The intrinsic abundance ratio and X-factor of CO isotopologues in L1551 shielded from FUV photodissociation
Sheng-Jun Lin, Yoshito Shimajiri, Chihomi Hara, Shih-Ping Lai, Fumitaka Nakamura, Koji Sugitani, Ryohei Kawabe, Yoshimi Kitamura, Atsushi Yoshida, Hidefumi Tatei, Toshiya Akashi, Aya E. Higuchi, Takashi Tsukagoshi
PPreprint typeset using L A TEX style emulateapj v. 5/2/11
THE INTRINSIC ABUNDANCE RATIO AND X-FACTOR OF CO ISOTOPOLOGUESIN L 1551 SHIELDED FROM FUV PHOTODISSOCIATION
Sheng-Jun Lin , Yoshito Shimajiri , Chihomi Hara , Shih-Ping Lai , Fumitaka Nakamura , Koji Sugitani ,Ryohei Kawabe , Yoshimi Kitamura , Atsushi Yoshida , Hidefumi Tatei , Toshiya Akashi , Aya E. Higuchi ,and Takashi Tsukagoshi ABSTRACTWe investigate the intrinsic abundance ratio of CO to C O and the X-factor in L 1551 usingthe Nobeyama Radio Observatory (NRO) 45 m telescope. L 1551 is chosen because it is relativelyisolated in the Taurus molecular cloud shielded from FUV photons, providing an ideal environmentfor studying the target properties. Our observations cover ∼ (cid:48) × (cid:48) with resolution ∼ (cid:48)(cid:48) , which arethe maps with highest spatial dynamical range to date. We derive the X CO /X C O value on the sub-parsec scales in the range of ∼ ± Herschel observations, we found that the abundance ratio reaches its maximumat low A V (i.e., A V ∼ X CO /X C O value at the boundary of the cloud is most likely due to the selectiveFUV photodissociation of C O. This is in contrast with Orion-A where its internal OB stars keepthe abundance ratio at a high level greater than ∼
10. In addition, we explore the variation of theX-factor, because it is an uncertain but widely used quantity in extragalactic studies. We found thatX-factor ∝ N . which is consistent with previous simulations. Excluding the high density region, theaverage X-factor is similar to the Milky Way average value. Subject headings:
ISM: abundances — ISM: clouds — photo-dominated region — ISM: individualobjects (L1551) INTRODUCTIONThe ultraviolet (UV) radiation plays a crucial role inmany processes of the interstellar medium (ISM), suchas photoelectric heating, grain charging, photoionization,and photo-dissociation of molecules (Bethell et al. 2007),in which the far-ultraviolet (FUV: 6 eV < hν < [email protected], [email protected] Institute of Astronomy and Department of Physics, NationalTsing Hua University, Hsinchu 30013, Taiwan Laboratoire AIM, CEA/DSM-CNRS-Universit´e ParisDiderot, IRFU/Service d’Astrophysique, CEA Saclay, F-91191Gif-sur-Yvette, France The University of Tokyo, 7-3-1 Hongo Bunkyo, Tokyo113-0033, Japan National Astronomical Observatory of Japan, 2-21-1 Osawa,Mitaka, Tokyo 181-8588, Japan Graduate School of Natural Sciences, Nagoya City Univer-sity, Mizuho-ku, Nagoya 467-8501, Japan SOKENDAI (The Graduate University for Advanced Stud-ies), 2-21-1 Osawa, Mitaka, Tokyo 181-8588, Japan Institute of Space and Astronautical Science, JapanAerospace Exploration Agency, 3-1-1 Yoshinodai, Chuo-ku,Sagamihara, Kanagawa 252-5210, Japan Department of Earth and Planetary Sciences, Tokyo In-stitute of Technology, 2-12-1, Okayama, Meguro-ku, Tokyo152-8551, Japan College of Science, Ibaraki University, 2-1-1 Bunkyo, Mito,Ibaraki 310-8512, Japan
Shimajiri et al. 2014, 2015). For the FUV emission withenergy high enough to photodissociate CO, it rapidlybecomes optically thick when it penetrates into molecularclouds. In contrast, the self-shielding effect of C O is rel-atively less significant because of the shift of its absorp-tion lines and its low abundance. Therefore, the FUVemission with energy above the C O dissociation levelis expected to penetrate relatively deeper in a molecu-lar cloud, and the abundance ratio of CO and C O, X CO /X C O , will increase when the self-shielding ef-fect of C O does not yet dominate. When the self-shielding effect of both CO and C O become impor-tant, the abundance ratio should decrease toward the in-trinsic abundance value which derived from abundancesof the elements, C, C, O, and C. The intrinsicvalue may vary as the distance to the Galactic center(Wilson 1999).The X CO /X C O has been observed in regions withvarious conditions. In the typical massive star-formingregions, where the ISM is filled up with diffuse FUVflux from OB stars, the measurements of X CO /X C O are usually higher than the intrinsic value. Shimajiriet al. (2014) measured X CO /X C O in the Orion-A gi-ant molecular cloud (Orion-A GMC) and found that the X CO /X C O of most regions are a factor of two greaterthan 5.5, the typical value in the solar system. It ispossible that besides the interstellar FUV radiation, em-bedded OB stars provide strong FUV radiation so thatthe distance of the penetration will increase. Anotherpossibility is that the higher temperature in the mas-sive cores will also change the fractionation of C and C (R¨ollig & Ossenkopf 2013). In the intermediate-mass star-forming regions, Kong et al. (2015) showedthat the intensity ratio of CO ( J =2–1) to C O ( J =2– a r X i v : . [ a s t r o - ph . GA ] M a y X CO /X C O for optically thincase, rises to a peak up to ∼
40 at A V ∼ ∼ A V in the southeast ofthe California molecular cloud. Although the trend isconsistent with the theoretical expectations, the peak oc-curs at somewhat higher extinction; Warin et al. (1996)showed that the peak is at A V ∼ to 10 cm − . In the low-massstar-forming regions, Lada et al. (1994) observed a partof the IC 5146 filament and found that the X CO /X C O value is considerably greater than 5.5 in the outer parts( A V (cid:46)
10 mag).LDN 1551 (hereafter L 1551) is a relatively isolatednearby star-forming region located at a distance of 160 pc(Snell 1981; Bertout et al. 1999) at the south end of theTaurus-Auriga-Perseus molecular cloud complex. Twosmall clusters of young stellar objects (YSOs) were de-tected in L 1551. One contains two embedded Class Isources, L 1551 IRS 5 (hereafter IRS 5) and L 1551 NE(hereafter NE), and the other is a group of more evolvedYSOs (hereafter HL Tau group) located at the north ofIRS 5 and NE. The IRS 5/NE cluster co-host a parsec-scale bipolar outflow (Snell et al. 1980) which is likelymixed from outflows of each source (Moriarty-Schieven etal. 2006). Several Herbig-Haro objects along the outflowshave been identified with different origins (Devine et al.1999). Another east-west redshifted outflow was foundlater (Moriarty-Schieven & Wannier 1991; Pound & Bally1991), but since IRS 5 and NE are two binary systems(Looney et al. 1997; Reipurth et al. 2002; Takakuwa et al.2014; Chou et al. 2014) the origin of this east-west out-flow is still questionable (Moriarty-Schieven et al. 2006;Stojimirovi´c et al. 2006). The HL Tau group containsHL Tau, XZ Tau, LkH α ∼ µ G (Swift et al. 2005, 2006). Since L 1551dose not contain OB stars, L 1551 is a suitable target forstudying the FUV influence only from the interstellarradiation.In this paper, we aim to study the variation of X CO /X C O under the influence of the interstellar FUVradiation. In contrast to previous studies, we observed CO ( J =1–0), CO ( J =1–0), and C O ( J =1–0) linesusing the 25-BEam Array Receiver System (BEARS) re-ceiver equipped on the Nobeyama Radio Observatory(NRO) 45 m telescope to obtain data that could resolvethe sub-parsec scale with a complete coverage from theoutskirts of the cloud into its dense region. We also use Herschel archival data to obtain a visual extinction mapup to A V ∼
70 mag in order to examine the variation of X CO /X C O with the FUV attenuation. In §
2, we de-scribe our NRO observations and Herschel data. In § CO, CO, and C O maps of L 1551,and then derive the excitation temperature and opticaldepths of the CO and C O lines and the column den-sities of these molecules. In §
4, we discuss the depen-dence of the abundance ratio of CO to C O on visualextinction, A V , the influence of FUV radiation, and theX-factor in L 1551. At the end, we summarize our results in § OBSERVATIONS AND DATA REDUCTION2.1.
NRO 45m observations
We observed the L 1551 star-forming region using the45 m telescope at NRO. Three molecular lines were ob-served: CO ( J =1–0; 115.27120 GHz), CO ( J =1–0; 110.20135 GHz), and C O ( J =1–0; 109.78218 GHz).The CO data has been published in Yoshida et al.(2010). The observations were carried out between2007 to 2010 (Table 1). The telescope has beam sizes(HPBW), θ HPBW , of 15 (cid:48)(cid:48) at 115 GHz, and of 16 (cid:48)(cid:48) at110 GHz. The main beam efficiencies, η mb , were mea-sured as 32% in 2007–2008 season at 115 GHz, 38% in2009–2010 season at 110 GHz, and 43% in 2009 sea-son at 110 GHz. These efficiencies were measured everyseason by NRO staff with a superconductor-insulator-superconductor (SIS) receiver, S100, and an acousto-optical spectrometers (AOSs). The front end is BEARS,which has 25 beams configured in a 5-by-5 array with abeam separation of 41 . (cid:48)(cid:48) ∼ − at the rest frequencies of the three lines.We calibrated the variations in both beam efficiency andsideband ratio of the 25 beams with the values from theNRO website. These values were measured by NRO staffevery season through measurements of bright sourceswith S100 in SSB mode.For the observations, we used the on-the-fly (OTF)mapping technique (Sawada et al. 2008). The antennais driven at a constant speed to continuously scan ourobserving region toward L 1551. To get rid of artificialscanning patterns in the results, we scanned the observedregions twice, one in RA and one in Dec directions, andthen combined the OTF maps in the two orthogonal di-rections by PLAIT algorithm (Emerson & Graeve 1988).The antenna pointing was checked every ∼ (cid:48)(cid:48) during the observa-tions. Figure 1 (a), (b), and (c) show the observed areasof CO, CO, and C O, covering 44 (cid:48) × (cid:48) , 42 (cid:48) × (cid:48) ,and 30 (cid:48) × (cid:48) , respectively.The data were converted in terms of the main-beambrightness temperature in units of K, T mb = T ∗ A /η mb ,where T ∗ A is the antenna temperature in units of K. Inorder to maximize the energy concentration ratio, weapplied a spheroidal function with m = 6 and α = 1(Schwab 1984) to convolve the OTF data, and then ob-tained three dimensional ( α, δ, v LSR ) final cube data.Adopting Nyquist sampling for the 45 m telescope, weset the spatial grid size to 7 . (cid:48)(cid:48)
5, and the final effectivebeam sizes, θ eff , are 21 . (cid:48)(cid:48) CO and 22 . (cid:48)(cid:48) COand C O. The rms noise 1 σ levels are 1.23 K for CO,0.94 K for CO, and 0.67 K for C O in T mb for a veloc-ity resolution of 0.1 km s − . To achieve higher signal-to-noise ratios, we further convolved all maps with Gaus-sians to make channel maps, integrated intensity maps,mean velocity maps, and velocity dispersion maps, andthe effective beam size of each map is written in its figurecaption. We summarize the parameters for each line inTable 1.2.2. Herschel column density and dust temperaturemaps
We used the Herschel Science Archival SPIRE/PACSimage data (Observation ID: 1342202250/1342202251,Quality: level 2 processed) of 160, 250, 350, and 500 µ m to make dust temperature ( T dust ) and column den-sity maps in a similar way to K¨onyves et al. (2010). Sincethe 70 µ m emission seems to be not detected except to-wards IRS 5/NE and the HL Tau group, we used only 4bands (500, 350, 250, and 160 µ m).At first, we made convolutions for all Herschel images(other than the 500 µ m image) to smooth their resolu-tions to the 500 µ m resolution of 36 (cid:48)(cid:48) by using the IDLpackage developed by Aniano et al. (2011). Then, weresampled up or down all the images (other than the250 µ m) to the same grid size of the 250 µ m image (6 (cid:48)(cid:48) )and derived a spectral energy distribution (SED) at eachposition of the 250 µ m image. Here, we adopted anarea of ∼ (cid:48) × (cid:48) , centered at (RA J2000 , Dec
J2000 ) =(4 h m . s
9, +18 ◦ (cid:48) (cid:48)(cid:48) ), as an area of the zero pointof the surface brightness for the L 1551 cloud.Assuming a single temperature of the dust emission,the gray-body SED fitting was performed with a functionof I ν = B ν ( T dust )(1 − e τ ν ) , (1)where ν infers the frequency, B ν expresses Planck’s law, I ν is the observed surface brightness, and τ ν is the dustoptical depth. τ ν can be expressed as κ ν Σ, where κ ν isthe dust opacity per unit mass and Σ is the surface massdensity. We adopted κ ν = 0 . ν β cm g − , (2)where β = 2, following K¨onyves et al. (2010).We carried out the SED fitting at each point wherethe dust emission was detected above 3 times of the rmsnoise level, which was measured in the area of the zeropoint of the surface brightness at all these 4 bands. Inthe case that the SED fitting failed or the surface bright-ness was not high enough for SED fitting at some points,we set NaN there. As the data weight of the SED fit-ting, we adopted 1 /σ , where σ is the square sum ofthe rms noise and the calibration uncertainties of sur-face brightness (15% at 500, 350, and 250 µ m from Grif-fin et al. (2010); 20% at 160 µ m from Poglitsch et al.(2010)). The dust temperature of each pixel was derivedfrom the above SED fitting, and the result is shown inFig. 1 (f). We used χ statistic to determine the good-ness of the SED fitting per pixel (Press et al. 2007). Forthe SED fitting, each χ statistic has a χ distributionwith two degrees of freedom, and then the probability Qgives a quantitative measure for the goodness. In mostof the inner region of the L 1551 cloud, the SED fittingseems good (the probability Q ∼ (cid:48)(cid:48) (Fig. 1 (d)), we derived the column density directly from the 250 µ m surface brightness ( τ µ m (cid:28) (cid:48)(cid:48) (K¨onyveset al. 2010). RESULTS3.1. CO ( J =1–0) emission line Figure 1 (a) shows the integrated intensity mapof CO ( J =1–0) integrated in the v LSR range from −
15 km s − to 20 km s − for pixels with its signal-to-noiseratio greater than 3. The CO emission is distributedall over the observed area. A sharp edge at the south-east of the mapping area can be recognized. Moriarty-Schieven et al. (2006) suggested that the stellar windsfrom massive stars, Betelgeuse and Rigel in the Ori-Eridanus supershell, possibly compressed the edge of theL 1551 molecular cloud. Two elongated structures whichare extended in the southwest–northeast direction andcentered on the Class I source IRS 5 are identified asthe molecular outflows ejected from IRS 5 and anotherembedded Class I source NE (Snell et al. 1980; Moriarty-Schieven et al. 2006; Stojimirovi´c et al. 2006).3.2. CO ( J =1–0) emission line Figure 1 (b) shows the integrated intensity mapof CO ( J =1–0) integrated in the v LSR range from −
15 km s − to 20 km s − for pixels with its signal-to-noiseratio greater than 3. The overall distribution of CO isconsistent with those in previous studies of CO carriedout with position switching mode of the FCRAO 14 m(Stojimirovi´c et al. 2006) and NRO 45 m (Yoshida et al.2010) telescopes. There are a cavity structure at thenortheast of NE and a U-shaped wall structure at thesouthwest of IRS 5. In addition, two intensity peaks canbe seen toward IRS 5 and NE. These structures are notrecognized in the CO maps.3.3. C O ( J =1–0) emission line Figure 1 (c) shows the integrated intensity mapof C O ( J =1–0) integrated in the v LSR range from −
15 km s − to 20 km s − for pixels with its signal-to-noiseratio greater than 3. The C O emission line is likely totrace the inner part of the regions traced by CO and CO. The overall distribution of C O is consistent withthose in previous studies of C O carried out with theOTF mode of the Kitt Peak 12 m telescope (Swift et al.2005) and the position switching mode of the NRO 45 m(Yoshida et al. 2010) telescope.3.4.
Column densities of the CO and C O gas andabundance ratio of CO to C O In order to derive the optical depths and column densi-ties of CO and C O, we assume that (1) CO ( J =1–0) is optically thick, (2) CO and its rarer isotopicspecies trace the same component, and (3) these threelines reach Local Thermal Equilibrium (LTE). Thus, thetemperature of CO ( J =1–0) can be treated as the ex-citation temperature, T ex , of CO ( J =1–0) and C O( J =1–0) for deriving the optical depths and column den-sities of CO and C O. In order to make direct compar-ison between the different lines, we smooth all the dataso that they have the same effective beam size of 30 . (cid:48)(cid:48) CO ( J =1–0), CO ( J =1–0), and C O( J =1–0) cube data for pixels with their signal-to-noiseratio greater than 5.In general, the spectra of CO ( J =1–0) and C O( J =1–0) show single-component velocity structures, andthus we can apply single-Gaussian fitting as an appropri-ate method. However, since the spectra of CO ( J =1–0) often show multiple velocity components due to theprominent outflows, we adopt different fitting strategiesto obtain the excitation temperature. First, if the COspectrum has only one velocity component, which usuallyhappen in quiescent ambient region, the single-Gaussianfitting may gives an adequate result. If the residual ofthe fitting at the peak velocity is less than 3 σ , we takethe peak intensity to calculate the excitation tempera-ture. Second, if the single-Gaussian fitting fails, we applydouble-Gaussian fitting to separate the ambient and out-flow components. Except the regions around NE, onlyone redshifted or blueshifted component appears. Weidentify the component of which the peak velocity is simi-lar to the peak velocities of CO and C O spectra as theambient component and the other component as the out-flow component, if (1) the residuals of the fitting at boththe peak velocities are less than 3 σ and (2) the velocityranges of the two Gaussian components at half maximumare not overlapped with each other. The later criterionis included, because the determination of the peak inten-sity of the ambient component could be affected by theclose outflow component. In some cases, the peak COintensity is even smaller than the peak CO intensity,which is not physical and is rejected. Third, for the re-maining spectra, the velocity structure seems to containthree or more components. Since our goal is to obtainthe peak intensity of the ambient gas, we only apply asingle Gaussian to one peak having the peak velocityclosest to the ambient gas velocity determined from the CO and C O spectra at the same position. We fix thepeak velocity of this single Gaussian as the mean valueof the CO and C O peak velocities and we only fitthe Gaussian within a narrow velocity range bracketedby the two local minima around the peak. Following theabove procedure, we obtain the estimate of the excita-tion temperature from the peak intensity of the fittedGaussian to the ambient CO emission.Figure 1 (e) shows the excitation temperature map de-riving from the following equation with the previous as-sumption (1) and the beam filling factor of 1 (e.g., Pinedaet al. 2010; Kong et al. 2015) T ex = hν CO k (cid:20) log (cid:18) hν CO /kT mb , CO + J ν ( T bg ) (cid:19)(cid:21) − K= 5 . (cid:20) log (cid:18) . T mb , CO + 0 . (cid:19)(cid:21) − K , (3)where T mb , CO is the peak intensity of CO ( J =1–0) inunits of K from the above fitting, J ν ( T ) = hν/k exp( hν/ ( kT )) − is the effective radiation temperature (Ulich & Haas 1976), and T bg =2.7 K is the temperature of cosmic mi-crowave background radiation. Hereafter, we call thisexcitation temperature derived from Eq. (3), T CO . Wethen obtain the optical depths, τ , of the CO ( J =1–0)and C O ( J =1–0) emission and the column densities, N , of the CO and C O gas using the following equa-tions (e.g., Lada et al. 1994; Kawamura et al. 1998) τ CO = − log (cid:18) − T mb , CO /φ CO . . /T ex ) − − − . (cid:19) , (4) τ C O = − log (cid:18) − T mb , C O /φ C O . . /T ex ) − − − . (cid:19) , (5) N CO = 2 . × τ CO FWHM CO T ex − exp( − . /T ex ) cm − , (6)and N C O = 2 . × τ C O FWHM C O T ex − exp( − . /T ex ) cm − , (7)where T mb , CO and T mb , C O are the peak intensitiesin K, FWHM CO and FWHM C O are the FWHMsin km s − , and φ CO and φ C O are the beam fillingfactors. The beam filling factors can be expressed as φ = θ / ( θ + θ ), where θ source and θ beam are the source size and the effective beam size, respec-tively. In molecular clouds, the C O emission is usuallyconsidered to trace the dense cores, clumps, and/or fil-aments. On the other hand, the CO emission tracesmore extended regions than C O. In the case of the Tau-rus molecular cloud, Tachihara et al. (2002) and Qian etal. (2012) found that both the typical CO and C Ocore size are ∼ . (cid:48)(cid:48) ∼ φ CO and φ C O as 1.Figures 2 (a) and 2 (b) show the optical depth maps of CO ( J =1–0) and C O ( J =1–0), respectively. For the CO ( J =1–0) emission, its optical depth is more than1.5 at the center of L 1551, and drops to less than 1 at theouter edge. For the C O ( J =1–0) emission, its opticaldepth is less than 0.8 in the whole region, suggestingthat the C O ( J =1–0) emission is fully optically thin inL 1551.Figures 2 (c) and 2 (d) show the column density mapsof CO and C O, respectively. The cavity structureand the U-shaped wall structure can be seen in the COcolumn density map. The C O column density mapshows that C O concentrate in the region surroundedby IRS 5, NE, and L 1551 MC. Moreover, the U-shapedwall structure is also seen in the C O column densitymap as a low column density part.We can derive the abundances from X CO = N CO /N H and X C O = N C O /N H with the H col-umn density derived from the FIR dust continuum im-ages. Because we adopt the dust temperature map withan effective beam size of 36 (cid:48)(cid:48) (see § N CO and N C O maps with Gaussians to degrade theirresolutions down to 36 (cid:48)(cid:48) (corresponding to 0.027 pc atthe distance of 160 pc). Figures 3 (a) and 3 (b) show theabundance maps of CO and C O, respectively. Themean and standard deviation of X CO and X C O are(3.1 ± × − and (3.1 ± × − , respectively. Wecan see N H peaks toward IRS 5, NE, and L 1551 MCin Fig. 1 (d) and N C O also concentrates in the aboveregions in Fig. 2 (d) but the peaks of N C O are not asobvious as N H . As a result, the X C O map shows lowvalues in above three regions, which means the C O de-pletion occurs at these regions. It is reasonable for thedepletion at L 1551 MC because we find that L 1551 MCis dense and cold in Figures 1 (d) and (f). For IRS 5 andNE, the dust temperature at the envelop of protostarsare higher than other regions, but the dust temperatureis still less than the CO evaporation temperature of 20–25 K and the H column density is very high comparedwith other regions; therefore the envelop of protostarsalso show C O depletion. On the other hand, the low X CO value also appears at the C O depletion region;however, the low X CO value derived in this region maynot show the real situation because CO and dust emis-sion may not trace the exactly same layer in the line ofsight (see § CO to C O can be directlyderived from X CO /X C O = N CO /N C O . Figure 3(d) shows the spatial variation of X CO /X C O . The X CO /X C O value ranges from ∼ ∼
27 in the wholeregion, and the mean and standard deviation of the abun-dance ratios are 8.1 ± X CO /X C O value in the outskirts of the cloud is more than ∼ CO integrated intensity. DISCUSSION4.1.
Selective photodissociation of C O To study the influence of interstellar FUV radiation on X CO /X C O , we remove the outflow regions to avoidthe influence from outflow activities for later discussion.We define the outflow regions by the regions where the CO integrated intensities in either blueshifted or red-shifted parts are higher than 3 σ (see Fig. 3 (f)); here weregard −
15 to 5.5 km s − and 7.9 to 20 km s − as theblueshifted and redshifted outflow velocity ranges, re-spectively. Moreover, to avoid the influence from under-estimation of X C O by C O depletion, we also removethe region where the X C O value is less than the ISMstandard value of 1.6 × − (Frerking et al. 1982; Ford &Shirley 2011) (shown as the gray contour in Fig. 3 (b)).In the other words, we remove the region where C Odepletion factor = 1 . × − /X C O > O depletion regions. We can see that thespatial variation of the X CO /X C O value is outside-indeceasing and the minimum of the X CO /X C O valuecoincides with the dense part in L 1551 (see the H col-umn density map in Fig. 1 (d) as a comparison). Thisspatial variation might be caused by the penetration ca-pability of interstellar FUV radiation which is related tothe visual extinction, A V . In order to investigate the de-pendence of the abundance ratio on the visual extinctionwithin the molecular cloud, we derived the A V map fromthe H column density map with a relation from Bohlin et al. (1978), A V = N H . × cm − . (8)Figure 3 (c) shows the derived A V map. Since we adoptthe dust temperature map with an effective beam sizeof 36 (cid:48)(cid:48) (see § (cid:48)(cid:48) (corresponding to 0.027 pc at the distance of 160 pc).Figure 4 (a) shows the correlation between the X CO /X C O value and the A V value. The X CO /X C O value is higher than the typical solar sys-tem value of 5.5 in the A V range less than 10 mag, andthen decreases down to ∼ A V value,suggesting the dependence of the abundance ratio on A V value is significant. Although the sample of A V (cid:46) CO ( J =1–0) and C O ( J =1–0) more than 5 σ , we still can see that the maximumof X CO /X C O value occurs between A V ∼ X CO /X C O value, we averaged X CO /X C O value in 2.5 mag A V bins. The averaged X CO /X C O value reaches the max-imum of 10.3 which is ∼ A V range less than 2.5 mag. Theoretical studies pre-dict that in the mean ISRF, the X CO /X C O maxi-mum occurs at A V ∼ ∼ A V further in-creases (Warin et al. 1996); that is, the selective pho-todissociation of C O appears at A V ∼ X CO /X C O value is only ∼ X CO /X C O value just barely reaches 27.7 which is ∼ X CO /X C O value may becaused by non-uniform density distributions. Anotherexample is shown by Lada et al. (1994) who derived asimilar correlation in a filament of IC 5146 (see Fig. 19in Lada et al. 1994), which is also a low mass star-forming region. They found that the X CO /X C O ratioat A V <
10 mag is ∼ ∼ A V >
10 mag. Note that the X CO /X C O peakpositions in the A V axes in L 1551 and IC 5146 are differ-ent. Although Lada et al. (1994) used the Near-InfraredColor Excess (NICE) technique to derive A V and we usedthe gray-body SED fitting and Equ. 8 to derive A V , boththe NICE method and our method are based on the as-sumption that R V = A V /E ( B − V ) = 3 .
1. Thus thesedifference may be owing to that the difference beam sizebetween our data or there are still some different physicalcondition between them.As seen in Fig. 1 (c), the C O emission is rather weakor not detected at the periphery of the cloud, in whichhigher X CO /X C O value is expected. On the otherhand, the CO emission is strong at these regions. Sincewe only calculated X CO /X C O value as the signal-to-noise ratio of both CO and C O greater than 5 (see § Oabundance. In order to check this possibility, we canestimate the upper limit of X CO /X C O value by usingthe C O emission with the signal-to-noise ratio of 3 andassuming FWHM C O as the value at the outskirts inFig. 1 (c) ( ∼ − ) at the region where the C Oemission is weak ( (cid:46) σ level noise) but the CO emissionis still strong ( (cid:38) σ level noise). Then X CO /X C O value can reach ∼
45 at these regions, which is ∼ Influence of the excitation temperature
In order to derive the optical depths and column den-sities, we assumed that the CO, CO, and C O linestrace the same region, and derived the excitation temper-ature, T CO , using the peak intensity of CO ( J =1–0).However, the three lines may trace different regions, and T CO could not be used as the excitation temperatures of CO ( J =1–0) and C O ( J =1–0). To investigate the in-fluence of the excitation temperature, we re-estimate theabundance ratio by using the dust temperature (shown inFig. 1 (f)), T dust , as the excitation temperature of COor/and C O. Since the thermal emission from dust isusually optically thin and can traces inner regions, the CO or/and C O may trace the same region as dust.We found that T dust is generally smaller than T CO in theinner region of L 1551 and T dust in the area we calcu-lated abundance ratio has a mean value of 13.2 ± CO emission andthe dust emission may trace different regions in the innerregion of L 1551. The H column density traced by thedust emission is also a very useful measure to examinewhether the CO lines trace the same region as the dustor not. In fact, we can see the distributions of the N H map and the N C O map are similar, compared to the N CO map. Thus, we assume the excitation tempera-ture of C O is the dust temperature, but we try both T CO and T dust for CO for comparison. On the otherhand, since Stojimirovi´c et al. (2006) used the two rota-tional transitions of CO ( J =1–0 and 3–2) to derive amean excitation temperature of the main NE/IRS 5 out-flow to be 16.5 K, which is generally higher than T CO , wealso use this temperature for comparison. Therefore, wecompare the derived abundance ratios under the follow-ing four assumptions:1. T ex , CO = T CO and T ex , C O = T CO . (Original)2. T ex , CO = T CO and T ex , C O = T dust .3. T ex , CO = T dust and T ex , C O = T dust .4. T ex , CO = T ex , C O = 16 . X CO /X C O and A V under one of the above four as-sumptions. The mean and standard deviation of theabundance ratios are 8.0 ± ± ± ± X CO /X C O value on the A V value under the assumption (2) and (4) are consistentwith that under the original assumption (1). We noticethat the X CO /X C O value at A V (cid:38) T CO usedin Fig. 4 (a) has a mean value of 15.3 ± X CO /X C O and A V . Thus, thisderivation of abundance ratio is not sensitive to the ex-citation temperature. In the case of the assumption (3),the trend of X CO /X C O is similar to those under theother assumptions in the A V range less than 10 mag, butthe X CO /X C O value increases in the A V range morethan 10 mag. In fact, the ambient region in which the A V value larger than 10 mag is corresponding to L 1551 MCin which the dust temperature is less than ∼
10 K. Sincethe only difference between the assumption (2) and (3)is the excitation temperature of CO, this behavior canbe interpreted as follows: the excitation temperature of CO in L 1551 MC is underestimated in the case of T ex = T dust , and thus the column density of CO is over-estimated. This suggests that the CO line does nottrace the dense region of L 1551 MC traced by the dustemission: L 1551 MC cannot be recognized in the COintegrated intensity map of Fig. 1 (b).Consequently, even when we take into account the un-certainties in the excitation temperature estimation, ourresults still lead to the same conclusion. That is, the X CO /X C O value reaches ∼
11 at the low A V value( A V ∼ A V value ( A V ∼ ∼
10 K) excitation temperature for the COline.4.3.
Comparison between the L 1551 molecular cloudand the Orion-A giant molecular cloud
Orion-A GMC is located at a distance of 400 pc(Menten et al. 2007; Sandstrom et al. 2007; Hirota etal. 2008) and is under strong influence of the FUV emis-sion from the Trapezium cluster and NU Ori (Shimajiriet al. 2011). Several photon dissociation regions (PDRs)have been identified from comparison of the distributionsof the 8 µ m, 1.1 mm, and CO ( J =1–0) line emission(Hollenbach & Tielens 1997; Shimajiri et al. 2011, 2013).Shimajiri et al. (2014) found that the abundance ratiosof CO to C O in PDRs and non-PDRs are both higherthan 5.5, and concluded that these high abundance ratiosare due to the selective FUV photodissociation of C O.In order to investigate the influence of the different en-vironments on the abundance ratio of CO to C O, wecompare the abundance ratio between L 1551 and Orion-A. Figure 5 shows the X CO /X C O value as a functionof N C O for L 1551 and Orion-A (both the PDR andnon-PDR regions). The data of Orion-A are from Shi-majiri et al. (2014). Here, Shimajiri et al. (2014) usedthe C O column density to estimate the total columndensity. Although the C O column density and the A V value do not have a perfect linear relation because of theselective photodissociation (Frerking et al. 1982), we canstill consider that the higher column density is generallycorresponding to the higher A V value and vice versa. Ourresults show that the abundance ratio in L 1551 is gener-ally lower than that in Orion-A. The mean X CO /X C O value in L 1551 is 8.0 ± ± ± X CO /X C O value has a maximum in the low N C O regime ( < × cm − ), and then decreases to ∼
10 inthe high N C O regime (see Fig. 7 (d), (e), (f), and (g) inShimajiri et al. 2014). This trend is similar to the resultof L 1551 but the selective photodissociation of C O inL 1551 only occurs at the outskirts ( A V ∼ ∼ CO-to-H conversion factor across the L 1551molecular cloud Measurements of the mass distribution in molecularclouds help to understand their physical and chemicalcharacteristics. However, in ISM, the most abundantmolecular species, H , is hard to be observed because H lacks its electric-dipole moment and its quadruple tran-sition is hard to occur in the typical molecular cloud en-vironment. The secondary abundant molecular species,CO, is not the case, because CO is much easier to beobserved and the CO ( J =1–0) emission is consideredas the most available mass tracer. Thus, the CO-to-H conversion factor, also called “X-factor”, is defined as,X–factor = N H W CO cm − K − km − s , (9)where W CO is the integrated intensity of the CO( J =1–0) emission. Bolatto et al. (2013) showed thatthe averaged X-factor = 2 × cm − K − km − s with ±
30% uncertainty in the Milky Way disk. Nevertheless,the X-factor can vary by a factor of ∼
100 in differentregions, because CO ( J =1–0) is often optically thick(Lee et al. 2014). For example, Pineda et al. (2010) mea-sured X-factor ∼ (1.6–12) × cm − K − km − s in theTaurus molecular cloud, Lee et al. (2014) measured X-factor ∼ × cm − K − km − s in the Perseus molec-ular cloud, and Kong et al. (2015) measured X-factor ∼ × cm − K − km − s in the southeastern part ofthe California molecular cloud. These discrepancies arealso found in ISM numerical simulations which show thatX-factor is likely dependent on extinction, volume den-sity, temperature, metallicity, turbulence, star formationfeedback, and so on (Shetty et al. 2011a,b; Lee et al.2014; Clark & Glover 2015). Even within a molecularcloud, the X-factor can still vary with a wide range. Onthe other hand, the variation of the X-factor of the COand C O can be smaller (see Appendix D).Hereafter we calculate the X-factor for pixels in Fig. 1(a) with the signal-to-noise ratio of the CO integratedintensity greater than 3. Figure 6 shows the correla-tion between the W CO value and the A V value, andthe points are color-coded by T ex = T CO . We can seethat there are multiple trends. The averaged X-factorof the whole region is 1.08 +1 . − . × cm − K − km − s.This is about a factor of two smaller than the averageX-factor in Milky Way, but the dispersion still coversthe average X-factor in Milky Way. In order to revealthese different trends, although we have already definedtwo regions, the ambient and outflow components, in the previous discussion, we divide into more regions basedon the previous division here: (a) the diffuse componentwhich is the ambient component excluding the L 1551MC component, (b) the L 1551 MC component which isthe green polygon in Fig. 3 (f), and (c) the outflow com-ponent which is the same as defined in Sec. 4.1. In eachpanel, we use the chi-squre fitting method with errors inboth coordinates to find the best fit line.The diffuse component (Fig. 6 (a)) shows a rel-atively narrow distribution, compared with theother two regions, and the fitted X-factor =1.24 +0 . − . × cm − K − km − s is similar to thefitted X-factor of the whole region and is the sameorder of magnitude as the average X-factor in MilkyWay. The L 1551 MC component (Fig. 6 (b)), however,shows a more extended distribution and a larger fittedX-factor = 5.59 +3 . − . × cm − K − km − s. This isdue to the fact that the CO emission is optically thickand thus saturated in this dense starless core. Theoutflow component (Fig. 6 (c)) has a fitted X-factor =7.43 +1 . − . × cm − K − km − s, which is smaller thanthat of the diffuse component, but shows three differentdistributions. Two of the distributions are extended andhave shallow slopes in the correlation, which belong tothe outflow region overlapped with IRS 5 and NE shownin Fig. 3 (f) as two green incomplete circles and markedin the panel (c) of Fig. 6. The CO emission from theIRS 5 and NE region is likely saturated because theH column densities of IRS 5 and NE are comparableto that of L 1551 MC ( N H ∼ cm − ). The othernarrow distribution belongs to the remaining outflowsand dominates the fitting of X-factor, which may bedue to the fact that the entrained energy more stronglyexcites the CO emission and makes its high excitationtemperature. Consequently, in the low A V range( A V (cid:46)
10 mag), the X-factor of the diffuse component isconsistent with the Milky Way average value, and theX-factor of the outflow component is smaller. However,in L 1551 MC ( A V (cid:38) A V value, and the points are color-coded by T ex . We can see that the X-factor of the whole regionis dependent on A V with a fitted power law, X-factor ∝ N . ± . . Especially, the L 1551 MC component has awell-correlated distribution (less dispersion of the powerindex), X-factor ∝ N . ± . . A similar distributionbut a shallower power law, X-factor ∝ N . , was alsofound by Kong et al. (2015) in the southeastern part ofthe California molecular cloud excluding the hot regionaround the massive star LkH α A V (cid:46) A V (cid:38) to 10 cm − and initial turbulence, and showed thatthe low density case has the decreasing trend with a wider A V range compared to the high density case. For thehighest density case in their simulation (n ∼ cm − ),the result only shows one characteristic trend, X-factor ∝ N H , which means the CO emission is saturated(optically thick) to be constant and then the X-factordirectly relates to N H , which is consistent with our re-sults. In L 1551, we do not find the decreasing trend evenat the low A V range, which indicates the FUV strengthin L 1551 is less than that in Perseus. Under the meanISRF, CO is self-shielded at A V ∼ T ex value, and the points are color-coded by A V . Our data do not show obvious correlated distri-butions. However, an inverse power law, X-factor =2 × ( T ex / − . , was found in the southeastern partof California cloud (Kong et al. 2015), which is consis-tent with X-factor ∝ T − . in simulations of Shetty etal. (2011b). Although this relation passes through thediffuse component of L 1551, we can not confirm this re-lationship due to the narrow temperature range of ourdata.Figure 9 shows the correlation between the X-factorand the CO velocity dispersion, σ CO , and the pointsare color-coded by T ex . The simulations predict thatX-factor ∝ σ − . CO (Shetty et al. 2011b). However, ourdata do not have a correlated distribution. Kong et al.(2015) also found no correlation in the southeastern partof California cloud. SUMMARYWe have carried out wide-field OTF observations inthe CO, CO, and C O ( J =1–0) lines toward theL 1551 molecular cloud using the BEARS receiver of theNRO 45 m telescope. The main results are summarizedas follows:1. The abundance ratio of CO to C O( X CO /X C O = N CO /N C O ) is derived tobe ∼ ± X CO /X C O map from the outskirts to the dense part of L 1551,the starless core L 1551 MC.2. We found that the X CO /X C O value reaches itsmaximum in A V ∼ A V value, suggesting that theselective photodissociation of C O from the in-terstellar FUV radiation occurs in L 1551. PDRtheoretical models with the mean ISFR suggestthat the selective photodissociation of C O causesthe increase of X CO /X C O value up to ∼ A V regionswith ∼ X CO /X C O value in 2.5 mag bins reaches its maximum value of 10.3 which is ∼ A V range less than 2.5 mag. Since the modelcalculations consider non-turbulent uniform clouddensity, this lower maximum X CO /X C O valuemay be caused by non-uniform cloud density.3. The mean of the excitation temperatures of COand C O derived from CO in the non-outflow re-gions is 15.3 ± T ex , CO = T ex , C O =16.5 K or assume T ex , CO = T dust , we also find asimilar trends of X CO /X C O . Therefore, our re-sults described in item 2 are not sensitive to theuncertainties in the excitation temperature deter-mination.4. The trends in the X CO /X C O variations inL 1551 and Orion-A are similar to each other. Thatis, both X CO /X C O values reach their maxi-mums at low N C O value, and then decrease as N C O value increase; however, the X CO /X C O value in Orion-A converges to a higher value thanthat in L 1551. This is due to the FUV radiationfrom the embedded OB stars.5. We calculated that the averaged X-factor = N H /W CO( J =1 − in L 1551 is1.08 +1 . − . × cm − K − km − s which is some-what smaller but consistent with the Milky Wayaverage value ∼ × cm − K − km − s.6. The X-factor of the different regions in L 1551 hasa large variation. For the diffuse region, we found aX-factor = 1.24 +0 . − . × cm − K − km − s whichis similar to the Milky Way average value.For the outflow region, we found a smaller X-factor = 7.43 +1 . − . × cm − K − km − s. Theregion of L 1551 MC has a larger X-factor =5.59 +3 . − . × cm − K − km − s with a large dis-persion because of the saturation of the CO emis-sion. For the whole region, we found a correlationbetween the X-factor and the A V value with X-factor ∝ N . ± . at A V = ∼ CO emission.The 45 m radio telescope is operated by Nobeyama Ra-dio Observatory, a branch of National Astronomical Ob-servatory of Japan. Part of this work was achieved usingthe grant of NAOJ Visiting Fellow Program supported bythe Research Coordination Committee, National Astro-nomical Observatory of Japan (NAOJ). S.J.L. and S.P.L.acknowledge support from the Ministry of Science andTechnology of Taiwan with Grants MOST 102-2119-M-007-004-MY3.APPENDIX
A. VELOCITY STRUCTURE OF THE CO ( J =1–0) EMISSION LINE Figure A1 shows the velocity channel maps of CO ( J =1–0). In the velocity range from − − to 6 km s − ,an extended elongated emission can be recognized at the southwest of IRS 5. This component is identified as theblueshifted lobe of the outflows ejected from IRS 5 and NE (Emerson et al. 1984; Moriarty-Schieven et al. 2006;Stojimirovi´c et al. 2006). In the velocity range from 4 km s − to 10 km s − , a narrower elongated structure appears atthe northeast of IRS 5 and through NE, which could still be a mix of the outflows from both sources. In the velocityrange from 2 km s − to 6 km s − , a clam-shaped emission shows up around the HL Tau group, which are identified asthe collective outflows from the HL Tau group (Mundt et al. 1990; Stojimirovi´c et al. 2006; Yoshida et al. 2010). Thisoutflow structure extends to ∼
14 km s − in the redshifted lobe. In the velocity range from 10 km s − to 20 km s − ,there is an emission between NE and the HL Tau group. This component is located between the redshifted outflowsfrom IRS 5 and NE and the collective outflows from the HL Tau group (Stojimirovi´c et al. 2006), which could be theresults of the outflow interaction. In the velocity range from 8 km s − to 16 km s − , another narrow elongated emissionis distributed in the east–west direction pointing back to IRS 5 and NE, but the origin of this east–west outflow isstill unclear (Moriarty-Schieven & Wannier 1991; Pound & Bally 1991; Reipurth et al. 2002; Stojimirovi´c et al. 2006).At 10 km s − , a diffused emission with an intensity of ∼ − , two individual small coreswith an extent of ∼ (cid:48) at the northern side probably do not associate with the L 1551 molecular cloud because theirvelocity of ∼ − is different from the systemic velocity of ∼ − in L 1551 (see the 6 km s − panel in Fig. A1).Figures A2 (a) and (d) show CO mean velocity and velocity dispersion maps, respectively. The mean velocityover whole region is 6.6 ± − (Yoshida et al. 2010). We calculate that the mean velocity dispersion toward allobserved area is 0.44 ± − . Note that the velocity dispersion at the area of the outflows ( ∼ − ) isrelatively higher than the mean value toward the overall observed area. B. VELOCITY STRUCTURE OF THE CO ( J =1–0) EMISSION LINE Figure B1 shows the velocity channel maps of CO ( J =1–0). In the velocity range from 4.25 km s − to 7.25 km s − ,we can see the U-shaped wall clearly at the southwest of IRS 5 which is distributed around the southwestern blueshiftedlobe of the outflows traced in CO and extending to the outside of those outflows. Note that although the U-shapedwall seems to surround the blueshifted outflows, its velocity covers both redshifted and blueshifted ranges, which mayhint that the blueshifted outflow axis is almost parallel to the plane of sky. In the velocity range from 6.25 km s − to 7.25 km s − , we can see the cavity at the northeast of NE. At the velocity of 6.25 km s − , we can see a narrowfilamentary structure sticking out from IRS 5 and extending toward L 1551 MC and beyond. At the velocity of6.75 km s − , the CO emission is distributed over the whole region, which is consistent to the centroid velocity of6.7 ± − measured by Yoshida et al. (2010). At the velocity of 7.75 km s − , we can barely see an east–westelongated structure corresponding to the east–west outflow seen in the CO map.Figure A2 (b) shows the CO mean velocity map. The cavity is clearly seen in the blueshifted velocities withrespect to the mean velocity of 6.7 km s − . Because the cavity is on top of the redshifted outflows from IRS 5/NE,we speculate that the gas on the cavity region is pushed toward us by the redshifted outflows. There is a redshiftedstreamer ( v LSR (cid:38) . − , see the black dashed line in Fig. A2 (b)) at the southwest of the narrow filamentarystructure (see the grey dashed line in Fig. A2 (b)) and at the north of east–west outflow. However, the integratedintensity of this streamer is very low in Fig. 1 (b), an thus this component is not discussed hereafter. Figure A2 (e)shows the CO velocity dispersion map. The velocity dispersion of the U-shaped wall ( ∼ − ) is relativelyhigher than the mean value in the overall observed area (0.20 ± − ). C. VELOCITY STRUCTURE OF THE C O ( J =1–0) EMISSION LINE Figure C1 shows the velocity channel maps of C O ( J =1–0). These maps show a filamentary structure in thenorthwest–southeast direction in the velocity range from 6.25 km s − to 7.25 km, and L 1551 MC traced by NH coincides with this filament (Swift et al. 2005). This filamentary structure can also be recognized in the CO channelmaps (the 6.25 km s − panel of Fig. B1) and the H column density map (Fig. 1 (d)). A U-shaped wall structure ispresent in the velocity range from 6.75 km s − to 7.05 km, which coincides with the redshifted part of the U-shaped walltraced in CO. Another fainter filamentary structure can be seen in the velocity range from 7.05 km s − to 7.35 km s − ,and its position is just above the upper arm of the U-shaped wall. Figure A2 (c) and (f) are the C O mean velocityand the velocity dispersion maps, respectively. Within the U-shaped wall, the integrated intensity becomes low andthe gas is mostly blueshifted (see Fig. 1 (c)).
D. RELATIONSHIP BETWEEN CO ISOTOPES AND H COLUMN DENSITY
Figures D1 and D2, respectively, show the correlation between the W CO value, the integrated intensity of the CO( J =1–0) emission, and the A V value, and the W C O value, the integrated intensity of the C O ( J =1–0) emission,and the A V value, in which the points are color-coded by T ex = T CO . In order to make direct comparison with theircolumn density, we only plot the same regions where the pixels have their signal-to-noise ratio greater than 5 (see § CO and C O emission are 7.48 +5 . − . × and 7.23 +3 . − . × cm − K − km − s,respectively. In Fig. D1 and D2, the distribution of points in each panel is similar to that for CO but has a narrowerdistribution because their optical depths are smaller than that of the CO emission. Especially for the C O emission,0since τ C O < .
8, even the distribution of L 1551 MC has a linear correlation, compared with the CO and CO.
REFERENCESAlonso-Albi, T., Fuente, A., Crimier, N., et al. 2010, A&A, 518,A52Aniano, G., Draine, B. T., Gordon, K. D., & Sandstrom, K. 2011,PASP, 123, 1218Bergin, E. A., & Tafalla, M. 2007, ARA&A, 45, 339Bertout, C., Robichon, N., & Arenou, F. 1999, A&A, 352, 574Bethell, T. J., Zweibel, E. G., & Li, P. S. 2007, ApJ, 667, 275Bohlin, R. C., Savage, B. D., & Drake, J. F. 1978, ApJ, 224, 132Bolatto, A. D., Wolfire, M., & Leroy, A. K. 2013, ARA&A, 51,207Burrows, C. J., Stapelfeldt, K. R., Watson, A. M., et al. 1996,ApJ, 473, 437Caselli, P., Walmsley, C. M., Tafalla, M., Dore, L., & Myers, P. C.1999, ApJ, 523, L165Devine, D., Reipurth, B., & Bally, J. 1999, AJ, 118, 972Dickman, R. L. 1978, ApJS, 37, 407Chou, T.-L., Takakuwa, S., Yen, H.-W., Ohashi, N., & Ho,P. T. P. 2014, ApJ, 796, 70Clark, P. C., & Glover, S. C. O. 2015, MNRAS, 452, 2057Emerson, J. P., Harris, S., Jennings, R. E., et al. 1984, ApJ, 278,L49Emerson, D. T., & Graeve, R. 1988, A&A, 190, 353Ford, A. B., & Shirley, Y. L. 2011, ApJ, 728, 144Frerking, M. A., Langer, W. D., & Wilson, R. W. 1982, ApJ, 262,590Griffin, M. J., Abergel, A., Abreu, A., et al. 2010, A&A, 518, LL3Gutermuth, R. A., Megeath, S. T., Myers, P. C., et al. 2009,ApJS, 184, 18Herbig, G. H., Andrews, S. M., & Dahm, S. E. 2004, AJ, 128,1233Hirota, T., Ando, K., Bushimata, T., et al. 2008, PASJ, 60, 961Hollenbach, D. J., & Tielens, A. G. G. M. 1997, ARA&A, 35, 179Kawamura, A., Onishi, T., Yonekura, Y., et al. 1998, ApJS, 117,387K¨onyves, V., Andr´e, P., Men’shchikov, A., et al. 2010, A&A, 518,LL106Kong, S., Lada, C. J., Lada, E. A., et al. 2015, ApJ, 805, 58Lada, C. J., Lada, E. A., Clemens, D. P., & Bally, J. 1994, ApJ,429, 694Lada, C. J., Lombardi, M., & Alves, J. F. 2009, ApJ, 703, 52Lee, M.-Y., Stanimirovi´c, S., Wolfire, M. G., et al. 2014, ApJ,784, 80Liszt, H. S. 2007, A&A, 476, 291Looney, L. W., Mundy, L. G., & Welch, W. J. 1997, ApJ, 484,L157Menten, K. M., Reid, M. J., Forbrich, J., & Brunthaler, A. 2007,A&A, 474, 515Moriarty-Schieven, G. H., & Wannier, P. G. 1991, ApJ, 373, L23Moriarty-Schieven, G. H., Johnstone, D., Bally, J., & Jenness, T.2006, ApJ, 645, 357Mundt, R., Buehrke, T., Solf, J., Ray, T. P., & Raga, A. C. 1990,A&A, 232, 37Pineda, J. L., Goldsmith, P. F., Chapman, N., et al. 2010, ApJ,721, 686Poglitsch, A., Waelkens, C., Geis, N., et al. 2010, A&A, 518, LL2 Pound, M. W., & Bally, J. 1991, ApJ, 383, 705Press, W. H., Teukolsky, S. A., Vetterling, W. T., & Flannery,B. P. 2007, Numerical Recipes (Cambridge: Cambridge Univ.Press), 778Qian, L., Li, D., & Goldsmith, P. F. 2012, ApJ, 760, 147Reipurth, B., Rodr´ıguez, L. F., Anglada, G., & Bally, J. 2002, AJ,124, 1045R¨ollig, M., & Ossenkopf, V. 2013, A&A, 550, AA56Sandstrom, K. M., Peek, J. E. G., Bower, G. C., Bolatto, A. D.,& Plambeck, R. L. 2007, ApJ, 667, 1161Sawada, T., Ikeda, N., Sunada, K., et al. 2008, PASJ, 60, 445Schwab, F. R. 1984, in Optimal Gridding of Visibility Data inRadio Interfer- ometry, Indirect Imaging, ed. J. A. Robert(Cambridge: Cambridge Univ. Press), 333Shetty, R., Glover, S. C., Dullemond, C. P., & Klessen, R. S.2011, MNRAS, 412, 1686Shetty, R., Glover, S. C., Dullemond, C. P., et al. 2011, MNRAS,415, 3253Shimajiri, Y., Kawabe, R., Takakuwa, S., et al. 2011, PASJ, 63,105Shimajiri, Y., Sakai, T., Tsukagoshi, T., et al. 2013, ApJ, 774,LL20Shimajiri, Y., Kitamura, Y., Saito, M., et al. 2014, A&A, 564,AA68Shimajiri, Y., Kitamura, Y., Nakamura, F., et al. 2015, ApJS,217, 7Snell, R. L., Loren, R. B., & Plambeck, R. L. 1980, ApJ, 239, L17Snell, R. L. 1981, ApJS, 45, 121Sorai, K., Sunada, K., Okumura, S. K., et al. 2000, in Society ofPhoto-Optical Instrumentation Engineers (SPIE) ConferenceSeries, Vol. 4015, Radio Telescopes, ed. H. R. Butcher, 86-95Stojimirovi´c, I., Narayanan, G., Snell, R. L., & Bally, J. 2006,ApJ, 649, 280Swift, J. J., Welch, W. J., & Di Francesco, J. 2005, ApJ, 620, 823Swift, J. J., Welch, W. J., Di Francesco, J., & Stojimirovi´c, I.2006, ApJ, 637, 392Swift, J. J., & Welch, W. J. 2008, ApJS, 174, 202Sunada, K., Yamaguchi, C., Nakai, N., et al. 2000, in Society ofPhoto-Optical Instrumentation Engineers (SPIE) ConferenceSeries, Vol. 4015, Radio Telescopes, ed. H. R. Butcher, 237-246Sz˝ucs, L., Glover, S. C. O., & Klessen, R. S. 2014, MNRAS, 445,4055Tachihara, K., Onishi, T., Mizuno, A., & Fukui, Y. 2002, A&A,385, 909Takakuwa, S., Saito, M., Saigo, K., et al. 2014, ApJ, 796, 1Ulich, B. L., & Haas, R. W. 1976, ApJS, 30, 247van Dishoeck, E. F., & Black, J. H. 1988, ApJ, 334, 771Visser, R., van Dishoeck, E. F., & Black, J. H. 2009, A&A, 503,323Warin, S., Benayoun, J. J., & Viala, Y. P. 1996, A&A, 308, 535Wilson, T. L. 1999, Reports on Progress in Physics, 62, 143Yoshida, A., Kitamura, Y., Shimajiri, Y., & Kawabe, R. 2010,ApJ, 718, 1019 K km s -1 K km s -1 K km s -1 cm -2 K CO integrated intensity CO integrated intensity C O integrated intensityN(H ) T dust L 1551 MCNEIRS 5 V1213 TauLkHα 358HL TauXZ Tau WallCavity Wall T ex Fig. 1.—
L 1551 maps of integrated intensity and derived proper-ties. The integrated intensity maps of (a) CO ( J =1–0), (b) CO( J =1–0), and (c) C O ( J =1–0) in v LSR =[ −
15 km s − , 20 km s − ]in units of K km s − with effective beam sizes of 24 . (cid:48)(cid:48)
9, 25 . (cid:48)(cid:48)
2, and25 . (cid:48)(cid:48)
2, respectively. The crosses denote the Class I sources NE, andIRS 5, and the Class I/II source HL Tau. The circles denote XZTau, LkH α
358 and V1213 Tau. The black ellipse denotes the star-less core L 1551 MC. The blue and white ellipses denote a U-shapedwall structure and the magenta one a cavity structure. The linearscale of 0.1 pc is shown at the bottom-left corner of each panel. (d)The column density map of H in units of cm − with an effectivebeam size of 18 (cid:48)(cid:48) , derived from Herschel data. (e) The excitationtemperature map in units of K with an effective beam size of 30 . (cid:48)(cid:48) CO ( J =1–0) peak intensities. (f) The dusttemperature map in units of K with an effective beam size of 36 (cid:48)(cid:48) .Each beam size is denoted as the white ellipses at the bottom-rightcorner of each panel. cm -2 cm -2 τ( CO) τ(C O)N(C O)CO) N( Fig. 2.—
L 1551 maps of optical depth and column density. The optical depth maps of (a) CO ( J =1–0) and (b) C O ( J =1–0) andthe column density maps of (c) CO and (d) C O in units of cm − . The effective beam sizes are 30 . (cid:48)(cid:48)
4. Note that these maps are madefrom the data with the signal-to-noise ratio larger than 5. The symbols in the panels have the same meaning as in Fig. 1. cm -2 A V X( CO) X(C O) (cid:0) X( CO) X(C O) (cid:0) without outflow & C O depletion regions X( CO) X(C O) Fig. 3.—
Comparison between the abundances, X CO and X C O , the abundance ratio, X CO /X C O , and the visual extinction, A V . (a) The X CO map and (b) the X C O map with effective beam sizes of 36 (cid:48)(cid:48) . The gray contour in the panel (b) indicates the ISMstandard value, X C O = 1.6 × − . (c) The A V map in units of mag, derived from Herschel data. (d) The X CO /X C O map and(e) the X CO /X C O map without outflow and C O depletion regions. The effective beam sizes are 36 (cid:48)(cid:48) . The white, magenta, andblack contours in the panels (d) and (e) indicate 5.5, 8.25, 11 mag, respectively. (f) The CO ( J =1–0) integrated intensity contours withan effective beam size of 36 (cid:48)(cid:48) superposed on the column density map of H with an effective beam size of 18 (cid:48)(cid:48) . The contours indicatethe blueshifted outflows (blue, v LSR = [ −
15 km s − , 5.5 km s − ]), the redshifted outflows (red, v LSR = [7.9 km s − , 20 km s − ]), and theambient component (black, v LSR = [5.5 km s − , 7.9 km s − ]). The contour levels start from 3 σ with a step of 6 σ where the rms noise 1 σ levels are 1.44 K km s − (red), 2.68 K km s − (black), 1.23 K km s − (blue). The green polygon is a selected region of L 1551 MC and thetwo incomplete green circles are selected regions of IRS 5 and NE, respectively. The symbols in the panels have the same meaning as inFig. 1. X C O / X C O (a) T ex , CO = T CO T ex , C O = T CO (b) T ex , CO = T CO T ex , C O = T dust Undepleted ambient component .Binned above componentRest component A V [mag] X C O / X C O (c) T ex , CO = T dust T ex , C O = T dust A V [mag] (d) T ex , CO = 16.5K T ex , C O = 16.5K N H [10 cm − ] N H [10 cm − ] Fig. 4.—
Correlation between the visual extinction, A V , and the abundance ratio, X CO /X C O . The top x-axis is the column densityof H , N H . The four panels (a), (b), (c), and (d) correspond to the four cases of the excitation temperatures of the CO and C O( J =1–0) lines. The blue and green dots denote the undepleted ambient and rest (depletion factor > X CO /X C O in the 2.5 mag A V bin. The yellow contours indicate the surface density of ambient points at the 15%, 45%, and 75% levels of the maximum surface densityof the undepleted ambient points. The horizontal dashed lines indicate the solar system value of 5.5. . . . . . . . . . N C O [10 cm − ] X C O / X C O Undepleted ambient component in L1551Rest component in L1551Orion A (Shimajiri et al. 2014)
Fig. 5.—
Correlation between the column density of C O, N C O , and the abundance ratio, X CO /X C O in L 1551 and Orion A. Theblue and green dots denote the undepleted ambient and rest components in L 1551, respectively. The gray dots denote the Orion A datafrom Shimajiri et al. (2014). The yellow and black contours indicate the surface density of the blue and grey points at the 15%, 45%, and75% levels of their maximum surface density of the ambient component in L 1551 and Orion A, respectively. The horizontal dashed lineindicates the solar system value of 5.5. Fig. 6.—
Correlation between the visual extinction, A V , and the integrated intensity of CO ( J =1–0), W CO . The gray dots denotesall data points in the map. The color-coded dots show the data points in the labeled regions, and the color represents the excitationtemperature, T ex . The black contours indicate the surface density of the color-coded points at the 15%, 45%, and 75% levels of themaximum surface density in each panel. Fig. 7.—
Correlation between the visual extinction, A V , and the X-factor. The gray dots denotes all data points in the map. Thecolor-coded dots show the data points in the labeled regions, and the color represents the excitation temperature, T ex . The blue contoursindicate the surface density of the color-coded points at the 5%, 15%, 45%, and 75% levels of the maximum surface density of all data. Fig. 8.—
Correlation between the excitation temperature, T ex , and the X-factor. The gray dots denotes all data points in the map. Thecolor-coded dots show the data points in the labeled regions, and the color represents the visual extinction, A V . The blue contours indicatethe surface density of the color-coded points at the 5%, 15%, 45%, and 75% levels of the maximum surface density of all data. Fig. 9.—
Correlation between the velocity dispersion of CO ( J =1–0), σ CO , and the X-factor. The gray dots denotes all data pointsin the map. The color-coded dots show the data points in the labeled regions, and the color represents the excitation temperature, T ex .The blue contours indicate the surface density of the color-coded points at the 5%, 15%, 45%, and 75% levels of the maximum surfacedensity of all data. Fig. A1.—
L 1551 velocity channel maps of CO ( J =1–0) in units of K. The numbers at the top-left corners denote the central LSRvelocities of the individual channels in units of km s − . The effective beam sizes are 30 . (cid:48)(cid:48)
0. The linear scale of 0.25 pc is shown at thebottom-left corner of the first panel. km s -1 km s -1 km s -1 CO mean velocity km s -1 km s -1 km s -1 CO velocity dispersion CO mean velocity CO velocity dispersion C O mean velocity C O velocity dispersion
Fig. A2.—
L 1551 maps of mean velocity and velocity dispersion. The mean velocity maps of (a) CO ( J =1–0), (b) CO ( J =1–0),and (c) C O ( J =1–0) in units of km s − with effective beam sizes of 28 . (cid:48)(cid:48)
3, 28 . (cid:48)(cid:48)
6, and 28 . (cid:48)(cid:48)
6, respectively. The gray and black dashedlines indicate in (b) the narrow filamentary structure and the redshifted streamer, respectively. The velocity dispersion maps of (d) CO( J =1–0), (e) CO ( J =1–0), and (f) C O ( J =1–0) in units of km s − with effective beam sizes of 28 . (cid:48)(cid:48)
3, 28 . (cid:48)(cid:48)
6, and 28 . (cid:48)(cid:48)
6, respectively.The mean velocities and velocity dispersions of CO, CO, and C O maps are calculated in v LSR = [ −
15 km s − , 20 km s − ], [3 km s − ,8 km s − ], and [5.7 km s − , 7.5 km s − ], respectively. The rest symbols in the panels have the same meaning as in Fig. 1. Fig. B1.—
L 1551 velocity channel maps of CO ( J =1–0) in units of K. The numbers at the top-left corners denote the central LSRvelocities of the individual channels in units of km s − . The effective beam sizes are 30 . (cid:48)(cid:48)
3. The linear scale of 0.25 pc is shown at thebottom-left corner of the first panel. Fig. C1.—
L 1551 velocity channel maps of C O ( J =1–0) in units of K. The numbers at the top-left corners denote the central LSRvelocities of the individual channels in units of km s − . The effective beam sizes are 30 . (cid:48)(cid:48)
4. The linear scale of 0.25 pc is shown at thebottom-left corner of the first panel. Fig. D1.—
Correlation between the visual extinction, A V , and the integrated intensity of CO ( J =1–0), W CO . The gray dots denotesall data points in the map. The color-coded dots show the data points in the labeled regions, and the color represents the excitationtemperature, T ex . The black contours indicate the surface density of the color-coded points at the 15%, 45%, and 75% levels of themaximum surface density in each panel. Fig. D2.—
Correlation between the visual extinction, A V , and the integrated intensity of C O ( J =1–0), W C O . The gray dots denotesall data points in the map. The color-coded dots show the data points in the labeled regions, and the color represents the excitationtemperature, T ex . The black contours indicate the surface density of the color-coded points at the 15%, 45%, and 75% levels of themaximum surface density in each panel. TABLE 1Parameters of our observations.
Molecular line CO ( J =1–0) CO ( J =1–0) C O ( J =1–0)Rest Frequency [GHz] 115.27120 110.20135 109.78218Primary beam HPBW [ (cid:48)(cid:48) ] ∼ ∼ ∼ ] 44 ×
44 42 ×
43 30 × (cid:48)(cid:48) ] 7.5 7.5 7.5Effective beam size [ (cid:48)(cid:48) ] 21.8 22.2 22.2Velocity resolution [km s − ] 0.1 0.1 0.1Typical noise level in T mbmb