HI Narrow-Line Self-Absorptions Toward the High-Mass Star-Forming Region G176.51+00.20

Using the Five-hundred-meter Aperture Spherical radio Telescope (FAST) 19-beam tracking observational mode, high sensitivity and high-velocity resolution HI spectral lines have been observed toward the high-mass star-forming region G176.51+00.20. This is a pilot study of searching for HI narrow-line self-absorption (HINSA) toward high-mass star-forming regions where bipolar molecular outflows have been detected. This work is confined to the central seven beams of FAST. Two HINSA components are detected in all seven beams, which correspond to a strong CO emission region (SCER; with a velocity of $\sim$ $-$18 km s$^{-1}$) and a weak CO emission region (WCER; with a velocity of $\sim$ $-$3 km s$^{-1}$). The SCER detected in Beam 3 is probably more suitably classified as a WCER. In the SCER, the HINSA is probably associated with the molecular material traced by the CO. The fractional abundance of HINSA ranges from $\sim 1.1 \times 10^{-3}$ to $\sim 2.6 \times 10^{-2}$. Moreover, the abundance of HINSA in Beam 1 is lower than that in the surrounding beams (i.e., Beams 2 and 4--7). This possible ring could be caused by ionization of HI or relatively rapid conversion from HI to H$_2$ in the higher-density inner region. In the WCER (including Beam 3 in the SCER), the HINSA is probably not associated with CO clouds, but with CO-dark or CO-faint gas.


INTRODUCTION
Stars form in molecular clouds, and it is important for star formation to synthetize molecular H 2 from atomic H. However, it is difficult to directly extract individual components from H i emission and correlate them with molecular clouds (Li & Goldsmith 2003;Liu et al. 2022). In such conditions, H i self-absorption (HISA) lines have become a very useful tracer of H i. HISA occurs in environments where cold H i gas is in the front of a warmer emission background (e.g. Knapp 1974); such conditions are ubiquitous in the Milky Way (e.g., Heeschen 1955;Gibson et al. 2000;Wang et al. 2020). HISA is associated with both CO-dark clouds and molecular clouds with strong CO emission (e.g., Hasegawa et al. 1983;Gibson et al. 2000;Gibson 2010). H i narrow-line self-absorption (HINSA), a special case of HISA, is usually associated with CO clouds, and its linewidth is comparable or smaller than that of CO (see Li & Goldsmith 2003;Goldsmith & Li 2005;Tang et al. 2016;Wang et al. 2020). HINSA has been found to be an excellent tracer of molecular clouds (Li & Goldsmith 2003), and this tight correlation has been confirmed by numerous studies (e.g., Krčo & Goldsmith 2010;Tang et al. 2020;Liu et al. 2022).
Numerous surveys have provided abundant knowledge about the environment and physical properties of the regions producing HINSA. For instance, the detection rate of HINSA is ∼ 77% (Li & Goldsmith 2003) in optically selected dark clouds (Lee & Myers 1999), ∼ 80% (Krčo & Goldsmith 2010) in molecular cores (i.e., the Lynds dark cloud, which has a peak optical extinction of 6 mag or more; see the catalog of dark clouds in Lynds 1962), and ∼ 58%  to ∼ 92% (Liu et al. 2022) in Planck Galactic cold clumps (PGCCs; cold and quiescent clumps in very early evolution stages of star formation; see Planck Collaboration et al. 2011;Wu et al. 2012;Planck Collaboration et al. 2016). These results indicate that HINSA is associated with molecular clouds, especially molecular cores and/or clumps. These papers also constrained the abundance of HINSA, which ranges from ∼ 10 −4 to ∼ 10 −2 (e.g., Krčo & Goldsmith 2010;Tang et al. 2020;Liu et al. 2022); the relationship of the central velocity, linewidth, and column density between HINSA and the molecular gas traced by 12 CO, 13 CO, OH, etc. (e.g., Li & Goldsmith 2003;Tang et al. 2020;Liu et al. 2022) and the atomic gas traced by CI (e.g., Li & Goldsmith 2003); and even a ring of enhanced HINSA abundance inside of a dark molecular cloud (i.e., in B227; see Zuo et al. 2018).
Star-forming regions characterized by molecular outflows have also been incorporated into samples arising from the search for HINSA (e.g., Liu et al. 2022). In this study, the corresponding detection rate of HINSA was only 25% (i.e., one in four sources characterized by molecular outflows). As such, one of our goals is to enlarge the sample of known star-forming regions characterized by molecular outflows. A good choice is the nine regions where outflowing gases traced by 12 CO, 13 CO, HCO + , and CS have been recently detected with high sensitivity (i.e., main beam rms noise of tens of mK) with the 13.7 m millimeter telescope of the Purple Mountain Observatory in Delingha by Liu et al. (2021).
In one of the nine regions, i.e., G176.51+00.20, a neutral stellar wind traced by atomic hydrogen (H i wind) has also been detected (Y. J., Li et al., 2022, ApJ, accepted). Therefore, studying HINSA toward G176.51+00.20 not only would be a bridge to connect the atomic and molecular gas therein, but may also be helpful for studying the relationship between H i winds and molecular outflows.
G176.51+00.20, an active high-mass star-forming region, is located 1.8 kpc from Earth (Moffat et al. 1979;Snell et al. 1988). The dense NH 3 core in the center of this region has become synonymous with this region (e.g., Torrelles et al. 1992a;Chen et al. 2003;Jiang et al. 2013). The central engine of this massive star-forming region was identified as a zero-age main-sequence B3 star by using a 3.6 cm map produced from Very Large Array observations (Torrelles et al. 1992b). A more detailed description of this region was provided by Dewangan (2019), where the center of the region of study in this work was marked as "H ii region" in their figure 1(a).
The remainder of this paper is organized as follows. In Section 2, we describe the data used in this work. Section 3 presents the results of the HINSA survey toward G176.51+00.20, and the physical properties of the HINSA features and those of the corresponding molecular lines. In Section 4, we discuss the association between the HINSA features and the molecular clouds, and estimate the abundance of the HINSA features, and summarize our results.

H i Observations with FAST
The most sensitive ground-based single-dish radio telescope is the Five-hundred-meter Aperture Spherical radio Telescope 1 , which located in Guizhou Province of southwest China (Nan 2006;Nan et al. 2011;Qian et al. 2020). By using the 19-beam receiver equipped on FAST (see Li et al. 2018;Jiang et al. 2019Jiang et al. , 2020, we observed highly sensitive and high-velocity resolution H i spectra toward G176.51+00.20. The mean main beam rms noise for the 19 obtained spectra is ∼ 7 mK @ 0.1 km s −1 resolution. The observations were conducted on August 19th and 20th, 2021, with total integration time is 335 minutes. The half-power beam width (HPBW) is ∼ 2.9 at 1.4 GHz, and the pointing error is ∼ 0.2 (see Jiang et al. 2019Jiang et al. , 2020  in 2020 May (for more details, see tables 2-3 of Liu et al. 2021). The velocity resolution of these lines ranges from 0.159 to 0.212 km s −1 (the corresponding channel width is 61 kHz), and the HPBW ranges from 49 to 61 . The lines were observed with the on-the-fly mode, and gridded to 30 pixels.
The rms noises range from ∼ 60.9 to ∼ 15.3 mK (for more details see Liu et al. 2021).

DATA ANALYSIS AND RESULTS
Because the molecular line observations are confined to the central ∼ 15 × 15 region centered at (R.A., Decl.) (J2000) = (05 h 37 m 52 s .1, 32 • 00 03 .0), the HINSA survey is also limited to the corresponding regions; i.e., central seven beams (Beams 1-7; see Figure 1(a)). We followed the method of Liu et al. (2022) to extract HINSA features. To extract HINSA, it was required to understand the physical properties of the emission regions, at least the central velocity, v c,m , velocity dispersion, σ m , and excitation temperature, T ex (i.e., the initial conditions to extract HINSA; see Liu et al. 2022). These three physical parameters were obtained by using the high-sensitivity molecular line observations of the high-mass star-forming region G176.51+00.20 made by Liu et al. (2021).

Overview of the Region with High-Sensitivity Molecular-Line Observations
From the study of molecular outflows toward G176.51+00.20 by Liu et al. (2021), the velocity of the most conspicuous components is ∼ −18 km s −1 , which corresponds to a high-mass star-forming region with a strong bipolar outflow (see Snell et al. 1988;Zhang et al. 2005;Liu et al. 2021). In fact, there is also a weak component with a velocity of ∼ −3 km s −1 in the region close to the edge of the observed region. Throughout this work, we have divided the observed region into two parts: a strong CO emission region (hereafter SCER) with a velocity of ∼ −18 km s −1 and a weak CO emission region (WCER) with a velocity of ∼ −3 km s −1 . Figure 1 presents an integrated intensity map of the WCER (left) and the SCER (right). In the WCER, the 13 CO emission is weak and no C 18 O emission is detected. Therefore, only the integrated intensity map of 12 CO, with a velocity range of [−5, −1] km s −1 , is presented here. There are three peaks in the CO map of the WCER; i.e., Peaks 1-3 (see the labels in Figure 1(a)). In the SCER, there is strong 12 CO (blue contours), 13 CO (red contours), and C 18 O (background map) emission, where the velocity range of the maps is [−30, −10] km s −1 . Because the various emission components are all connected, and it is difficult to distinguish each of them clearly, we marked the position centered at the peak intensity of C 18 O as Peak 4. We note that this large emission region covers, in fact, almost the entire observed region (including the regions covered by all seven FAST beams; see below). In the WCER, because the line emission of 13 CO is weak, and lines from other molecules (i.e., C 18 O, HCO + and CS) are not detected, we only conducted Gaussian fits for the 12 CO spectra. The values of v c,m and σ m are given based on those Gaussian fits. The peak brightness temperatures of the 12 CO spectra, T 12 CO peak , were used to calculate the value of T ex (e.g., Garden et al. 1991) as (1) The results of v c,m , σ m , and T ex , with 1σ errors, are −3.92 ± 0.01, 0.35 ± 0.01 km s −1 , and 7.18 ± 0.18 K for Peak 1, and −2.20 ± 0.01, 0.44 ± 0.01 km s −1 , and 4.76 ± 0.16 K for Peak 2, and −2.54 ± 0.04, 0.42 ± 0.04 km s −1 , and 5.08 ± 0.31 K for Peak 3, respectively. In the SCER, the value of T ex was also calculated based on T 12 CO peak . The values of v c,m and σ m are based on the Gaussian fit of the 13 CO spectrum. The values of v c,m , σ m , and T ex are −18.53 ± 0.01, 1.13 ± 0.01 km s −1 and 20.30 ± 0.08 K, respectively. These values were used to extract the HINSA features.

HINSA Features
To extract as many HINSA features as possible, we assumed that, for the spectrum of each FAST beam, HINSA features corresponding to each of the four components of the molecular clouds (i.e., Peaks 1-4) exist. The specific description of the method used to extract the HINSA feature is presented in Section A.
We extracted fourteen HINSA features in total from the seven beams (see Figure 2). Their velocity, V LSR , peak intensity, T ab , optical depth, τ 0 , and velocity dispersion, σ HI , with 1σ errors are listed in Table 1.
The fitted values of τ 0 and σ HI were used to calculate the column density of the HINSA features, N HI (see Li & Goldsmith 2003) as where ∆V = 8 ln(2)σ HI . The results of N HI and the corresponding T ex with 1σ errors are listed in Table 1. The values of N HI range from ∼ 5.3 to 33.1 × 10 18 cm −2 in the SCER, and from ∼ 0.5 to 3.4 × 10 18 cm −2 in the WCER.

Physical Properties of the Molecular Line Emission
To make a comparison between the HINSA features and the molecular line emission, we resampled the molecular data observed by Liu et al. (2021) to pixel sizes of 3 × 3 , and aligned pixel center of molecular emission with the corresponding FAST beam center. the SCER covers almost the entire observed region of the molecular lines. The line centers and velocity dispersions of the molecular lines are from the Gaussian fits to the 12 CO, 13 CO, and/or C 18 O spectra. We also calculated the column density of all observed molecular species, i.e., 12 CO, 13 CO, C 18 O, HCO + , and CS when these were detected (see Section B). The velocity center, velocity dispersion, and column density as well as the corresponding T ex , are listed in Table A1.  The association between the HINSA features and the molecular line emissions can be seen in Figure   3. In the SCER (shown as the first entry for each beam in Table 1), the HINSA features are probably associated with the molecular line emission. In the WCER (shown as the second entry in Table 1), no molecular line emission associated with HINSA features is detected toward Beams 1, 3, 4, and 6.
The only detected 12 CO emission, corresponding to Beams 2, 5, and 7, is also weak. The HINSA The red, green, blue, black, and yellow lines show the spectra of 12 CO, 13 CO, C 18 O, HCO + , and CS, respectively, and the purple lines represent the HINSA features. We fitted the profiles of both 12 CO and 13 CO with a Gaussian profile for each component, except for spectra that showed strong self-absorption or were contaminated by other components. For Beam 6, only the line emission of C 18 O was fitted because there were strong self-absorption in the spectra of 12 CO and 13 CO.
features in the WCER are probably associated with CO-dark or CO-faint gas (e.g., Wolfire et al. 2010;Tang et al. 2016;Liu et al. 2022). This is an issue worthy of further observations. In addition, a further quantitative comparison was conducted regarding the relationship of the center velocity and the non-thermal velocity dispersion between the HINSA features and the molecular gases. In the SCER, the differences between the center velocity of the HINSA features (see Table 1) and those of the 12 CO, 13 CO, and/or C 18 O lines are within the velocity resolution of the CO spectra (i.e., ∼ 0.2 km s −1 ). This indicates that the HINSA features are associated with the molecular clouds traced by the CO. The non-thermal velocity dispersion of the HINSA features, σ nt,HI = σ 2 HI − kT ex /m H , is smaller than that of the 12 CO lines, σ nt, 12 CO = σ 2 12 CO − kT ex /m12 CO , and is similar to that of the 13 CO lines, σ nt, 13 CO = σ 2 13 CO − kT ex /m13 CO , except for that in Beam 3, where the value of σ nt,HI is larger than that of σ nt, 12 CO by ∼ 20% (see σ HI in Table 1, and σ12 CO and σ13 CO in Table A1). This result is in agreement with that of Li & Goldsmith (2003), indicating that the HINSA in the SCER is probably mixed with the gas in cold and well-shielded regions of the molecular clouds.
In the WCER, the association between the HINSA features and the CO cloud appears to be poor.
First, the difference between the center velocity of the HINSA features (see Table 1) and that of 12 CO is larger than in the case of the SCER, and the maximum difference reaches ∼ 0.8 km s −1 (i.e., in Beam 7). Second, the value of σ nt,HI is larger than that of σ nt, 12 CO . No 13 CO emission is detected toward the regions corresponding to all the seven beams after resampling. In fact, we also do not detect any 13 CO emission in Beam 3 in the SCER, and the value of σ nt,HI in this region is larger than that of σ nt, 12 CO . Therefore, it is probably more accurate to classify Beam 3 in the SCER as a WCER component. The discussion below is based on this reclassification. We also find that the value of σ HI in the WCER (including Beam 3 in the SCER) is comparable to some of the values found by Tang et al. (2016) for CO-dark clouds. These facts indicate that the HINSA toward the WCER is probably not mixed with CO clouds, but probably associated with more diffuse CO-dark or CO-faint clouds.
Further observations of CI, CII, and/or OH may be helpful in clarifying the relationship between the HINSA features and the CO-dark or CO-faint gas (e.g., Li & Goldsmith 2003;Wolfire et al. 2010;Tang et al. 2016) toward the WCER.

Abundance of H i the HINSA Features
The fractional abundance of HINSA, X HI , based on the column density is valuable. In this case is a good measure of the column density of the molecular cloud. In Beam 6, the dilution of the dense gas tracers is relatively stronger (see Figure 1( to represent the column density of molecular clouds in the SCER. The abundance of HINSA, X HI , in most cases of the SCER (i.e., Beams 1 and 4-7), ranges from ∼ 1.1 × 10 −3 to ∼ 4.8 × 10 −3 (see the red points in Figure 4(b)). The one exception is the value of X HI is ∼ 2.6 × 10 −2 in Beam 2, where the value of N 13 CO H 2 is one order of magnitude smaller than those in the other five beams (i.e., Beams 1 and 4-7), and the relative error of N 13 CO H 2 is large (see Figure   4(b) and Table A1). The abundances of HINSA in Beams 1 and 4-7 are similar to those toward optically selected dark clouds and molecular cores (i.e., with an average value of ∼ 1.5 × 10 −3 , Li & Goldsmith 2003;Krčo & Goldsmith 2010). The HINSA abundances in the SCER (including Beam 2) fall into the HINSA abundance range toward PGCC sources (i.e., from 5.1 × 10 −4 to 1.3 × 10 −2 ) of Tang et al. (2020), but larger than the HINSA abundances derived from other PGCC sources (i.e., ∼ 3 × 10 −4 varied by a factor of ∼ 3) by Liu et al. (2022).  The HINSA abundance toward Beam 1 is consistent with the abundance of H i in the H i wind (i.e., the ratio of the column density of the H i wind to that of the total outflowing gas) in the same beam. This fact indicates that HINSA is probably a physical bridge that connects H i winds and molecular outflows.
We also find that the value of X HI in Beam 1 is smaller than the values obtained from the surrounding beams (i.e., Beams 2-7; see Figure 1(a)). This structure seems to be a ring, but unlike that observed by Zuo et al. (2018), the abundance of HINSA, X HI , diminishes toward the center of the cloud. One of the reason for this could be that the central high-mass zero-age main sequence B3 star (see Torrelles et al. 1992a) ionizes the H i detected in Beam 1 in the SCER, which is located in an H ii region (see Dewangan 2019). This then decreases the abundance of H i. The ionization can be studied using radio recombination lines (e.g., Zhang et al. 2014Zhang et al. , 2021.
The mean densities of H i,n HI , and total proton,n H , in Beam 1 are ∼ 7.2 and ∼ 6.1 × 10 3 cm −3 , respectively, assuming that the cloud diameter is ∼ 1.5 pc, corresponding to the diameter being the HPBW of FAST. We assume that the density profile has a constant density core surrounded by an envelope of the form where n 0 and r 0 (set to 10 17 cm) are the central density and the radius of the constant density core.
In this case, n 0 ∼ 9.4n H would be larger than that of model 1 in Goldsmith et al. (2007), indicating that the density is high enough to initiate the conversion of H i → H 2 . Following the model developed by Goldsmith et al. (2007), the time that has elapsed since the material was UV irradiated is ∼ 2 × 10 5 yr (see Section C), during which H i could convert to H 2 . This time is consistent with the timescale of the star formation in this region (i.e., ∼ 2 × 10 5 yr) derived from near-infrared H 2 knots by Chen et al. (2003). The model of Goldsmith et al. (2007) also implies another possible reason why X HI diminishes toward the center of the cloud, i.e., the conversion to molecular form occurs more rapidly in the higher-density inner region.
In the WCER, a comparison of N HI and N H (only for the regions where CO emission is detected) is plotted as blue and green points in Figure 4(b). The abundance of the HINSA features, X HI , is ∼ 0.04, ∼ 0.14, and ∼ 0.17 in Beams 2, 4, and 7, respectively. The value of X HI is ∼ 0.14 in Beam 3 in the SCER where 13 CO emission is not detected. The high abundances are most likely caused by an underestimation of the column density of the molecular clouds, because the associated molecular clouds (including the region in Beam 3 in the SCER) are probably CO-dark or CO-faint gas (e.g., Bolatto et al. 2013;Tang et al. 2016).
The comparison of the HINSA features and the molecular line emission indicates that the HINSA is probably mixed with the gas in cold and well-shielded CO clouds in which 13 CO emission is detected.
There are HINSA features that have no detected 13 CO and/or 12 CO emission counterparts, meaning they are probably associated with CO-dark or CO-faint gas. In addition, HINSA is a promising physical bridge to connect the H i winds with molecular outflows. There is a possible ring where the abundance of HINSA diminishes toward the center of the cloud.

APPENDIX
A. EXTRACTING HINSA The method used here to extract HINSA is based on the methodology developed by Liu et al. (2022). The first step of this method is to obtain the initial values of the optical depth, τ i , velocity dispersion, σ i , and central velocity, v i . τ i is initially set to 0.1. Obtaining the initial values of v i and σ i requires knowledge derived from the corresponding high-sensitivity CO spectrum, including the central velocity, v c,m , velocity dispersion, σ m , and excitation temperature, T ex .To extract HINSA as much as possible, we assume that every component of CO emission (i.e., Peaks 1-4) corresponds to a HINSA feature. The initial value of v i equals v c,m , and the initial value of σ i reads as where k is the Boltzmann constant, m H the mass of atomic hydrogen, and m CO the mass of the 12 CO molecule for Peaks 1-3 and of the 13 CO molecule for Peak 4.
The optical depth of a HINSA feature can be written (Krčo et al. 2008) as where τ i , σ i , and v i are the initial values of the optical depth, velocity dispersion, and central velocity of the i-th component of the HINSA features, respectively.
The H i emission free from self-absorption, T HI , and the self-absorption of H i, T ab , satisfies (Li & Goldsmith 2003) where p is a parameter representing the proportion of the background H i optical depth, T r is the observed H i spectrum, T c is the H i continuum brightness temperature, T ex is the excitation temperature (derived from the spectrum of 12 CO), and τ f is the foreground H i optical depth.
Following the treatment of Liu et al. (2022), p was set to unity, and then τ f = 0. T ex and T c were simultaneously ignored. T HI then can be expressed as which can be rewritten as where T f HI,smooth is the polynomial fit of T HI . τ i , σ i , and v i (Krčo et al. 2008) could be obtained by minimizing R where and where ∆ ch is the channel width of the H i spectrum and j the index of the channel. Similar to Liu et al. (2022), the L-BFGS-B algorithm from Scipy 2 (Virtanen et al. 2020) was used to obtain τ i , σ i , and v i . For more details of the methodology of extracting HINSA, see Liu et al. (2022 where X13 CO = [ 13 CO]/[H 2 ] = 2 × 10 −6 is used (Dickman 1978 where X12 CO = [ 12 CO]/[H 2 ] = 10 −4 is used in this work (Snell et al. 1984). If 13 CO emission is not detected, the optical depth correction in the third term is set to unity.

C. TRANSITION BETWEEN ATOMIC AND MOLECULAR PHASE
We follow the model developed by Goldsmith et al. (2007) to study the transition between H i and H 2 . We only focus on the central constant core, and assume that the central density is n 0 ∼ 5.7 × 10 4 cm −3 , line width is 1.14 km s −1 , radius is 10 17 cm, and T = 10 K. The other parameters follow these in Goldsmith et al. (2007). The equations (3)-(5) in Goldsmith et al. (2007) would produce the evolution of the density of H i, n HI , and H 2 , n H 2 . Figure 5 shows n HI and X HI as a function of time, where X HI set to n HI /(n HI + 2n H 2 ) assuming that the sizes of cold H i and CO cloud are similar (see e.g., Goldsmith et al. 2007). The time that has elapsed since the material was UV irradiated would be ∼ 2 × 10 5 yr assuming X HI ∼ 1.1 × 10 −3 .
This work made use of the data from FAST. FAST is a Chinese national mega-science facility, operated by National Astronomical Observatories, Chinese Academy of Sciences. We would like to thank all the staff members of the Delingha 13.7 m radio telescopes for their support making the various molecular line observations. We would like to thank the anonymous referee for the helpful comments and suggestions that helped to improve the paper. This work was sponsored by the n HI X HI X HI = 1.1×10 3 Figure 5. n HI and X HI as a function of time. The H i density in the cloud core drops to its steady state value of ∼ 2 cm −3 with a time of ∼ 4 × 10 5 yr.