H}}+{\mathrm{NH}}_{3}(98.6 \% )$	${\mathrm{NH}}_{3}+{{\rm{H}}}_{3}^{+}\to {{\rm{H}}}_{2}+{\mathrm{NH}}_{4}^{+}\,(63.7 \% )$
			${\mathrm{NH}}_{3}+{\mathrm{HCO}}^{+}\to \mathrm{CO}+{\mathrm{NH}}_{4}^{+}\,(7 \% )$
5 × 104	CCS	${\mathrm{HC}}_{2}{{\rm{S}}}^{+}+{{\rm{e}}}^{-}\to {\rm{H}}+\mathrm{CCS}(63 \% )$	$\mathrm{CCS}+{{\rm{H}}}_{3}^{+}\to {{\rm{H}}}_{2}+{\mathrm{HC}}_{2}{{\rm{S}}}^{+}\,(71.2 \% )$
		${\rm{S}}+\mathrm{CCH}\to {\rm{H}}+\mathrm{CCS}(20.2 \% )$	$\mathrm{CCS}+{{\rm{H}}}^{+}\to {\rm{H}}\,+\,{{\rm{C}}}_{2}{{\rm{S}}}^{+}\,(7.6 \% )$
	HC7N	${{\rm{C}}}_{7}{{\rm{H}}}_{2}{{\rm{N}}}^{+}+{{\rm{e}}}^{-}\to {\rm{H}}+{\mathrm{HC}}_{7}{\rm{N}}(50.9 \% )$	${\mathrm{HC}}_{7}{\rm{N}}\,+\,{{\rm{H}}}_{3}^{+}\to {{\rm{H}}}_{2}+{{\rm{C}}}_{7}{{\rm{H}}}_{2}{{\rm{N}}}^{+}\,(76.7 \% )$
		$\mathrm{CN}+{{\rm{C}}}_{6}{{\rm{H}}}_{2}\to {\rm{H}}+{\mathrm{HC}}_{7}{\rm{N}}(44.7 \% )$	${\mathrm{HC}}_{7}{\rm{N}}\,+\,{{\rm{H}}}^{+}\to {\rm{H}}+{\mathrm{HC}}_{7}{{\rm{N}}}^{+}\,(8.2 \% )$

Download table as:  ASCIITypeset image

The observed fractional abundances of NH₃ and CCS at the outskirts/CCS peak-intensity position of L1495A-N and L1521D also show good agreement with the outskirts model at an age between 0.2 and 0.4 Myr, which gives NH₃ and CCS fractional abundances 6 × 10⁻¹⁰ and 2 × 10⁻⁹, respectively. This age range is relatively consistent with the chemical age at the core centers of L1495A-N and L1521D and also similar to the chemical ages of the other starless cores in the L1495-B218 filaments 0.3 Myr or older.

The comparisons between observations and the chemical models suggest that most regions of the L1495-B218 filaments have likely formed in an environment with the initial C/O ratio in the gas around 0.35, which is close to the solar value (Le Gal et al. 2014). However, for the two starless cores L1495A-N and L1521D, we were not able to explain high abundance of CCS and HC₇N unless we adopt chemical models with higher initial gas-phase C/O ratio of ≥1. Because most of the cores in the L1495-B218 filaments agree with the chemical model with a low initial C/O, the initial C/O ratio across the Taurus molecular cloud was likely homogeneous with a value close to solar. Therefore, it appears likely that L1495A-N and L1521D underwent different formation mechanism or different evolutionary path compared to the rest of the cores in the L1495-B218 filaments.

Although the physics of having a high initial gas-phase C/O ratio is out of the scope of this article, we suggest two possible mechanisms that could produce results similar to the chemical model with a high initial C/O ratio in the gas (e.g., Hincelin et al. 2011, and references therein). The first possible mechanism is that a core forms by a collision of filaments. Generally, the gas-phase C/O ratio varies significantly depending on the gas density, extinction, and age. At a molecular cloud clump density >1000 cm⁻³, a gas-phase C/O ratio can increase significantly after ∼1 Myr because of oxygen depletion. Thus, if molecular filaments are older than ∼1 Myr and the collision of the filaments triggers core formation, the resulting cores would form in a high C/O ratio environment and have bright CCS and HC₇N emission. The second possible mechanism is that a core forms in a low C/O ratio environment (similar to the solar value), spends a long time in a quasi-static stage, and collapses while accreting gas from surrounding filament materials. The aged filament gas would have a high C/O ratio in the gas, so the accreted gas on the surface of the core may develop abundant CCS and HC₇N and can make bright CCS and HC₇N rings.