1: \documentclass[12pt,preprint]{aastex}
2:
3: \newcommand{\vdag}{(v)^\dagger}
4: \newcommand{\myemail}{jjsohn@astro.snu.ac.kr}
5: \def\today{\rightline{\ifcase\month\or
6: January\or February\or March\or April\or May\or June\or
7: July\or August\or September\or October\or November\or December\fi
8: \space\number\day, \number\year}}
9:
10: \begin{document}
11: \title{Probing Inward Motions in Starless Cores Using
12: The HCN $J = 1 - 0$ Hyperfine Transitions : A Pointing Survey
13: Toward Central Regions}
14: \author{Jungjoo Sohn\altaffilmark{1,2}, Chang Won Lee\altaffilmark{2},}
15: \author{Yong-Sun Park\altaffilmark{1}, Hyung Mok Lee\altaffilmark{1},}
16: \author{Philip C. Myers\altaffilmark{3}, and Youngung Lee\altaffilmark{4}}
17:
18: \altaffiltext{1}{Astronomy Department, School of Physics and
19: Astronomy, Seoul National University, Shillim-dong, Kwanak-gu,
20: Seoul 151-749, Korea}
21: \altaffiltext{2}{International Center for Astrophysics, Korea
22: Astronomy \& Space Science Institute, 61-1 Hwaam-dong, Yusung-gu,
23: Daejeon 305-348, Korea}
24: \altaffiltext{3}{Harvard-Smithsonian
25: Center for Astrophysics, 60 Garden Street, MS 42, Cambridge, MA
26: 02138, USA}
27: \altaffiltext{4}{Korea Astronomy \& Space Science
28: Institute, 61-1 Hwaam-dong, Yusung-gu, Daejeon 305-348, Korea}
29: \clearpage
30: \begin{abstract}
31: We have carried out a survey toward the central regions of 85
32: starless cores in HCN $J=1-0$ to study inward motions in the
33: cores. Sixty-four cores were detected with HCN lines. The infall
34: asymmetry in the HCN spectra is found to be more
35: prevalent, and more prominent than in any other previously used infall
36: tracers such as CS $J=2-1$, $\rm DCO^+$ $J=2-1$, and $\rm N_2H^+$
37: $J=1-0$. We found close relation between the intensities of the
38: HCN and $\rm N_2H^+$ lines. This implies that the HCN is not much
39: depleted in the central regions of the cores. In some cores, the
40: HCN spectra show different sign of asymmetry from other molecular
41: lines. A few cores show various signs of asymmetry in individual
42: HCN hyperfine lines. The distribution of the velocity shift
43: $\delta V$ of the HCN profiles with respect to the systemic
44: velocity of the optically thin tracer is found to be more shifted
45: toward bluer side than those of other infall tracers, indicating
46: that the HCN traces inward motions more frequently. The $\delta V$
47: distribution of each HCN hyperfine line for all sources is
48: similar. Moreover the $\delta V$ values obtained from different
49: HCN hyperfine lines for each source are nearly similar. These may
50: mean that most of starless cores are in similar kinematic states
51: across the layers of the cores. We identify 17 infall candidates
52: using all available indicators such as the velocity shift $\delta
53: V$ and the blue to red peak intensity ratio of double peaked
54: profiles for HCN $J=1-0$, CS $J=2-1$, $J=3-2$, DCO$^+$ $J=2-1$,
55: and $\rm N_2H^+$ $J=1-0$. Four of them, L63, L492, L694-2, and
56: L1197 are found to show higher blue to red ratio in the HCN
57: hyperfine line along the lower opacity, suggesting that infall
58: speed becomes higher toward the center.
59: \end{abstract}
60: \keywords{ISM: globules --- ISM: kinematics and dynamics --- ISM:
61: molecules --- stars: formation}
62:
63: %\today
64: \section{Introduction}
65:
66: Starless cores are dense molecular condensations without Yong
67: Stellar Objects (YSOs) (Beichman et al. 1986; Lee \& Myers 1999,
68: hereafter LM99; Di Francesco et al. 2006). Some of the starless
69: cores are called ``prestellar cores'' which show more evolved
70: features of the starless cores such as relatively higher $\rm
71: H_2$ column density, more compact density distribution,
72: deuterium enrichment, more pronounced CO depletion, and inward
73: motions (Crapsi et al. 2005). Because of this, they are believed
74: to be in a stage close to form a star. On the other hand there are
75: other types of starless cores which do not show any good hints of
76: star formation activity. It seems that not every starless core is
77: destined to deliver a baby star. In most cases, however, the
78: starless cores are believed to be one of the best targets to study
79: the initial conditions of star formation and its related physical
80: and chemical processes (Benson \& Myers 1989; Ward-Thompson et al.
81: 1994; Shirley et al. 2000; Andr\'{e}, Ward-Thompson, \& Motte
82: 1996).
83:
84: Inward motions are one of the key processes to study physical and
85: kinematical status of the starless core, and to test models of
86: star formation. The first evidence of the existence of such
87: motions in starless cores was discovered from a study of the
88: spatial distribution of the ``spectral infall asymmetry'' toward a
89: starless core L1544, the spectral line combination of a self
90: absorbed profile with double peaks of a blue component being
91: brighter than red one in an optically thick tracer and a single
92: peak profile in an optically thin tracer (Tafalla et al. 1998).
93: The infall asymmetry found in L1544 was spatially extended ($\sim
94: 0.1$ pc) with its deduced infall speed of up to 0.1 km s$^{-1}$.
95: Lee, Myers \& Tafalla (1999, LMT99 hereafter) found from a
96: pointing survey of starless cores in CS $J = 2 - 1$ and $\rm
97: N_2H^+$ $J = 1 - 0$ lines that the inward motions are the most
98: dominant kinematics in starless cores with a density higher than
99: about $10^{4}$ $\rm cm^{-3}$. From their follow-up mapping
100: observations of starless cores in the same lines, Lee, Myers \&
101: Tafalla (2001, LMT01 hereafter) have showed that the infall
102: asymmetry found in L1544, spatially extended up $0.2$ pc with the
103: line of sight speed of order 0.1 km s$^{-1}$, is a general feature
104: in the collapsing starless cores. The extended nature of the
105: infall asymmetry has an implication that the inward motions might
106: exist before the formation of a central star. This is inconsistent
107: with the inside-out model by Shu (1977) (Tafalla et al 1998). Lee,
108: Myers, \& Plume (2004, LMP04 hereafter) reported that several
109: cores may have higher infall speed in the inner denser regions
110: from the comparison between the CS $J = 3 - 2$ and CS $J = 2 - 1$
111: line surveys. Recently, Williams, Lee, \& Myers (2006) showed
112: spectral signatures of rotation and collapse in two starless
113: cores, L1544 and L694-2, by comparison of interferometric map
114: observations in $\rm N_2H^+$ $J = 1 - 0$. They found that the less
115: evolved core, L694-2, has smaller rate of velocity increase toward
116: the center than the more evolved core, L1544, indicating that
117: development of the spatial structure of the infall speed may be
118: related to the evolutionary stage of the starless cores. These
119: observational results on the inward motions in starless cores will
120: be useful to improve the collapsing models explaining the features
121: of inward motions in the early stage of star formation (e.g.,
122: Whitworth \& Ward-Thompson 2001, Ciolek \& Basu 2000, and Myers
123: 2005).
124:
125: One obstacle in this kind of study is a possible depletion of the
126: infall tracer in the dense central regions of the cores (Caselli
127: et al. 1999; Tafalla et al. 2002; Alves, Lada \& Lada 2001;
128: Charnley 1997). Leger (1983) and Bergin \& Langer (1997) have
129: shown that in a cold core denser than a few $10^{4}$ $\rm
130: cm^{-3}$, polar molecules such as CS are easily depleted by being
131: adsorbed on the dust grains. This property may often limit the
132: study of inward motions occurring in the nuclear regions of the
133: cores. For example, the typical size of CS depleted region in
134: starless cores observed by Tafalla et al. (2002) is about 0.1 pc
135: which is almost comparable to the size of the region showing
136: extended infall asymmetry. This implies that most of the blue
137: skewed CS profiles seen in the infalling cores may not be formed
138: from the central region, but mainly from envelopes of the cores.
139: Therefore, CS line may not be the best to trace infall motions
140: occurring at the nucleus of the cores.
141:
142: This study introduces a survey of infall motions toward starless
143: cores with HCN $J = 1 - 0$ line. Though the intensity of HCN is
144: weaker than that of other transitions, this line has an advantage
145: of tracing the central motion more sensitively than CS (e.g., Park
146: et al. 1999). Our follow-up mapping observations of several
147: starless cores in HCN $J = 1 - 0$ (Sohn et al. in preparation)
148: indicate a similar distribution to that of $\rm N_2H^+$ which
149: usually traces the density distribution of inner part of the cores
150: (Tafalla et al. 2006, Di Francesco et al. 2006). Aikawa et al.
151: (2005) have shown that HCN $J = 1 - 0$ does not suffer from
152: significant depletion effect in the central density of the core
153: less than 10$^{5}$ cm$^{-3}$. The calculation of the evolution of
154: the HCN abundance by Lee et al. (2004) indicates that HCN is
155: slightly depleted in a very small (of order $10^{-3}$ pc) region
156: near the center of the core, at the moment right near the stage of
157: the star formation, but not as much as the CS molecule. This
158: ``small'' depletion is seen in two starless cores, L1498 and
159: L1517B (Tafalla et al. 2006).
160:
161: Another advantage of HCN is the presence of three hyperfine lines
162: with different relative opacities of F(0-1) : F(1-1) : F(2-1) = 1
163: : 3 : 5 under the LTE condition. Because the hyperfine lines have
164: almost equal excitation temperature, these are expected to trace
165: somewhat different regions mainly depending on their given optical
166: depths. HCN line can be useful to study spatial variation of
167: inward motions along the line of sight of the cores.
168:
169: In this paper we report our systematic survey of 85 starless cores
170: with HCN $J = 1 - 0$ hyperfine lines to investigate the occurrence
171: and pattern of inward motions. In \S 2, we explain observations of
172: the starless cores. In \S 3 we describe the detection statistics,
173: spectral features and analysis of the observed HCN lines.
174: Implications of the observational results are discussed in \S 4.
175: Finally, we summarize our results in \S 5.
176:
177: \section{Observations}
178:
179: We have conducted single-pointing observations for 85 starless
180: cores in HCN $J = 1 - 0$ with the 13.7 m telescope of the Taeduk
181: Radio Astronomical Observatory (TRAO) in Korea. Observations were
182: carried out during January, February, November and December in
183: 2002 and January to May in 2003. We used the SIS receiver and
184: autocorrelation spectrometer with 10 KHz (corresponding to a
185: velocity resolution of 0.033 km s$^{-1}$ at 88.6 GHz) or 20 KHz
186: resolution. We adopted a frequency switching mode with frequency
187: throw of 5 or 8 MHz. Calibration was frequently done by measuring
188: the ambient temperature. System temperature during the observations
189: was typically between 400 - 600 K for single sideband. Pointing
190: accuracy during the observation was within $\pm$ 10 \arcsec (rms).
191: HPBW of TRAO telescope is 64\arcsec, and beam efficiency is 50$\%$
192: at 86 GHz frequency (Roh \& Jung 1999). We also observed 10
193: starless cores in $\rm H^{13}CN$ $J = 1 - 0$ line with similar
194: instrument configuration and observing mode between April and May,
195: 2003. The system temperature during the observation was between
196: 280 and 330 K.
197:
198: Targets were selected primarily based on the strengths of
199: N$_{2}$H$^{+}$ $J = 1 - 0$, CS $J = 2 - 1$ (LMT99, LMT01), and
200: HCO$^{+}$ $J = 1 - 0$ (Lee et al. 2007 in prep.) lines. Table 1
201: lists the names and positions of the sources and the observed line
202: parameters. All sources are known to be nearby (within a few
203: hundreds pc), dense (around 10$^{4-5}$ cm$^{-3}$), compact (r
204: $\approx$ 0.05 $\sim$ 0.35 pc) (LMT99), and have narrow line
205: widths ($\Delta$V$_{\rm FWHM}$ of N$_{2}$H$^{+}$ $\sim 0.2$ - $0.4
206: $ km s$^{-1}$, LMT99).
207:
208: For a quantitative analysis of inward motions using HCN lines, it
209: is necessary to determine the accurate frequency of HCN. A usual
210: way to determine the frequency of the HCN is to observe a source
211: showing a Gaussian shape in both HCN and other reference lines
212: such as C$^{18}$O $J = 1 - 0$ or N$_{2}$H$^{+}$ $J = 1 - 0$ whose
213: frequencies are already accurately known. However, most of HCN
214: profiles turned out to be highly affected by high optical depths,
215: showing either double-peaked or skewed features, making it
216: difficult to decide its frequency from the observations. So, we
217: use the value of the Cologne Database for Molecular Spectroscopy
218: giving the best estimation of HCN frequency with the lowest
219: measurement uncertainty of $\rm \sim 1~KHz$ (M\"uller et al. 2005)
220: which is smaller than the spectrometer resolution by a factor 10
221: $\sim$ 20. The value for HCN $J = 1 - 0$ hyperfine lines are as
222: follows; 88630.4157 ($\pm 0.001$) MHz for ($F$=1$-$1), 88631.8473
223: ($\pm 0.001$) MHz for HCN ($F$=2$-$1), and 88633.9360 ($\pm
224: 0.001$) MHz for ($F$=0$-$1). The reduction of the spectral line
225: data was performed using the SPA developed by FCRAO and the CLASS
226: reduction software (Buisson et al. 1994).
227:
228: \section{Result}
229: In total sixty four sources out of 85 targets are detected in HCN
230: $J = 1 - 0$ and eight of ten observed sources are detected in $\rm
231: H^{13}CN$ $J = 1 - 0$.
232: The observed line parameters of the sources are listed in Table 1.
233: The antenna temperature (T$_{A}^*$) of the HCN line ranges
234: 0.17 $\sim$ 1.50 K for ($F$=2$-$1). The $\rm H^{13}CN$ is usually
235: very weak : $\rm T_{A}^{*} = $ $0.06\sim 0.23$ K. Note that most
236: of the detected sources are observed in the HCN line with high S/N
237: ($>$ 10). The profiles of the observed lines for all detected
238: sources are displayed in Figure 1.
239:
240: \subsection{The HCN Spectra}
241:
242: One of the most noticeable features seen in Figure 1 is that the
243: HCN profiles hardly show any clear single Gaussian feature.
244: Instead, most of the HCN $J = 1 - 0$ spectra show asymmetric
245: feature such as double peaks with a dip or a single peak with
246: shoulder. The peak velocity of the optically thin tracer $\rm
247: N_2H^+$ $J = 1 - 0$ line is located close to a dip between the asymmetric profiles.
248: The estimated velocity distance between the peak velocity of $\rm
249: N_2H^+$ and the dip of HCN profile for 17 sources showing clear dip
250: in the HCN profiles is $0.1\pm 0.08$ $\rm km~s^{-1}$.
251: Therefore such a dip is
252: thought to be caused by self-absorption of emissions from the warm
253: nucleus by the cool envelope of the cores. It is interesting to
254: note that the self-absorption dip tends to be deeper at the
255: hyperfine component with large opacity (HCN, $F$=2$-$1), and at
256: the same time the infall asymmetry tends to be more clearly seen
257: at this component. Based on the details line asymmetry, the HCN
258: spectra can be classified into four groups (See Fig. 2) ;
259:
260: Group 1: All three hyperfines have two peaks, with a blue peak
261: brighter than a red peak. It occupies around 9 \% of the detected
262: sources ( Fig. 2a: L1495AN, L1521F, L1544, L1689B, L492, and
263: L694-2).
264:
265: Group 2: One or two of the hyperfines has two peaks, with a blue
266: peak being brighter than a red peak or a red shoulder component.
267: About 34 \% of the detected sources fall into this group (Fig. 2b:
268: L1333, L1355, L1498, L1495A-S, B217-2, B18-1, TMC2, TMC1, L1507A,
269: CB23, L1552, L1582A, L1622A2, L1622A1, CB45, L1696A, L1696B, L63,
270: L673-7, L1155C1, L1251A2, and L1197).
271:
272: Group 3: At least one of hyperfines has two peaks, with a red peak
273: being brighter than a blue peak or a blue shoulder. It occupies around
274: 22 \% (Fig. 2c: B5-1, L1399-2, L1521E, L1400A, B18-5, CB22,
275: L1527B1, L1517B, L1512, L1621-1, L134A, L1704-1, L204F, and
276: L1049-1).
277:
278: Group 4: Red and blue asymmetric profiles are mixed in three
279: hyperfines. This group has around 33 \% in total share (Fig. 2d:
280: L1521B2, L1521B, B18-3, L1517C1, L183, L1709B2, L204C-2, L234E-S,
281: L462-1, L462-2, L429-1, L490, CB130-1, L648-1, L1041-2, L922-1,
282: L1049-2, L1155C2, L944-2, L1251 and CB246-2).
283:
284: L1633 does not fall into any of the above mentioned groups because
285: its profiles have broad peaks without clear double peaks at two
286: hyperfines. The sources in groups 1 and 2 are referred to having
287: blue asymmetry as the optically thin $\rm N_2H^+$ line show a
288: single peak located near the dip or shoulder of the HCN profiles.
289: These groups occupy around 43 \% of the detected sources. On the
290: other hand the sources in group 3 have red-asymmetry.
291: The cause of such a red asymmetry is not clear: it could be due
292: to expansion as opposed to the collapse.
293:
294: We note that there are also the sources that have some interesting
295: features in HCN spectra. There are two distinct groups. There is a
296: group of sources having extraordinarily broad line width, almost
297: twice of (FWHM of $\sim$ 0.8 km s$^{-1}$) those ($\sim$ 0.4 km
298: s$^{-1}$) of typical starless cores (See Fig. 2e). These are B5-1,
299: L1582A, L1633, and L204F ($\sim 6$ \% of the detected 64 sources).
300:
301: Another interesting group is the sources showing anomalous
302: intensity ratio among the hyperfine lines. If the sources are in
303: LTE and the optically thin condition, the relative intensities of
304: the HCN $J = 1 - 0$ hyperfine transitions should have the ratio of
305: 1 : 3 : 5 for $F$=0$-$1 : 1$-$1 : 2$-$1. However, L1333, L1621-1,
306: L1527B1, L1633, L1696B, L1709B2, L429-1, L462-2, L492-2, L694-2,
307: L922-1, CB130-1, and L1197, have almost similar intensities
308: between $F$=0$-$1 and $F$=2$-$1 lines (See Fig. 2f). Furthermore
309: the intensity of $F$=0$-$1 hyperfine line in L694-2, CB130-1, and
310: L1197 is even stronger than that of $F$=2$-$1.
311: About twenty percent of the detected sources shows anomalous intensity ratio among
312: the hyperfine lines. This intensity ratio anomaly is not quite
313: well understood. One of plausible explanations is that the
314: $F$=2$-$1 hyperfine line which have higher opacity and therefore
315: trace the regions where the excitation temperature is lower, while
316: the $F$=1$-$0 lines which have much lower opacity and those trace
317: deeper regions where the excitation temperature is higher, as
318: Cernicharo et al. (1984) and Fuller et al. (1991) have pointed
319: out.
320:
321: \subsection{Comparison of the HCN Spectra with other spectra}
322:
323: The systematic surveys for searching for inward motions in
324: starless cores have been carried out in several molecular lines
325: such as CS $J = 2-1$ (LMT99), $J = 3-2$ , and DCO$^+$ $J = 2-1$
326: (LMP04). Compared with these lines for over 50 sources, the
327: spectra of HCN are found to have following important features.
328:
329: The first distinctive feature of the HCN line is the
330: self-absorption. The HCN line seems to be the most opaque among
331: the tracers so that the double-peaked shape of the spectrum is
332: more often seen in HCN (for more than 20 sources) than any other
333: tracers (about 10 sources in CS $J = 2-1$ and $J = 3-2$ lines).
334: Moreover, the self-absorbed dip of the HCN is more pronounced and
335: the separation between the peaks is wider than those of other
336: tracers. Fig. 3$-$(a) shows the line for L1544 where the
337: self-absorbed dip of the HCN line almost reaches the zero
338: temperature level and the separation between two peaks is the
339: widest among four different tracers.
340:
341: The second noticeable feature of the HCN profile is that the
342: profiles are more asymmetric than other tracers. For example, a
343: red component in the HCN profile L1498 is almost totally absorbed
344: while DCO$^+$ $J = 2-1$ is close to Gaussian, and CS $J = 2-1$ and
345: $J = 3-2$ show a blue peak with red shoulder ({\bf Fig. 3$-$(b)}).
346: HCN profile in L492 shows the double peaked infall asymmetry while
347: other tracers show the skewed asymmetry or Gaussian shape. More
348: than 10 sources show this feature. More asymmetric feature in HCN
349: is also seen in the red profile. For example, the HCN in L1512
350: shows an extreme red profile so that the blue component is nearly
351: self-absorbed out while other lines show either Gaussian
352: (DCO$^+$), or red profile with slight asymmetry (CS $J = 2-1$, and
353: $J = 3-2$) (Fig. 3$-$(c)).
354:
355: The last feature to note in the HCN is that some of the HCN
356: profiles show different asymmetry from other ones, i.e., blue
357: profile in HCN and red profile in other tracers, or vise versa.
358: Furthermore HCN profile itself often shows different asymmetry
359: among three HCN hyperfine lines. Fig. 3$-$(d) displays an example
360: for L183 showing the typical red profile in HCN $F$=2$-$1 (and
361: also in DCO$^+$), but the blue profile in CS $J = 2-1$ and $J =
362: 3-2$ lines. Note that, as shown in Fig. 4, L183 shows a blue
363: profile in HCN $F$=0$-$1, but red profiles in $F$=2$-$1 and
364: $F$=1$-$1. L429-1 is also a similar case: red profiles in HCN
365: $F$=2$-$1 and CS $J = 2-1$ while blue profile in CS $J = 3-2$ and
366: Gaussian one in DCO$^+$. The HCN lines show blue profiles in
367: $F$=0$-$1 and 1$-$1, but a red profile in $F$=2$-$1.
368:
369: \subsection{Analysis of $\delta$V for HCN $J = 1 - 0$ hyperfine transitions }
370:
371: A shape of each HCN hyperfine line, especially a line asymmetry
372: frequently seen in HCN, can be well quantified by estimating the
373: amount by which HCN $J = 1 - 0$ hyperfine spectrum is blue- or
374: red-shifted with respect to a line center of the optically thin
375: tracer defined in the following equation (Mardones et al. 1997):
376:
377: \begin{equation}
378: \rm \delta V = [V_{HCN(F=i-j)}-V_{N_{2}H^{+}}]/ \Delta
379: V_{N_{2}H^{+}},
380: \end{equation}
381: where V$_{{\rm HCN}( F = i-j )}$ is the HCN peak velocity of each
382: hyperfine ($F$=i$-$j) component obtained from a Gaussian fit.
383:
384: The HCN hyperfine lines of double components or a single peak with a shoulder
385: were fitted only for a brighter component with a Gaussian
386: after masking the fainter part or the shoulder
387: to obtain the velocity of the peak component of each line.
388:
389: The peak velocities of N$_{2}$H$^{+}$ lines, V$_{\rm N_{2}H^{+}}$, and
390: the FWHM of N$_{2}$H$^{+}$ $J = 1-0$, $\Delta$V$_{\rm
391: N_{2}H^{+}}$, are taken from LMT99. Note that V$_{\rm N_{2}H^{+}}$
392: was slightly shifted with respect to more accurate frequency
393: determined by LMT01. The velocity shifts for HCN lines for 50
394: sources are calculated and listed in Table 1. Figure 4 shows three
395: histograms for the distribution of $\delta$V for each hyperfine
396: component. For comparison, we also show that of $\delta$V$_{\rm
397: CS}$, which was similarly defined as e.q. (1) for CS lines from
398: LMT99. Distributions of $\delta$V$_{{\rm HCN}(F = 0-1)}$,
399: $\delta$V$_{{\rm HCN}(F = 1-1)}$ and $\delta$V$_{{\rm HCN}(F =
400: 2-1)}$ are clearly skewed to the blue side ($\delta$V$_{\rm HCN}$
401: $<$ 0). Mean values of the distributions are $-0.51 \pm 0.06$ km
402: s$^{-1}$, $-0.59 \pm 0.08$ km s$^{-1}$, and $-0.41 \pm 0.10$ km
403: s$^{-1}$, for $F$=0$-$1, $F$=1$-$1, and $F$=2$-$1 respectively.
404: Note that L429-1 and L183 have different asymmetry among three
405: hyperfine lines, causing the $\delta$V$_{{\rm HCN}(F = 2-1)}$ to
406: fairly differ from other $\delta$V values for two other hyperfine
407: lines. Excluding these two sources, $\delta$V values of three
408: hyperfines ($-0.49 \pm 0.06$ km s$^{-1}$, $-0.56 \pm 0.08$ km
409: s$^{-1}$, and $-0.50 \pm 0.08$ km s$^{-1}$, respectively) become
410: very close. A statistical \emph{t}-test with unknown standard
411: deviation shows that the probability of drawing the observed
412: $\delta$V$_{{\rm HCN} (F = i-j )}$ from the zero mean normal
413: distribution is far less than 1 \% for all three hyperfine lines.
414: Therefore the skewed distribution for all hyperfine lines is
415: statistically very significant. As noted in LMT99, and LMP04, the
416: skewed $\delta V_{{\rm HCN}(F = i-j)}$ distributions to the blue
417: can be produced only if the cores are predominantly in inward
418: motions. Thus we conclude that inward motions of the gas are
419: prevalent in the starless cores. Note that all $\delta$V$_{{\rm
420: HCN}}$ distributions are more skewed than those of CS $J = 2-1$
421: ($<\delta V>= -0.13 \pm 0.04 ~ {\rm km~s^{-1}}$, LMT99) and CS $J
422: = 3-2$ ($<\delta V>= -0.13 \pm 0.03 ~ {\rm km~s^{-1}}$, LMP04),
423: indicating that the hyperfine lines of HCN better trace inward
424: motions in starless cores than CS, as already implied in the last
425: section from more asymmetric characteristics of HCN profiles than
426: those of any other tracers.
427:
428: It is not quite clear whether $\rm \delta V$ distribution of
429: HCN component with lower optical depth is more skewed to the blue
430: side than that of the HCN component with higher optical depth as
431: shown in Figure 4.
432: Rather the mean values of $\delta V$ distributions among all HCN
433: hyperfine components appear to be similar within the error of the
434: distribution.
435: Moreover, Figure 5 shows correlations among $\delta$Vs of each hyperfine
436: line of the HCN. As noted above, L429-1 and L183 show different
437: asymmetry among three hyperfine lines and also greatly affect the
438: quality of correlation among $\delta $Vs of three hyperfine lines.
439: Excluding these two unusual sources, the correlation among the
440: $\delta$Vs of each hyperfine line becomes much better (the
441: correlation coefficient of the fit becomes better than 0.8).
442: This may imply that the gas in starless cores has
443: inward motions in a systematic rather than random fashion across
444: the different regions within a factor of optical depths 1 $\sim$
445: 5.
446:
447: \subsection{Infall Candidates}
448:
449: In this section we identify infall candidates based on the
450: spectral properties of the HCN line profiles together with those
451: of other lines (CS $J = 2-1$ and N$_{2}$H$^{+}$ $J = 1-0$ from
452: LMT99, and CS $J = 3-2$ and DCO$^{+}$ $J = 2-1$ from LMP04). The
453: spectral properties that we adopt for this purpose are (1) values
454: of $\delta$V for HCN $J = 1-0$ main component $F$=2$-$1 from this
455: study, for CS $J = 2-1$ and N$_{2}$H$^{+}$ from LMT99, and CS $J
456: = 3-2$ and DCO$^{+}$ from LMP04, and (2) peak intensity ratios of
457: the blue (T$_{b}$) to the red (T$_{r}$) component in double-peaked
458: spectra for HCN $F$=2$-$1 from this study, for CS $J = 2-1$ and
459: N$_{2}$H$^{+}$ from LMT99, and for CS $J = 3-2$ and DCO$^{+}$ from
460: LMP04.
461:
462: The values of $\rm \delta V$ enable us to decide how much the
463: profiles are shifted with respect to the peak velocity of the
464: optically thin tracer. However, $\delta$V is sometimes affected by
465: the weak N$_{2}$H$^{+}$ $J = 1-0$ emission, even in the case that
466: the optically thick tracers have a double peaked infall asymmetry.
467: Peak intensity ratios are used to add more information in this
468: case. Moreover the peak intensity ratios select other cases where
469: the higher density tracers such as N$_{2}$H$^{+}$ and DCO$^{+}$
470: show a double peaked infall asymmetry. This is why we adopt
471: multiple spectral properties to choose the infall candidates.
472:
473: The $\delta$V value is flagged with `1' for $\delta V \leq
474: -5\sigma_{\rm \delta V}$, `0' for $\mid \delta V \mid < 5
475: \sigma_{\rm \delta V}$, and `-1' for $\delta V \geq 5 \sigma_{\rm
476: \delta V}$. A good infall candidate should be given a number `1'.
477:
478: In a similar way the intensity ratio is also flagged with `1' for
479: T$_{b}$/T$_{r}$ $>$ 1+ $\sigma_{\rm T_{b}/T_{r}}$, `0' for
480: $1-\sigma_{\rm T_{b}/T_{r}}$ $<$ T$_{b}$/T$_{r}$ $<$
481: $1+\sigma_{\rm T_{b}/T_{r}}$, and `-1' for T$_{b}$/T$_{r}$ $<$ $1-
482: \sigma_{\rm T_{b}/T_{r}}$, where the values of T$_{b}$ and T$_{r}$
483: were obtained from Gaussian fits of blue and red components of the
484: profile. Again, a dense core flagged with 1 should be considered
485: to be an infall candidate. The above criteria used for selecting
486: infall candidates are similar to those used by LMT99 and LMP04,
487: but have more data from this HCN study.
488:
489: Table 2 summarizes ``color codes" of each source coupled with
490: above two spectral properties. These values can be used as
491: quantitative indicators for the selection of infall candidates
492: more accurately. Each source can be given the total of 9 color codes
493: although most of color codes for property (2) of DCO$^{+}$ and
494: N$_{2}$H$^{+}$ remain unassigned because these lines are usually
495: Gaussian except for several cores, such as L1521F, L1544, L492,
496: L694-2 and L1197. In the last column we list the total color codes
497: which are the algebraic sum of all color codes. We defined
498: ``Probable'' infall candidate which has a total color codes
499: between 3 $\sim$ 4. Such candidates are B18-3, TMC2, TMC1, L1522, L1696A,
500: L1696B, L1689B, L234E-S, L63 and L1197. L183 was classified as a
501: possible infall candidate in LMP04 and a strong infall candidate
502: in LMT99. But it is not a strong infall candidate in our study
503: because it shows minus signs of asymmetry from both spectral
504: properties in the HCN. On the other hand, L63 and L1197 were not
505: grouped as infall candidate in LMP04 and LMT99, but they are
506: probable candidates from both spectral properties in this paper.
507: We classified 7 strong infall candidates which have a total color
508: code $ > $ 4: L1355, L1498, L1521F, L1544, L492, L694-2, and
509: L1155C1. All the strong infall candidates have been classified as
510: previously in LMP04, LMT99.
511:
512:
513: \section{DISCUSSION}
514:
515: \subsection{Infall Structure}
516: As noted in the previous section, the HCN lines seem to be the
517: best tracer for the study of infall kinematics in the core not
518: only because they show infall asymmetry with clear dip more often
519: than other tracers, but also because they are composed of three
520: hyperfine lines with different opacities. These features enable us
521: to trace different layers of the core and makes it useful to study
522: infall structure along the line of sight of the core. Probably the
523: hyperfine line $F$=2$-$1 with the highest opacity would trace the
524: far side from the center of the core and the other hyperfine line
525: $F$=0$-$1 with the least opacity would trace the closer side from
526: the center of the core. So, if the hyperfine lines show the infall
527: asymmetry, then we may be able to see the variation of infall
528: velocities with the distance from the center by {\bf analyzing}
529: the HCN hyperfine lines. For this purpose, we choose the infall
530: candidates showing infall asymmetry in at least two hyperfine
531: lines and derive peak intensity ratios of the blue (T$_{b}$) to
532: the red (T$_{r}$) components in HCN hyperfine lines of the infall
533: candidates. Suppose that the excitation temperature increases
534: toward the core, infall speed of the gaseous material is small
535: (less than the velocity dispersion of the gas in the core), and
536: the central dip in the infall asymmetry profile is formed by the
537: absorption of the emission lines from the approaching warm part
538: near the central part of the core by the cool envelope receding
539: from an observer. Then the ratio of blue to red components should
540: be proportional to the gaseous infall speed. So with the ratios of
541: the blue to the red components, it may be possible to deduce
542: qualitative features of the infall speed in the core, although it
543: may not be possible to give any explicit form of the variational
544: infall speed with radius. Fig. 6 plots the intensity ratios of the
545: blue to the red components in HCN for each infall candidates,
546: showing that four sources such as L63, L492, L694-2, and L1197 out
547: of 12 infall candidates have larger ratio in the hyperfine line
548: with the lower opacity. If the excitation temperature increases
549: toward the center, these four cores may have increasing infall
550: speeds toward center of the cores.
551:
552: However, the infall structure of the core and observational data
553: are not so simple in practice. LMP04 have shown from their CS
554: surveys that CS $J = 3-2 $ usually traces inner denser region in
555: faster inward motions than CS $J = 2-1$, indicating that infall
556: structures are developed in several cores (L1544, L1521F, L1445,
557: L183, L158, L1689B, and L694-2). Williams, Lee, \& Myers (2006)
558: also find faster inward motions in the inner denser region for
559: L1544 and L694-2 using their N$_{2}$H$^{+}$$ J = 1-0 $ data.
560: However, our HCN hyperfine lines have revealed such a structure
561: toward the L694-2 only.
562:
563: Because the optical depths and excitation conditions for the
564: tracers are different among the sources, different tracers would
565: trace different regions of the cores. Therefore it would be
566: natural not to expect the detection of the same features of inward
567: motions in every core with different tracers.
568:
569: Ambipolar Diffusion Models (e.g., Ciolek \& Basu 2000) usually
570: predict increasing infall speed toward the center up to certain
571: radius beyond which the infall speed decreases. The turning point
572: and the magnitude of the infall speed are dependent upon the
573: evolutionary stage of the core. Because the spectral line profile
574: always carries the average information through the line of sight,
575: we may be watching such a complication.
576:
577: A detailed Monte Carlo radiative transfer model may help to deduce
578: the infall structure of the core from the complicated line
579: profiles. For example, Keto et al. (2004) made an effort to derive
580: the information of the density, velocity, and molecular abundance
581: of three starless cores L1544, L1489 and L1517B. A Monte Carlo
582: radiative transfer calculation by Lee et al. (2007) was
583: able to fit the infall profiles of the HCN hyperfine lines for
584: L694-2 and L1197 assuming approximate Bonner-Ebert density
585: distribution and $\Lambda$ shaped infall velocity structure
586: (increasing inward infall velocity to a point somewhere
587: between the center and the envelope of the core from where the
588: infall velocity starts to decrease toward the center of the core).
589:
590: The mapping of the cores with HCN hyperfine lines will also help
591: us to understand spatial variation of the infall speed of the
592: cores to the direction orthogonal to the line of sight of the
593: core. Thus combining the mapping data and the detailed Monte Carlo
594: radiative transfer model would be very useful for better
595: understanding of the infall structure of the cores. The study in
596: this direction is underway.
597:
598: \subsection{Less depletion of HCN and its effect}
599:
600: Whether the HCN depletes out in the central region of the core is
601: an important factor in studying the inward motions in the cores.
602: We measure the line intensities of the HCN $F$=0$-$1 components
603: with the least optical depths among three hyperfine components for
604: 36 sources showing the least self-absorbed features, to compare
605: with those of N$_{2}$H$^{+}$ which hardly suffers from the
606: molecular depletion in most cores and generally well traces
607: the density distribution of the cores (Tafalla et al. 2002; Bergin
608: \& Langer 1997). Figure 7 shows the correlation between the
609: integrated intensities of each HCN hyperfine line and
610: N$_{2}$H$^{+}$ line.
611: The intensity of N$_{2}$H$^{+}$ line is the total integrated intensity
612: of its seven hyperfine components which is adopted from LMT99.
613: The integrated intensity of each HCN hyperfine line lines is simply
614: a summation of the channel intensity multiplied by
615: the channel width over each hyperfine line profile.
616:
617: It appears that HCN $F$=0$-$1 integrated intensities
618: have a very good correlation (correlation coefficient \emph{r} =
619: 0.82) with those of N$_{2}$H$^{+}$. We believe that a slight
620: deviation from a tight relation between the intensities of two
621: lines is partly due to a rather large optical depth of the HCN
622: $F$=0$-$1 component. Our $\rm H^{13}CN$ observations enable us to
623: estimate its optical depths of $1\sim 3$ for 8 sources. The worse
624: correlations of the intensities of the other HCN components with
625: higher optical depths (Fig. 7-b and c) support this
626: possible effect of high optical depth on the HCN lines.
627:
628: Our follow-up mapping observations in HCN also support the idea
629: that the HCN is less affected from the molecular depletion in the
630: central region of the core, from the fact that HCN $F$=0$-$1 has
631: similar spatial distribution to that of N$_{2}$H$^{+}$ (Sohn et
632: al. 2007 in prep.). Altogether the HCN suffers from less depletion
633: in the central region and therefore it is thought to trace inner
634: regions of the cores better than the CS.
635: This may result in the
636: higher occurrence of infall motions of gaseous molecules and more
637: negative $\delta V$ for HCN than that of CS.
638:
639: As another evidence for the HCN to be a good tracer of the central region
640: of the core, one can suggest some possible correlation between the degree of the self-absorption
641: in the HCN line and the optical depth of the $\rm N_2H^+$ line.
642: We tested the relation in 17 sources showing a clear self-absorption in HCN(1-0) hyperfine
643: by measuring $\rm (T_{peak} -T_{dip}) / T_{peak}$ of the HCN profiles and the ratio of the line widths of
644: the emission profile and the absorbed feature, as indicators of the amount of the self-absorption.
645: Here $\rm T_{peak}$ is the peak temperature
646: of the observed HCN spectrum and $\rm T_{dip}$ is the intensity of the self-absorbed dip.
647: These quantities were compared with the tau values of $\rm N_2H^+$ line taken from LMT09,
648: but there were found to be no clear correlation between them.
649: In fact the degree of self-absorption itself in a molecular line toward
650: the cores can be affected not only by the optical depth of the tracer,
651: but also by other properties such excitation conditions and kinematics in the cores
652: like infall motions.
653: In addition, once a spectrum is self-absorbed, it is practically difficult to measure
654: physical quantities of the real dip and the unabsorbed profile, without employing a proper model
655: dealing with radiation processes in the core which
656: is actually beyond the scope of this paper.
657: We suggest that at least the property of the less depletion of the HCN
658: in the inner region of the core
659: can be a preferable evidence for the HCN to be a useful tracer of the central regions of the core.
660:
661: \section{SUMMARY}
662: We performed a systematic pointing survey of starless cores in HCN
663: and/or $\rm H^{13}CN$ $ J = 1-0 $ hyperfine transitions to study
664: inward motions in starless cores. Sixty-four and ten of 85 targets
665: were detected in HCN and $\rm H^{13}CN$ lines, respectively. Main
666: results obtained from the survey are as follows;
667:
668: 1. Most of the HCN spectra detected toward the starless cores are
669: found to exhibit asymmetric features such as the blue asymmetry
670: (43 \%), the red asymmetry (20 \%), and the mixture of blue and
671: red asymmetry in the hyperfine lines (33 \%).
672:
673: 2. The HCN lines show the highest occurrence of double peaked
674: features (for more than 30 \% of the detected cores), more
675: asymmetric feature, and less depletion toward the central region
676: of the cores than other infall tracers. Especially three HCN
677: hyperfine lines with different optical depths are useful to study
678: kinematics of the different regions of the cores. Altogether HCN
679: lines are found to be one of the best infall tracers.
680:
681: 3. The statistical distribution of the velocity shift $\delta V$
682: of the HCN profiles with respect to the systemic velocity of the
683: optically thin tracer is found to be significantly shifted to the
684: blue side, implying the predominance of the inward motions in the
685: starless cores. The degree of the skewness in the distribution is
686: greater than that of any other infall tracers, indicating that the
687: HCN traces more frequently inward motions. The $\delta V$
688: distributions of HCN hyperfine lines are similar within the error.
689: Moreover the $\delta V$ values of HCN hyperfine lines for each
690: source are also similar. These may mean that most of starless
691: cores have inward motions in a systematic rather than random
692: fashion across the layers of the cores.
693:
694: 4. This survey enables us to identify infall candidates for
695: further study in the future, using all available ``color codes''
696: such as the velocity shift $\delta V$ and the blue to red peak
697: intensity ratio of double peaked profiles for HCN $ J = 1-0 $, CS
698: $ J = 2-1 $ , $ J= 3-2$, DCO$^+$ $ J = 2-1 $, and $\rm N_2H^+$ $ J
699: = 1-0 $. We identify 7 strong infall candidates: L1355, L1498,
700: L1521F, L1544, L492, L694-2, and L1155C1, and 10 probable infall
701: candidates: B18-3, TMC2, TMC1, L1552, L1696A, L1696B, L1689B, L234E-S,
702: L63, and L1197.
703:
704: 5. We find four infall candidates such as L63, L492, L694-2, and
705: L1197 with the higher blue to red ratio in the hyperfine line of
706: the lower opacity, suggesting that they may be in a stage of
707: development of infall structure of increasing infall speeds
708: inside.
709:
710: \acknowledgments
711:
712: This research was supported by the KOSEF grant
713: No. R14-2002-058-01000-0 of the Korea Science and Engineering
714: Foundation and in part by KOSEF grant No. KOSEF
715: R01-2003-000-10513-0 of the Korea Science and Engineering
716: Foundation. The authors express special thanks to the staffs of
717: Taeduk 14m observatory for their dedicated support.
718:
719: \begin{thebibliography}{}
720: \bibitem [ ]{ } Aikawa, Y., Herbst, E., Roberts, H., \& Caselli, P.,
721: 2005, \apj, 620, 330
722: \bibitem[ ]{ } Alves, J., Lada, C. J., \& Lada, E. F. 2001,
723: Nature, 409, 159
724: \bibitem[ ]{ } Andr\'{e}, P., Ward-Thomson, D., \&
725: Motte, F. 1996, A\&A, 314, 625
726: \bibitem[ ]{ }Beichman, C. A., Myers, P.C., Emerson, J.
727: P., Harris, S., Mathieu, R., Benson, P. J., \& Jennings, R.E.
728: 1986, \apj, 307, 337
729: \bibitem[ ]{ }Benson, P. J., \& Myers, P.C. 1989, ApJS, 71,
730: 89
731: \bibitem[ ]{ }Bergin, E. A., \& Langer, W. D. 1997, \apj, 486, 316
732: \bibitem[ ]{ } Buisson, G., Desbats, L., Duvert, G.,
733: Forveille, T., Gras, R., Guilloteau, S., Lucas, R., \& Valiron, P.
734: 1994, CLASS Manual (Granoble:IRAM)
735: \bibitem[ ]{} Caselli , P., Walmsley, C.M., Tafalla, M., Dore, L.,
736: \& Myers, P.C. 1999, \apj, 523, L165
737: \bibitem[ ]{} Cernicharo, J., Castets, A., Duevrt, G., \&
738: Guilloteau, S. 1984, A\&A, 139, 13
739: \bibitem[ ]{} Charnley , S. B., 1997, MNRAS, 291, 455
740: \bibitem[ ]{} Ciolek, G. E., \& Basu, S., 2000, \apj, 529, 925
741: \bibitem[ ]{} Crapsi, A., Caselli, p., Walmsley, C. M., Myers, P. C., Tafalla, M., et al.,
742: 2005, \apj, 619, 379
743: \bibitem[ ]{} Di Francesco, J., Evans II, N. J., Caselli, P.,
744: Myers, P. C., Shirley, Y., Aikawa, Y., \& Tafalla, M., 2006,
745: astro-ph/0602379
746: \bibitem[ ]{} Fuller, G.A., Myers, P.C., Welch, W.J., Goldsmith, P.F.,
747: Langer, W.D., Campbell, B.G., Guilloteau, S., \& Wilson, R.W.
748: 1991, \apj, 376, 135
749: \bibitem[ ]{} Keto, E., Rybicki, G. B., Bergin, E. A., \& Plume,
750: R. 2004, \apj, 613, 355
751: \bibitem[ ]{} Lee, C.W., Myers, P.C., \& Plume, R. 2004, ApJS, 153, 523L
752: (LMP04)
753: \bibitem[ ]{} Lee, C. W., \& Myers, P. C. 1999, ApJS, 123, 233 (LM99)
754: \bibitem[ ]{} Lee, C. W., Myers, P. C., \& Tafalla, M. 1999, \apj, 526, 788
755: (LMT99)
756: \bibitem[Lee et al.(2001)]{lee01} Lee, C. W.,
757: Myers, P. C., \&, Tafalla, M. 2001, \apj S, 136, 703 (LMT01)
758: \bibitem[]{} Lee, J.-E., Bergin, E. A., \&, Evans II, N. J. 2004, \apj, 617,
759: 360
760: \bibitem[ ]{} Lee, S. H, Park, Y-S, Sohn, J., Lee, C. W., \& Lee,
761: H. M. \apj, in press
762: \bibitem[ ]{} Leger, A. 1983, A\&A, 123, 271L
763: \bibitem[ ]{} Mardones, D., Myers, P.C., Tafalla, M., Wilner, D.J.,
764: Bachiller, R., \& Garay, G. 1997, \apj, 489, L719
765: \bibitem[ ]{} Myers, P. C. 2005, \apj, 623, 280
766: \bibitem[ ]{} M\"uller, H. S. P., Schlo\"der, F., Stutzki, J.,
767: \& Winnewisser, G. 2005, J. Mol. Struct. 742, 215
768: \bibitem[ ]{} Park, Y. -S, Kim, J. \& Minh. Y. C. 1999, ApJ, 520, 223
769: \bibitem[ ]{} Roh, D. K., \& Jung, H. S. 1999, PKAS, 14, 123
770: \bibitem[ ]{} Shirley, Y., Evans, N.J., Rawlings, J.M., \&
771: Gregersen, E.M. 2000, ApJS, 131, 249
772: \bibitem[ ]{} Shu, F. H. 1997, \apj, 214, 488
773: \bibitem[ ]{} Tafalla, M., Mardones, D., Myers, P. C., Caeslli, P.,
774: Bachiller, R., \& Benson, P. J. 1998, \apj, 504, 900
775: \bibitem[ ]{} Tafalla, M., Myers, P. C., Caselli, P., Walmsley, C. M.,
776: \& Comito, C. 2002, \apj, 569, 815
777: \bibitem[ ]{} Tafalla, M., Santiago-Garc\'{i}a, J., Myers, P.
778: C., Caselli, P., Walmsley, C. M., \& Crapsi, A. 2006, A\&A, 455,
779: 577
780: \bibitem[ ]{} Ward-Thompson, D., Scott, P. K., Hills, R. E.,
781: \& Andre, P. 1994, MNRAS, 268, 276
782: \bibitem[ ]{} Whitworth, A. P. \& Ward-Thompson, D. 2001, \apj, 547, 317
783: \bibitem[ ]{} Williams, J. P., Lee, C. W., \& Myers, P. C. 2006, \apj, 636,
784: 952
785: \end{thebibliography}
786:
787:
788: %\clearpage
789:
790: %\input{tab1.tex}
791:
792: %\clearpage
793:
794:
795: \begin{figure}[btp]
796: \epsscale{0.8} \plotone{f1-1.ps} \caption{HCN $J= 1-0$ profiles
797: for 64 sources with strong detection. All profiles are shifted so
798: that the $\rm V_{N_2H^+}$ or Gaussian fit velocity of HCN
799: $F$=2$-$1 component, in case of no existence of $\rm V_{N_2H^+}$,
800: is zero. \label{fig1}}
801: \end{figure}
802:
803: \begin{figure}[btp]
804: \epsscale{0.9}
805: \plotone{f1-2.ps}
806: \end{figure}
807:
808: \begin{figure}[btp]
809: \epsscale{0.9}
810: \plotone{f1-3.ps}
811: \end{figure}
812:
813: \begin{figure}[btp]
814: \epsscale{0.9}
815: \plotone{f1-4.ps}
816: \end{figure}
817:
818:
819: \begin{figure}[btp]
820: \plotone{f2.ps} \caption{Six prototypical HCN $J=1-0$ spectra. (a)
821: All three components have two peaks, with the blue peak brighter
822: than the red. (b) One or two of hyperfines has two peaks or single
823: peak of the infall asymmetry. (c) At least one of hyperfines has
824: red asymmetry. (d) Red asymmetry and blue asymmetry are mixed in
825: three hyperfines. (e) The line profiles are broad. (f) Hyperfine
826: intensity ratios are anomalous. \label{fig3}}
827: \end{figure}
828:
829:
830: \clearpage
831: \begin{figure}[btp]
832: \plotone{f3.ps} \caption{HCN profiles compared with those of other
833: infall tracers. (a) The HCN profile is usually the most opaque
834: among other infall tracers. (b) Either the blue profiles or (c)
835: the red profiles of the HCN show more asymmetric features. (d) The
836: HCN profiles sometimes show totally revered asymmetry to those of
837: other line profiles. \label{fig4}}
838: \end{figure}
839:
840:
841: \begin{figure}[btp]
842: \epsscale{0.7} \plotone{f4.ps} \caption{Histograms of the
843: normalized velocity difference ($\delta V_{\rm HCN \ \it J \rm
844: =1-0}$) between $V_{\rm HCN(F=i-j)}$ and $V_{\rm N_{2}H^{+} \ \it
845: J \rm =1-0}$ for a sample of 49 starless cores. The distribution
846: of $\rm \delta V_{CS \ \it J \rm =2-1}$ from LMT99 is also given
847: after adopting the frequency set for CS and $\rm N_2H^+$
848: determined from LMT01. The $\rm \delta V$ distribution of each HCN
849: hyperfine transition appears to be similar, but more skewed to the
850: blue than that of $\rm \delta V_{CS \ \it J \rm =2-1}$.
851: \label{fig4}}
852: \end{figure}
853:
854: \begin{figure}[btp]
855: \epsscale{1.0} \plotone{f5.ps} \caption{Correlations among
856: $\delta$Vs of each HCN hyperfine line. The solid lines indicate
857: the line for perfect correlation between different hyperfines. The
858: numeric number at right of each panel is a correlation coefficient
859: of the fit. Parenthesized numbers are the correlation coefficients
860: without L183 and L429-1. \label{fig5}}
861: \end{figure}
862:
863: \clearpage
864: \begin{figure}[btp]
865: \epsscale{0.9} \plotone{f6.ps} \caption{Intensity ratios of the
866: blue (T$_{b}$) to the red (T$_{r}$) components in HCN hyperfine
867: lines. Note that four infall candidates, L63, L492, L694-2, and
868: L1197 show the larger value for T$_{b}$/T$_{r}$ in the hyperfine
869: line with the lower opacity. \label{fig6}}
870: \end{figure}
871:
872: \clearpage
873: \begin{figure}[btp]
874: \epsscale{0.3} \plotone{f7.ps} \caption{The comparison between the
875: integrated intensities of each HCN $J=1-0$ hyperfine and
876: N$_{2}$H$^{+}$ $J=1-0$. The diagrams show that the intensity of
877: the HCN ($F$=0$-$1) has better correlation (correlation
878: coefficient \emph{r} = 0.82) with the intensity of N$_{2}$H$^{+}$
879: than those of other HCN hyperfine lines. \label{fig7}}
880: \end{figure}
881:
882: \end{document}
883: