Kinematics of the Envelope and Two Bipolar Jets in the Class 0 Protostellar System L1157
Woojin Kwon, Manuel Fernandez-Lopez, Ian W. Stephens, Leslie W. Looney
KKinematics of the Envelope and Two Bipolar Jets in the Class 0Protostellar System L1157
Woojin Kwon , , , Manuel Fern´andez-L´opez , , Ian W. Stephens , , and Leslie W. Looney [email protected] ABSTRACT
A massive envelope and a strong bipolar outflow are the two main struc-tures characterizing the youngest protostellar systems. In order to understandthe physical properties of a bipolar outflow and the relationship with those ofthe envelope, we obtained a mosaic map covering the whole bipolar outflow ofthe youngest protostellar system L1157 with about 5 (cid:48)(cid:48) angular resolution in COJ=2–1 using the Combined Array for Research in Millimeter-wave Astronomy.By utilizing these observations of the whole bipolar outflow, we estimate itsphysical properties and show that they are consistent with multiple jets. We alsoconstrain a preferred precession direction. In addition, we observed the centralenvelope structure with 2 (cid:48)(cid:48) resolution in the λ = 1 . O, C O, CO, CS, CN, N H + , CH OH, H O, SO,and SO . All the CO isotopes and CS, CN, and N H + have been detected andimaged. We marginally detected the features that can be interpreted as a rotat-ing inner envelope in C O and C O and as an infalling outer envelope in N H + .We also estimated the envelope and central protostellar masses and found thatthe dust opacity spectral index changes with radius. Subject headings: circumstellar matter — stars: formation — stars: individual(L1157) — stars: pre-main sequence — techniques: interferometric SRON Netherlands Institute for Space Research, Landleven 12, 9747 AD Groningen, The Netherlands Astronomy Department, University of Illinois, 1002 West Green Street, Urbana, IL 61801 Korea Astronomy and Space Science Institute, 776 Daedeok-daero, Yuseong-gu, Daejeon 34055, Republicof Korea Instituto Argentino de Radioastronom´ıa, CCT-La Plata (CONICET), C.C.5, 1894, Villa Elisa, Argentina Institute for Astrophysical Research, Boston University, Boston, MA 02215, USA a r X i v : . [ a s t r o - ph . S R ] O c t
1. Introduction
Bipolar outflows are the most energetic and remarkable phenomenon of star formation.Particularly, the youngest protostellar systems, the so-called Class 0 young stellar objects(YSOs), are characterized by a well-developed bipolar outflow with a massive envelope (An-dre et al. 1993). At the earliest stage of star formation, material strongly accretes ontothe central protostar from the envelope, presumably through an accreting disk. Meanwhile,a bipolar outflow is launched and helps in removing angular momentum of the accretingmaterial. Therefore, investigating the two main structures (envelopes and bipolar outflows)together is crucial for understanding the early stages of star formation and evolution. Notethat what we call a bipolar outflow in this paper is a jet ejecting material rather than a slowmolecular outflow mainly consisting of the interacting features between the ejected jet andthe ambient gas like a bow shock (e.g., Frank et al. 2014). Because of this, the terms outflowand jet are used interchangeably in this paper.Although much is known about bipolar outflows, such as their general physical andchemical properties, impacts on the environments, and evolutionary changes (e.g., Arceet al. 2007), there remains many unanswered, fundamental questions: What is the launchingmechanism? Is it rotating? Is it precessing and if so, what causes the precession? Theo-retical studies agree that bipolar outflows are launched magnetocentrifugally (Blandford &Payne 1982; Ferreira et al. 2006; Ferreira 1997; Cerqueira et al. 2006), but the launchingmechanism and the launching regions are not yet clear (c.f., X-wind and disk-wind models,Shu et al. 1994; Konigl & Pudritz 2000). Bipolar outflows are expected to be rotating dueto magnetocentrifugal launching and angular momentum conservation for the accreting ma-terial. Indeed, a small number of previous observational studies toward Class 0 YSOs havereported a velocity gradient across bipolar outflow widths, which can be interpreted as rota-tion (Pety et al. 2006; Launhardt et al. 2009; Zapata et al. 2010; Choi et al. 2011). However,different interpretations are possible such as multiple jet events (Soker & Mcley 2013). Inaddition, bipolar outflows often show a wiggling feature, which can be interpreted as pre-cession. However, with the exception of close binary systems, the precession mechanism hasnot fully been understood (e.g., Teixeira et al. 2008).L1157-mm is one of the archetypical outflow sources. It is embedded in an isolatedstar forming cloud located in Cepheus at a distance of 250 pc (Looney et al. 2007, andreferences therein). With a bipolar outflow oriented nearly on the plane of the sky, L1157is one of the most extensively investigated Class 0 YSOs over a wide range of wavelengthson various topics. Looney et al. (2007) detected a flattened envelope structure in silhouetteusing the IRAC data of the
Spitzer Space Telescope (hereafter
Spitzer ). Kwon et al. (2009)found that grains have significantly grown already at this earliest stage, and the density 3 –distribution has a power-law index of ∼ ∼ (cid:48)(cid:48) resolution of CO and SiO in the southern lobe have suggested multiple bow shocks (e.g.,Gueth et al. 1996; Gueth et al. 1998). In addition Nisini et al. (2010) found that in theshocked region, 20% of all cooling is due to water molecules, and Santangelo et al. (2013)distinguished warm and hot components in the shocked regions using the far-infrared spec-troscopic imaging data of the Herschel Space Observatory .In this paper, we study the bipolar outflow and the envelope structure of L1157 usingCARMA. We report various molecular line observations as well as continuum at λ = 1 . (cid:48) with 5 (cid:48)(cid:48) angular resolution. In addition, basedon a multi-jet model fitting to the data cube, we find the physical parameters of the bipolaroutflow. We first describe the details of the observations in Section 2. Then, the resultsof the central envelope area are presented and discussed in Section 3. In Section 4 and 5,we show the outflow mapping image, analyze the data, and discuss the physical meanings.Finally, we enumerate the main conclusions in Section 6.
2. Observations and data reduction
We have obtained continuum and molecular line data toward L1157 using CARMA. Thedata were taken between 2010 July and October. An additional data set, originally taken asa CARMA summer school project to carry out polarimetric observations toward L1157 in2011 July (Stephens et al. 2013), has also been included. The observations are summarizedin Table 1. For each molecular line transition, Table 2 shows the rest frequencies, synthesizedbeam size, velocity resolution, and the root-mean square (RMS) noise per channel.We have observed the entire bipolar outflow spanning over 5 (cid:48) in CO J=2–1 using 25mosaic pointings. When designing the mosaic pattern, we used the
Spitzer µ m image 4 –(Looney et al. 2007). All 25 mosaic points were observed between each of the phase calibratorobservations to achieve a constant sensitivity over the entire field. We have employed themost compact CARMA E-array configuration in order to maximize sensitivity for the largestscales possible, i.e., to minimize the missing flux associated with the absence of zero spacingvisibilities for interferometric observations. The configuration provides an angular resolutionof about 5 (cid:48)(cid:48) at λ = 1 . (cid:48)(cid:48) scales given the uv coverageof CARMA (Kwon et al. 2009). Four bands among 8 spectral bands available in CARMAwere set to a wide bandwidth of 500 MHz for continuum, and the other four bands wereused to observe the spectral lines CO, C O, CN, CH OH, and H O. We have configured theCARMA correlator with a bandwidth of 125 MHz ( ∼
370 km s − ) and a velocity resolutionof about 0.5 km s − at λ = 1 . (cid:48)(cid:48) resolution. We employed the samecorrelator set-up as the bipolar outflow observations.In addition, we have observed at λ = 3 mm toward L1157-mm in the C-array config-uration to match the angular resolution of the D-array data at λ = 1 . O and N H + (in the lower and upper sideband respectively) and four31 MHz bands for the other molecular line transitions such as CO, CH OH, SO, and SO .The last three bands were set to the wide band of 500 MHz for observing continuum.We also included the E-array data of the CARMA summer school project, which hadthe primary goal of studying magnetic fields associated with L1157. The data consist ofcontinuum, CO, and CS at λ = 1 . λ = 1 . λ = 1 . (cid:48)(cid:48) and for the bipolar outflowand the summer school data (CS) about 5 (cid:48)(cid:48) . The velocity resolutions and sensitivities ofindividual continuum and molecular lines are also listed in the table. We have not detectedany structures in CH OH, H O, SO, and SO for the achieved sensitivities.
3. Central envelope region
To date, it is still nearly impossible to achieve high angular resolution data with agood sensitivity for directly observing outflow launching regions. A reasonable approach,therefore, is to model the overall features of the bipolar outflow and obtain an insight for thesmall scale features such as launching regions. In addition, the properties of the launchingregion may be related to the physical properties of the envelope, which is the reservoir of theaccreting material. For example, the kinematics of the envelope are expected to match thoseof the bipolar outflow launching regions. In this section we show and discuss the results ofthe envelope properties, and the observational and modeling results of the bipolar outfloware presented in the next section with an interpretation connecting the two structures.
The λ = 1 . β ) map with the assumption ofoptically thin emission: dust mass absorption coefficient κ ν ∝ ν β . As the Rayleigh-Jeanslimit could be invalid at a low temperature even for millimeter wavelengths, we use theformula with the exponential terms of blackbody radiation: β = log (cid:16) I ν I ν exp( hν /kT ) − hν /kT ) − (cid:17) / log (cid:16) ν ν (cid:17) − , (1)where I ν is an intensity at a frequency ν , h is the Planck constant, k is the Boltzmannconstant, and T is a temperature. We adopted an envelope temperature of 20 K, whichis reasonable for a Class 0 YSO envelope and consistent with radiative transfer modeling(private communication with H.-F. Chiang). At a temperature of 20 K, the β value estimatedusing the Planck function is ∼ . λ = 3 mm continuum has asynthesized beam of 2 . (cid:48)(cid:48) × . (cid:48)(cid:48) − ◦ , a position angle measured counterclockwise fromnorth) as listed in Table 2. In order to compare with the λ = 1 . β map properly, we used a Gaussian-function tapering during map constructionand achieved a beam of 2 . (cid:48)(cid:48) × . (cid:48)(cid:48) − ◦ ), which is nearly the same as the λ = 1 . . (cid:48)(cid:48) and is located at R.A. (J2000) = 20 h m . s and Dec (J2000) = +68 ◦ (cid:48) . (cid:48)(cid:48)
77, which isconsistent with higher angular resolution observations (e.g., Chiang et al. 2012).The total continuum fluxes at λ = 1 . (cid:48)(cid:48) × (cid:48)(cid:48) box about thephase center are 0 . ± .
018 and 0 . ± .
002 Jy, respectively. Note that the absoluteflux calibration uncertainties are not included. These continuum fluxes provide an estimateof circumstellar material mass (circumstellar disk and envelope, but mainly envelope mass)assuming the optically thin case (e.g., Looney et al. 2000): M ≈ F ν D /κ ν B ν ( T d ) (Hildebrand1983), where F ν is the integrated flux density, D is the distance to L1157 of 250 pc, κ ν is themass absorption coefficient, and B ν ( T d ) is the Planck function at dust temperature T d . Weassume κ ν is 0.01 cm g − at λ = 1 . M env ≈ . ± . M (cid:12) , which is consistent with previous studies ofradiative transfer modeling using a similar mass absorption coefficient (e.g., Kwon et al.2009). Free-free emission of low-mass YSOs is minimal (at most a few percent) at millimeterwavelengths, and in the case of L1157 it is approximated at a negligible level of 0.5–0.7 mJyat λ = 1 to 3 mm (Tobin et al. 2013). Note that the mass estimate could be uncertain upto about a factor of two due to uncertain mass absorption coefficients. The mass estimateerror is purely from the statistical error of the total fluxes: absolute flux calibration errorsup to 15% (Section 2) is not included. In addition, we estimated the deconvolved sizes byGaussian fitting: 2 . (cid:48)(cid:48) × . (cid:48)(cid:48)
98 (PA = − ◦ ) and 1 . (cid:48)(cid:48) × . (cid:48)(cid:48)
54 (PA = − ◦ ) at λ = 1 . λ = 3 mm aremade in the tapered map explained in the previous paragraph.The right panel of Figure 1 shows how the derived dust opacity index, β , changesthroughout the envelope of L1157 and is shown only at locations where both wavelengthcontinua have a signal-to-noise higher than three. As shown in Figure 1, the central regionof the envelope has a smaller β which is close to 0.3 (suggesting larger dust grains), and theboundary region has a large value around 2 (similar to those estimated in the interstellar 7 –medium, implying small sub-micron grains). Note that β is most sensitive to the grainsize (e.g., Draine 2006). This trend is understandable in the sense that grains grow fast in adense central region, and the boundary has similar conditions to the interstellar medium, e.g.,density and radiation field. The mean β , based on the total fluxes of the two frequencies, is0.76. In addition we present both wavelength data and β with uv distances in Figure 2, whichalso shows a variation of β . We shifted the phase center to the continuum peak for the plot,and calculated β by Equation 1 assuming an optically thin case. In fact, β decreases with uv distance; the dense central region has a smaller β suggesting large grains. Large opticaldepth can also cause changes in β ; however, radiative transfer modeling could not explainsuch a large variation of β only with optical depth effects even in a more massive envelopeobject (Kwon et al. 2009). Indeed, although the uncertainties were large, data presentedin Kwon et al. (2009) also show the same trend. In contrast, Chiang et al. (2012) did notdetect such a variation of β for the envelope of L1157. We suspect that the discrepancywas caused by absolute flux calibration uncertainties, especially when Chiang et al. (2012)combined long baseline data taken in A and B configurations. In the common uv ranges,both data are in agreement within the absolute flux calibration uncertainties. In this section, we present the results of various molecular line observations toward thecentral envelope region with about 2 (cid:48)(cid:48) resolution. Figure 3 shows the integrated intensitymaps of six molecular lines overlaid with the CO J=2–1 map . The blue- and red-shiftedcomponents of the central region of the outflow are shown in blue and red contours, respec-tively. Since the bipolar outflow is nearly on the plane of the sky ( i ≈ ◦ , Bachiller et al.2001), the blue- and red-shifted components are clearly distinguished on opposite sides. Inaddition, the integrated intensity peaks are not located in the middle of both blue- andred-shifted lobes. Instead, they trace two edges in a conic shape.The three CO isotopes are differently distributed. C O J=2–1 shows a strong signalat the envelope and seems to trace the bipolar outflow edges, presumably entrained gas. Incontrast, CO J=1–0 is detected toward the envelope and the CO bipolar outflow peaks.C O J=1–0 does not trace the bipolar outflow and instead shows the envelope elongatedperpendicular to the outflow. Interferometric observations suffer missing flux issues due tolack of short baselines. In the case of CO J=1–0, the brightness temperature of single dish We indicate the rotational transitions only for CO and its isotopes. The spectral line details are foundin Table 2. (cid:48)(cid:48) . As the total flux of our data over the same region is about2 Jy, the missing flux is approximately 87%. As shown in Figure 5 and 6 of Gueth et al.(1997), however, the large missing flux is mostly due to the large-scale low level emission.Missing flux is small toward the central peak and narrow jet features. For C O J=1–0 andC O J=2–1, whose features are likely more compact, the missing flux is expected to be evensmaller.Differences among these three isotopes are also found in the line profiles. Figure 4 showsthe six line profiles at their peak positions. CO J=2–1, which arises from both the envelopeand the bipolar outflow, has a self-absorption feature (or extended features filtered by theinterferometric observations) and a broad wing component. C O J=2–1 and C O J=1–0have narrow line profiles, and a Gaussian profile fitting gives a full width half maximum(FWHM) of 1 . ± .
19 and 1 . ± .
17 km s − , respectively. The hyperfine structures ofC O span over about 1.8 MHz: 224.7135334, 224.7141870, 224.7147438, and 224.7153100GHz (the Cologne Database for Molecular Spectroscopy, M¨uller et al. 2001). However, thelowest frequency line at 224.7135334 GHz is negligible since its integrated intensity is lowerthan the others up to by an order of magnitude. Therefore, the effective frequency span ofthe hyperfine structures would be about 1.1 MHz, which corresponds to about 1.47 km s − at the rest frequency. This explains the broad linewidth of C O: √ . + 1 . ≈ . − . In Section 3.4, we will also show that there exists a velocity gradient for C OJ=2–1 and C O J=1–0 about the envelope that may suggest rotation.CS and CN also have clear detections. However, their peaks are located toward eitherthe red- or blue-shifted outflow lobe. The peak of CS is north of the continuum center, whichis toward the red-shifted outflow lobe, and the peak of CN is south, at the blue-shifted lobe.The line profile of CN has a FWHM of 0 . ± .
18 km s − and a velocity center of 2 . ± . − . CN is the narrowest line presented in this paper and is found to be blue-shiftedcompared to the systemic velocity of 2.5 km s − . The CS line profile is very similar to C Oin terms of the linewidth and the central velocity. However the angular resolution of theCS data is relatively poor, so the peak position is arguably consistent with the continuumpeak. As the critical densities of the two transitions we detected are similar (1.1–1.4 × cm − ), the discrepancy of CS and CN distributions may indicate the differences in formationand/or destruction. For example, CS and CN both trace bipolar outflows, but CN can bequickly destroyed, which results in a distribution closer to the heating source (e.g., Bachiller& P´erez Guti´errez 1997). Note that single dish observations detected CS and CN over largeregions of the bipolar outflow (e.g., Bachiller et al. 2001), but our observations are limitedonly toward the center. 9 –The distribution of N H + clearly shows anti-correlation with CO. N H + is found in theouter region of the envelope with a hole at the center, and in the southeastern region adjacentto the blue-shifted outflow lobe.This anti-correlation is well understood by the fact that N H + is destroyed by CO (e.g.,Stephens et al. 2015). As shown in Figure 4, the line profiles of N H + around the east andwest blobs marginally show multiple velocity components, as expected for infall features. Chiang et al. (2010) reported CARMA N H + observations with an angular resolution of7 (cid:48)(cid:48) and found that the N H + traces well the flattened envelope structure Looney et al. (2007)detected in silhouette by Spitzer . They did not detect the central hole destroyed by COat this angular resolution, but they did detect a double-peaked feature toward the centralregion with velocities around 2.7 and 3.1 km s − . They suggested that the two velocitycomponents indicate infall motion. In addition, a feature supporting a solid body rotationin a large scale of ∼ ∼ (cid:48)(cid:48) are not sensitive to the largescales of ∼ H + channel maps with the identical synthesizedbeam of Chiang et al. and compared the J F F = 101 →
012 hyperfine components (i.e.,the more isolated hyperfine component in the N H + spectra, located at a V LSR ∼ − . − in Figure 4). We found that the line profile peak of our data is about 0.7 Jy beam − while Chiang et al. obtained 0.9 Jy beam − , which implies that our observations filtersout ∼
22% of the N H + flux observed by Chiang et al. Despite the missing flux and poorsensitivity of our data, two velocity components are marginally detected toward each ofthe east and west blobs. For each blob, individual N H + spectra were fit with two sets ofGaussian velocity components. The 7 hyperfine structures were fit simultaneously; theirrest frequencies and relative strengths are adopted from Caselli et al. (1995) and Womacket al. (1992), respectively. Since the sensitivity is marginal, we simply assumed opticallythin emission and local thermal equilibrium. In addition, it is assumed that the two velocitycomponents have the same linewidth. The fits results in velocity components of 2 . ± . . ± .
06 km s − for the west blob and 2 . ± .
07 and 3 . ± .
09 km s − for theeast (Figure 5). The FWHM is 0 . ± .
06 km s − in the west blob and 0 . ± .
11 km s − in the east blob. The reported errors are statistical fitting uncertainties. Note that thespectra used for the fitting have a velocity resolution of 0.1 km s − . N H + is known to be 10 –optically thin for L1157, so the double peaks are unlikely due to a self-absorption feature.Due to the low signal-to-noise at high spectral resolution, we also investigate whether a singlevelocity component can fit the data. The reduced χ value for the single velocity fit is almostidentical to the two velocity fit. However, the line profiles show obvious signs of two velocitycomponents, particularly for the isolated, J F F = 101 →
012 hyperfine component (Figure5), suggesting two velocity components may be meaningful.These two velocities are consistent with previous studies of lower angular resolutionsdiscussing infall motion (e.g., Chiang et al. 2010). Since the two blobs present similar doublevelocity components rather than a velocity gradient, the outer envelope ( ∼ H + data is thought to be dominated by infall motion. The inner regionof the N H + hole is nicely imaged with CO isotopes and consistent with a rotation feature(Section 3.4). The outside wall of the bipolar outflow cavity is also traced by N H + in morecompact configurations (Chiang et al. 2010; Tobin et al. 2012). Hereafter, we will refer to theouter envelope as the region traced by N H + from 500 to 1000 AU, and the inner envelope asthe region traced by C O. Note that the dust continuum traces both these envelope scales.The outer envelope, as mapped by Chiang et al. (2010), extends to ∼ As shown in Figure 3, CO isotopes well trace the envelope structure. In particular C Opresents an elongated feature perpendicular to the bipolar outflow. In order to investigatethe kinematics of the inner envelope, we examined the position-velocity (PV) diagrams ofC O and C O cut through the phase center along two directions: the bipolar outflow (PA= 140 ◦ ) and the elongated envelope (70 ◦ ). The PV diagrams along the elongated envelopefeature show a velocity gradient, while the cut along the bipolar outflow does not. The leftand right panels of Figure 6 present the C O and C O PV diagrams cut in the elongatedenvelope. C O shows double peaks in the PV diagram with a positional offset. This featureis not due to the hyperfine structures of C O. Unresolved hyperfine lines produce broaderlinewidths (Section 3.2), and resolved hyperfine lines may show multiple peaks but not avelocity gradient. The velocity gradient corresponds to 1.1 km s − over 0 . (cid:48)(cid:48) O features are relatively complicated, but the signals are mainly located in thetop right and bottom left in the diagram, which means that the eastern part of the elongatedenvelope is redshifted and the western part is blueshifted. This overall velocity gradient is in 11 –agreement with C O. Note that the green lines in the two panels connect the same locationsin position and velocity. The features shown in both C O and C O could be a rotation,which is clockwise when looking down from the north. The 3.1 km s − component detectedaround an offset of − (cid:48)(cid:48) is at the boundary of the outer envelope traced by N H + , suggestingthis could be infalling material.The mass of the central objects can be estimated from the rotational features of themolecular lines. Although our data are not high enough quality for clear evidence of Keple-rian rotation, it is worthy to estimate the central object mass based on the interpretation.In Figure 6 we also overlaid Keplerian rotation curves of a central protostar with a mass of0.04 and 0 . M (cid:12) with blue and orange lines, respectively. The mass of 0 . M (cid:12) matches theoverall outline of the C O features in the PV diagram and the velocity gradients detected inboth C O and C O. The mass of 0 . M (cid:12) is marked to show how much the curve changeswith mass. The bipolar outflow is known to be nearly in the plane of the sky, which presum-ably suggests that the rotation feature is close to edge-on. Nevertheless, we show two dashedlines for the cases of 60 ◦ inclination from the plane of the sky. The 60 ◦ inclination line isequivalent to half the rotational velocities at a given radius. In other words, if the distancewere indeed a half of 250 pc, the edge-on case is shown by the dashed lines. Similarly, theopposite case of a 500 pc distance can simply be estimated in the plot.For the Keplerian rotation curves, we did not take into account the envelope mass withina given radius: M ( R ) ≈ M env ( R/R out ) − p , where M env is the total envelope mass (0.58 M (cid:12) ,as derived in Section 3.1), R out is the envelope size (5 (cid:48)(cid:48) ), and p is the power-law index of avolume density distribution. When assuming p = 1 . . (cid:48)(cid:48)
35 (the radius of the offset between the two peaks in the PV diagrams connectedby a green line in Figure 6) is 0.02 M (cid:12) . Due to the cavity swept by the bipolar outflow andthe sharply increasing dust temperature close to the center, the mass inside is likely smallerthan this estimate. This implies that Keplerian rotation tends to appear inside around 0 . (cid:48)(cid:48) . (cid:48)(cid:48)
55 ( ∼
140 AU at a distance of 250 pc) where the envelope mass inside is comparableto the protostellar mass (0 . M (cid:12) ): v = ( GM ( R ) /R ) / , i.e., v ∝ R − p . Again, on the scalesof 1000 AU, infall motion is detected in N H + .Although it is not clear due to the limitation of our data in sensitivity and resolution,Figure 6 indicates that the outer contours and the overall gradient in the green line can beexplained by Keplerian rotation around a 0 . M (cid:12) protostar. Note that we adopted thedistance of 250 pc, which is most reasonable (Looney et al. 2007), but the mass estimatedecreases or increases linearly depending on the distance. Gueth et al. (1997) estimated thecentral protostellar mass as 0 . M (cid:12) based on the velocity gradient detected in C O, but 12 –they assumed the distance of 480 pc and overestimated the interval across the 1 km s − velocity gradient roughly as ∼ (cid:48)(cid:48) , which both cause an overestimate of the protostellar mass.Since Gueth et al. (1997) interferometric observations have a comparable angular resolutionto our data, missing flux might not cause a significant bias. It is not easy to comparethe two interferometric data sets since they have different channel widths and a weighting.Their Figure 9 presents a channel map with uniform weighting and shows intensity peaksseparated by 3 (cid:48)(cid:48) , but the span only appears well in two adjacent channels (2.81 and 3.02km/s), not blue- and red-shifted channels. On the other hand, their Figure 4 that uses anatural weighting shows a similar position span for the velocities we observed. Consideringthe uncertainties and differences of both data sets, we argue that our interpretation is notso problematic, and that the two estimates are consistent. Based on the dust continuum,we estimated the envelope mass as ∼ . M (cid:12) (Section 3.1), which gives an envelope toprotostellar mass ratio of about 14.
4. Bipolar outflow4.1. Outflow activity
The left panel of Figure 7 shows the CO emission integrated in velocity (moment 0)observed in the E-array configuration (black contours) at a 5 (cid:48)(cid:48) spatial resolution overlaid onthe 8 µ m emission taken by the Spitzer telescope (Looney et al. 2007). The figure clearlyshows overlap of the 8 µ m and CO emission. Note that the Spitzer µ m channel maybe dominated by H (cid:48)(cid:48) from the central protostar, the COemission shows a point symmetric X-shape morphology composed of two narrow ridges of ∼ (cid:48)(cid:48) width. At further distances, the CO emission consists of several bright knots along anelongated curved structure spanning in total ∼ (cid:48)(cid:48) (75000 AU). Most of the bright knotshave been extensively reported and studied in the past (e.g., Bachiller et al. 2001), includingB0, B1 and B2 in the blue-shifted (southern) lobe and R, R0, R1 and R2 in the red-shifted(northern) lobe as indicated in Figure 7.The upper right panel of Figure 7 is a zoom-in of the central region, showing contoursof the CO emission observed in CARMA D configuration at 2 . (cid:48)(cid:48) Spitzer µ m emission and the yellow tones indicate the well-knownextinction lane perpendicular to the outflow. The observations show a clear X-shape pattern,with the strongest emission in the northwest and southeast arms. The average FWHM of 13 –these arms is about 3 . (cid:48)(cid:48)
3, which is about 2 . (cid:48)(cid:48) − to +12.5 km s − (notethat the velocity is indicated with respect to the cloud velocity V LSR = +2 . − , section3.4). By inspecting the image, it is evident that close to L1157-mm, the northwest-southeastarm of the X-shaped structure has an average absolute velocity 2–3 km s − higher than thenortheast-southwest arm. The arms of this X-shaped emission has been interpreted as thelimb-brightened cavity walls of the outflow (e.g., Gueth et al. 1997; G´omez-Ruiz et al. 2013).However, our data show certain similarities between the northwest and southeast arms inmorphology, symmetry, brightness, and kinematics of the emission, suggesting componentsof one jet. The same occurs with the northeast-southwest arms, though fainter, suggestingthese arms could be another jet. In Section 4.2 we discuss the possibility that the COemission is associated with multiple outflows or molecular jet ejections from the centraldisk-protostar system.Even though CARMA is filtering out the fainter emission between the two bright narrowridges shown by the single-dish and interferometer combined CO J=1–0 map of Gueth et al.(1996), which interpreted the system as a limb-brightened cavity, the possibility of two jets isnot ruled out. Gueth et al. (1996) reported a very high emission contrast between the ridgesand the region enclosed by the ridges, which they could not completely model even using anexponential decay in the density profile of the putative cavity. The compactness of ridgesallows CARMA to detect them: the largest angular scale of CARMA in this configurationgoes up to ∼ (cid:48)(cid:48) . The proposed multi-jet model does not need to recover all the extendedflux to be tested and has the advantage of explaining several observed features that cannotbe explained by the cavity model. As mentioned in Gueth et al. (1996), the cavity modelhas a difficulty in explaining (1) the edge-inner cavity emission contrast, (2) the morphologyof the emission at high velocity showing only the cavity walls, and (3) the non-homogeneousemission of the inner cavity showing several arches and small filaments.Figure 8 shows the CO emission integrated over 6 arbitrarily selected velocity intervalsso that the overall morphology and kinematics of the outflow are adequately depicted. Theextended emission at velocities close to the systemic velocity is filtered by the interferometerand hence not shown. Although the rest of the velocity channels are also affected by thepresence of extended emission, the main issue is the dynamic range of the images since thenarrow outflow ridges are much brighter than the emission from the putative inner cavitymaterial (Gueth et al. 1996). Nevertheless, we have checked that the sidelobes have littleimpact in the brightness distribution of the ridges themselves. We also noticed in the COmoment 0 image, there are some artifacts at the edges of the images. In particular, theemission features west of R1 and R0 and west and east of B1 are sidelobes of the strongest 14 –outflow ridge. These edge problems only appear in the channels (Fig. 8) where this ridge isstronger. On the contrary, the emission southeast of R1 is clearly not an artifact. One cansee for instance, that its curvature is the opposite from the main northwest ridge passingthrough R0.The redshifted high velocity CO emission (+16.2, +32.0 km s − ; Fig. 8 a) traces partof the northwest outflow lobe with some finger-like features protruding northwest from thecurved molecular jet path. At medium velocity (+9.6, +15.7 km s − ; Fig. 8 b) the COemission runs mainly from the protostar position and extends northward to knot R alonga western curved ridge. There is also emission along the eastern curved feature southeastof R1, although this feature was not well mapped due to the limited field of view of ourobservations. Some emission is also detected north of knot R. The redshifted low velocityemission (+4.0, +9.1 km s − ; Fig. 8 c) is similar to that of the medium velocity emission.The emission runs outward from the protostar along the knots R0, R1, R, and R2. COemission is also seen northeast of L1157-mm and southeast of R1 which may indicate adifferent curved jet path. This velocity channel shows emission at knots R and R2. Alsoevident is that the curved features cross each other at the infrared knots R and R1. Close tothe cloud velocity (+1.5, +3.5 km s − ; Fig. 8 d), the emission associated with knots R0 andR1 becomes fainter, while that at knot B2 in the southern blue lobe becomes obvious. For theblueshifted low velocity (–2.0, –0.5 km s − ; Fig. 8 e), knot B2 shows an extended morphology.Hints of arched structures apparently crossing each other are seen at B1 and B2. At bothlow and high blueshifted velocities, the main southeast curved ridge is seen from L1157-mmand along knots B0 and B1. In addition, there is a blueshifted emission component in themiddle of the northern lobe at ∼ (cid:48)(cid:48) north of L1157–mm. At high blueshifted velocities(–19.8, –2.6 km s − ; Fig. 8 f), an arch joining L1157-mm with the western part of B1 formsthe southwest curved ridge.We do not include velocity dispersion (second moment) maps in this paper because thespectral profiles of the CO emission show extended emission wings that appear masked bythe core emission of the line profiles. Pronounced wings (extending up to ∼
33 km s − )are commonly found along all the knots marked in Fig. 7, while the rest of the gas havelinewidths between (1–3) km s − .Figure 9 shows a zoom-in of the north lobe of the outflow. It shows the infrared 8 µ memission overlaid with the CO redshifted high velocity emission between +32 km s − and+18 km s − . This image shows two clear finger-like features at 140 (cid:48)(cid:48) and 150 (cid:48)(cid:48) from theL1157-mm position at a PA of –60 ◦ and –50 ◦ respectively. There is also possibly a thirdfinger at 155 (cid:48)(cid:48) away from L1157-mm at PA = − ◦ . The tip of these fingers spatially coincidewith some of the mid-IR knots studied by Takami et al. (2011). All three features are marked 15 –in the figure with a straight line joining each of them with the position of L1157-mm. Theirlinear trajectory along with their high velocity suggests that they can be molecular bulletsejected with different position angles. Therefore, these fingers indicate a possible rotationand precession of the ejecting source. Previous work based on molecular observations towards L1157 (Gueth et al. 1996, 1998;Bachiller & P´erez Guti´errez 1997) have proposed the existence of an episodic precessingjet which drags the environment and excavates a wide cavity. The cavity walls are limb-brightened and hence easily detected in molecular emission. However, the L1157 outflowshows a point reflection symmetry of the strongest clumps at the northwest and southeastputative cavity walls. The symmetry of the walls is seen via three physical parameters:shape, radial velocity, and clump structure. Thus, previous investigations (Zhang et al.2000; Bachiller et al. 2001) indicated the possibility that not only the CO emission butalso the SiO emission (a well known tracer of the shocked gas in molecular outflows andjets) in L1157 is tracing one narrow molecular precessing jet, coincident with the northwest-southeast CO arm seen close to the protostar. The idea of a precessing jet is reinforcedby the finding of high velocity shocks at different position angles in the north lobe of theoutflow (see Section 4.1). Bachiller et al. (2001) also speculated the existence of a milderwide-angle wind surrounding this narrow precessing jet, which would account for the weakCO molecular emission (undetected in SiO) to the northeast of the red (northern) lobe andthe southwest of the blue (southern) lobe. However, this scenario can poorly account forthis “anomalous”emission since it is not detected at both sides of the outflow simultaneously(i.e., there is no CO emission west of the narrow SiO molecular jet in the north lobe and tothe east in the south lobe). Moreover, the velocity of a wide-open angle wind is expectedto decrease with the opening angle from the outflow axis (Arce & Sargent 2004). However,such a characteristic is not found in our data.On the basis of the new CARMA CO observations, we explore the possibility of theejection of multiple jets from the L1157 protostar. In particular we fit a two precession jetmodel to the CO data. The idea of a precessing jet needs a mechanism to eject materialthat follows ballistic trajectories. Evidence for high velocity ballistic ejections are clearlyshown in Figure 9. In addition, there is evidence suggesting that some of these bullets arecometary-shaped and run parallel to each other in the southern lobe (Takami et al. 2011).This matches well with the idea of multiple jets or ejections being present in the outflow.We suggest that the initially supposed limb-brightened cavity edges (e.g., Gueth et al. 16 –1996, 1998) can possibly be the path of two molecular jets (see Figure 10 for a sketch ofthis new interpretation). The CO CARMA images (Figures 7 and 8) show a clear wigglingpath running northwest-southeast close to the central protostar. It coincides well with thepreviously identified SiO molecular jet, which we call
Jet 1 . The CO emission shows alsoanother S-shaped feature running northeast-southwest close to the protostar which we call
Jet 2 . Jet 2 shares similar physical characteristics with
Jet 1 : (1) a sinuous trajectory, (2)a similar physical width and length, (3) a similar linewidth of the line core, and (4) a pointreflection symmetry that is especially clear at the center of the system (Figure 7). However,the radial velocity of
Jet 2 is closer to the systemic velocity than for
Jet 1 . Another differenceis that
Jet 1 shows in general, lines with more extended wings than
Jet 2 .In Section 4.3, we present a numerical fit to the L1157 CO outflow based on a twoprecessing molecular jet model, which is in contrast to the wide-angle wind scenario. Wepropose this modeling as a proof of feasibility of a multiple jet model, although this kind ofsystem has been recently proposed for another Class 0 object, NGC 1333 IRAS 4A2 (Soker& Mcley 2013). In that case, the authors tried to explain the observed velocity gradientin the gas of the outflow (Choi et al. 2011), which is perpendicular to the outflow axis asproduced by the ejection of multiple jets.In the two molecular jet scenario, the jets can be launched at the same time or atdifferent epochs. A system simultaneously ejecting two jets may be possible if there are twoprotostars ejecting two corresponding jets (Murphy et al. 2008). A striking example of abinary system (or triple) in which both protostars are launching bipolar precessing ionizedjets is L1551 IRS 5 (Rodr´ıguez et al. 2003; Pyo et al. 2009; Lim & Takakuwa 2006; Itohet al. 2000; Fridlund & Liseau 1998). This system has a separation of 45 AU between thetwo circumstellar disks, each of 10 AU in radius. Recent VLA observations of L1157 (Tobinet al. 2013) have shown a compact although somewhat elongated disk-like structure with aradius of ∼
15 AU, indicating that only one protostar is present in the system at a 12 AUresolution. On the other hand, a re-evaluation of the same data with a weighting moresensitive to find structures (i.e., robust = 0) suggests that there could be a binary systemwith ∼
15 AU separation (J. Tobin, private communication). At this time, we cannotdetermine if L1157 is a close binary system or not.In the case that L1157 is a single protostellar system, the two molecular jets could beejected at different times, which could explain why the CO emission of
Jet 2 is weaker thanthat of
Jet 1 and why
Jet 2 does not show SiO emission (thought to disappear in older andslower shocks, see e.g., Codella et al. 1999; Miettinen et al. 2006; Fern´andez-L´opez et al.2013). Thus,
Jet 2 would be a remnant of an older ejection that entrained the molecularenvironment. Although L1551 IRS 5 observations support the scenario of two jets, there 17 –is one noticeable difference between the system and L1157. The former has jets with amarkedly dissimilar radial velocity (separated by ∼
140 km s − ), possibly indicating two wellseparated ejection axes, while in L1157 the two jets show similar radial velocities, which mayindicate similar ejection axes with only a small inclination difference. This axis would beinclined differently which may be due to secular motions of the system itself. On the otherhand, there could be different ejection mechanisms launching multiple jets simultaneously,e.g., a circumstellar disk with an accreting spiral pattern of multiple arms launching a jeteach (see Soker & Mcley 2013). This discussion lies beyond the scope of this paper. As stated in the previous section, we suggest that L1157 may be ejecting two precessingbipolar jets, which we called
Jet 1 and
Jet 2 . For each jet, we fit the data using a precessingjet model put forward in Raga et al. (2009). This model consists of a jet launched by adisk-protostar system with a precession determined by half the opening angle ( α ) and acharacteristic frequency, ω . The bipolar jet is affected by the precession of the launchingsystem and ejects material running into ballistic trajectories with constant velocity, v j . Forthis analysis, we did not consider additional perturbations due to an orbital binary motionsince we do not know if L1157 is a binary system and this kind of perturbation displaysa pattern only discernible at smaller spatial scales (Raga et al. 2009). In our model, theejecta shows a helix-like geometry in which the radius of the helix gets wider with time. Thefollowing equations, as derived from Raga et al. (2009), describe the outflow motion in areference frame centered in the disk-protostar system (see Figure 2 in Raga et al. 2009): xyz = ( t − τ ) v x v y v z , (2)where the velocity of the ejection is v x v y v z = v j sin α cos ( s prec ωτ − φ ) − sin α sin ( s prec ωτ − φ )cos α . (3)In these equations, t is the present time and τ is the “negative” time at which a parcelof the jet was launched such that t = 0 means “now”. s prec defines the direction of rotationwith +1 indicating clockwise (when seen from top of the Dec axis) and –1 indicating counter-clockwise rotation. φ is the phase angle in the xy-plane of the current ejection at τ = 0. 18 –Also we allow a possible inclination angle i with respect to the sky plane and a positionangle, pa , defined counter-clockwise from north. Thus, we doubly rotate the jet system in xyz coordinates to get the plane of the sky x (cid:48)(cid:48) y (cid:48)(cid:48) z (cid:48)(cid:48) coordinates, where x (cid:48)(cid:48) is in the RAdirection pointing west, z (cid:48)(cid:48) is in the Dec direction toward north and y (cid:48)(cid:48) is perpendicular tothe plane of the sky, pointing away from us. Given the data provided by the observations( x (cid:48)(cid:48) , z (cid:48)(cid:48) , and v (cid:48)(cid:48) y ) and assuming the velocity of the ejection does not change, one can obtainthe ejection time τ for each data-point using: τ = − (cid:115) ( x (cid:48)(cid:48) ) + ( z (cid:48)(cid:48) ) v j − ( v (cid:48)(cid:48) y ) . (4)As can be seen from the parametric equations above, τ can be used as the parameter , toderive a synthetic model, once a set of free variables ( s prec , v j independent for each jet lobe, α , ω , i , pa , and φ ) is determined. This can be used to speed-up the fitting process.With the model at hand, we extracted the information on position ( x (cid:48)(cid:48) , z (cid:48)(cid:48) ) and radialvelocity ( v (cid:48)(cid:48) y ) of the molecular emission for each jet by picking out points of peak emissionfrom each 2.5 km s − binned velocity channel image. All the selected points were above3 σ and followed an arbitrarily designed jet path. After repeating the data selection processseveral times for each jet, the final Jet 1 and
Jet 2 data points were defined (Figure 11).Figures 11, 12, and 13 show the radial velocity and trajectories of the model compared withthe observed data. As can be seen from these figures, there is good position and kinematicagreement between model and data. Table 3 shows the values obtained in fitting the twojets. The model gives higher velocities for the northern lobes (60–90 km s − ) than for thesouthern lobes (30–40 km s − ) in both jets, which may be expected since the northern lobeextends farther than the southern lobe. This could imply, for instance, that different lobesare ejected at different intrinsic velocities or with different inclination angles. We choose ourmodel to have two independent velocities for each north-going jet and south-going counter-jet. The CO may be tracing the jet directly ejected from the surroundings of the central star,or material swept up by the jet just in the surroundings of the bullet path, making possibledifferences between the intrinsic velocity in the jet and counter-jet. On the other hand, theCO emission may be tracing the material swept up by the jet far from the bullet path. Inthis case, Bachiller et al. (2001) proposed that the environment surrounding the southernlobe is denser than that surrounding the northern lobe, thereby reducing more efficiently thevelocity of the southern lobe. Our model shows also an agreement between the precessingperiod of the jets (derived from the model, they are about T =5000-8000 years) and theirprecessing angles (11 ◦ ). However, the position and inclination angles are different for bothjets, suggesting a slightly different ejection axis orientation. The fact that the period andprecessing angle of both Jet 1 and
Jet 2 are similar could support the scenario of the two jets 19 –ejected at different times described in Section 4.2. In such a case, one problem is to explainthe different ejection axis orientation that we previously explained as secular motions of thesystem itself. Such a problem may simply be explained by two protostars each ejecting anoutflow.A note of caution should be made here in establishing that our fitting is aimed uniquelyto probe the feasibility of a two precessing bipolar jet model in order to explain the observedL1157 CO emission. We have not tried to univocally fit the model to the data in any case,and we are aware that other models with multiple jets could fit the data as well. For instance,
Jet 2 has a solution with s prec = − s prec = 1 shown here. This isprobably due to the poor sampling of Jet 2 in the north lobe and also because the relativeweakness of its CO emission compared to that of
Jet 1 . We choose not to show the solutionwith s prec = − Jet 1 which has s prec very well constrained. Again, even though the model can be regarded as somewhat arbitrary,its purpose is to qualitatively show the possibility that the system is composed by at leasttwo precessing molecular jets and to provide an order of magnitude estimates of the physicalparameters.
5. Discussion5.1. Possible causes of precession in L1157
The cause of the wiggling (precession) pattern of the protostellar jets is not well known.There are theoretical models explaining this phenomenon as caused by (1) the orbital mo-tion of a binary system (e.g., Masciadri & Raga 2002), (2) the precession due to the tidalinteraction between the disk of one protostar and a companion protostar in a non-coplanarorbit (e.g., Terquem et al. 1999a; Montgomery 2009), (3) the warp of the inner disk (in prin-ciple caused by a perturbing companion star) from which the jet is thought to be launched,and/or (4) the misalignment between the disk rotation axis and the ejection engine, withthe latter usually understood as an MHD disk-wind (Frank et al. 2014). At this momentthere is no conclusive empirical evidence supporting any of these scenarios. The first threemodels would imply a multiple (at least binary) protostellar system, while model (4) wouldjust require a single protostar. Nonetheless, in many (if not all) of the protostellar systemswith observed precessing jets, evidence indicate that they are mostly binary systems (e.g.,HH30, Guilloteau et al. 2008; HH211, Lee et al. 2010; H111, Reipurth et al. 1999), and hencewe do not explore the model of the jet launching engine misaligned with the disk rotationaxis, although we note that it should remain as an alternative mechanism. We rule outthe orbital motion model because it entails a mirror symmetry for the jet, a small outflow 20 –opening angle, and a very short period spiraling outflow (Raga et al. 2009), all of which arenot observed in our data. We cannot test the possibility of a warped inner disk since wedo not have observational data of the dust emission at AU scales. We thus test the tidalprecession model with the L1157 jets. The tidal precession model has a good theoreticalbackground that we use to derive the orbital parameters of a hypothetical binary systemproducing the precession of the jets. In particular, we use an equivalent form of equation(37) from Montgomery (2009) for circular precessing Keplerian disks. This equation relatesthe angular velocity at the disk edge ( ω d ), the Keplerian orbital angular velocity of the com-panion around the primary protostar ( ω o ), and the retrograde precession rate of the diskand the jet ( ω p ): ω p = − ω o ω d cos α , (5)where α is the inclination of the orbit of the companion with respect to the plane of thedisk (or obliquity angle), and this angle is the same as the half–opening angle in equation 3(Terquem et al. 1999b). The equation also assumes that both objects of the binary systemhave the same mass. Introducing the values for the jet precessing angle α and the precessionperiod τ p derived from our fit (Table 3) and adopting some reasonable value for the radiusof the disk r d responsible for launching the jet, we can constrain the orbital period τ o andthe orbital radius r o of the putative binary system. Tobin et al. (2013) showed that the diskradius is less than 15 AU, and we take 1 AU as a lower limit for it. From this we derivedorbital periods and radii in terms of the primary and secondary protostar’s solar masses, M and M : here M = M . For Jet 1 the orbital period is (50–370) M − / yr, while for Jet 2 the orbital period is (60–450) M − / yr. The orbital radius for Jet 1 is (13–52) M / AUand for
Jet 2 the orbital radius is (15–60) M / AU. It is evident that there is a goodagreement between the τ o and r o estimates for Jet 1 and
Jet 2 , as expected from the similarprecession periods and angles, and intrinsic velocities. Moreover, the derived orbital radiusis consistent with being a few times the disk radius which is expected for disks affected bytidal truncation (e.g., Artymowicz 1994). Indeed, using the derived mass of 0.04 M (cid:12) forthe protostars (Section 3.4, M = M = 0 . M (cid:12) ), we then obtain τ o =130–1200 yr and r o =7–31 AU for a disk with r d =1–15 AU. The precession direction that our models prefer is clockwise when looking down fromnorth ( s prec = +1, Table 3). Since precession due to tidal interaction of a companionin a non-coplanar orbit occurs in the opposite direction to the rotation of the disk (the so-called retrograde precession), the outflow launching structure would rotate counter-clockwise. 21 –However, our C O and C O data show a clockwise rotation of the inner envelope (easternside redshifted). This implies the disk and the inner envelope would be counter-rotating.We discuss how to understand this counter-rotation of the inner envelope and the outflowlaunching structure and the caveats. When discussing the launching structure, we specificallymean protostars of a binary system or a circumbinary/circumstellar disk.The precession direction of
Jet 1 is significantly well constrained. In contrast,
Jet 2 direction is relatively unclear: the clockwise and counter-clockwise cases are comparable inour modeling. This is probably due to the fact that
Jet 2 is not as strong as
Jet 1 . It isunrealistic to assume that the two jets are precessing in opposite directions. Therefore, wedid not consider this case. On the other hand, the inner envelope rotation features discussedwith our C O and C O data are also not decisive due to lack of angular resolution: thevelocity gradient could be contaminated by the bipolar outflow. The possibility is low,however, as the profiles are along the elongated direction nearly perpendicular to the bipolaroutflow, and the same directional velocity gradient in the inner envelope was also reportedby Gueth et al. (1997). Based on the argument above, the counter-rotation of the innerenvelope and the jet launching structure could be accepted.Such counter-rotation does not seem to be realistic in terms of angular momentum instar forming molecular core scales. However, some planets and moons show counter-rotatingmotions in the solar system and counter-rotating binary systems may be theoretically feasi-ble, for instance, as a result of the interactions of a triple system (Li et al. 2014). In addition,the kinematics of the L1157 flattened envelope are complicated. The large-scale solid bodyrotation detected in N H + by Chiang et al. (2010) is counter-clockwise, while their inten-sity weighted velocity map shows an overall direction switch at a smaller scale (Figure 3 inChiang et al. 2010). Counter-rotation indeed helps a circumbinary disk last longer and ismore effective during the binary system eccentricity evolution (Dunhill et al. 2014). It isnot known yet, but L1157 might be a close binary system at 15 AU scale (J. Tobin, privatecommunication). Although it is not ideal in the viewpoints of angular momentum, the bi-nary protostellar system and the circumbinary disk could be counter-rotating. However, thedetailed studies on how such a counter-rotating system forms is beyond the scope of thispaper.On the other hand, the modeling formalism we used is identical to a rotation case, soit is possible to interpret the modeling results as a clockwise rotation instead of precession.However, the periods of several thousand years that our models constrain are too large for areasonable launching radius. Assuming Keplerian rotation around the protostar, the periodsgive a 70–80 AU launching radius.Finally, the jet modeling formalism we used could also be applied to the precession of a 22 –misaligned launching structure with respect to the system rotation axis (like the mentionedcase 4 on previous section 5.1), so it is possible to interpret the modeling results as a prograderather than a retrograde precession. A prograde precession is also expected in the case of awarped inner disk. Higher angular resolution and better sensitivity observations will providedecisive kinematic results toward this interesting Class 0 protostellar system whose bipolaroutflow launching structure appears to be rotating in the opposite direction of the innerenvelope.
6. Conclusions
We present CARMA millimeter observations of the youngest Class 0 protostellar systemL1157. The central envelope region (L1157-mm) was imaged in λ = 1 . O, CS, CN, CO, C O, and N H + with 2 (cid:48)(cid:48) resolution. We also observed CH OH,H O, SO, and SO , but these lines were undetected with the sensitivity of these observations.The continua and the various line features allow us to estimate the physical properties of theenvelope. In addition, we obtained a large (5 (cid:48) ) mosaic image covering the bipolar outflow ofL1157 in CO J=2–1 with 5 (cid:48)(cid:48) resolution. Our results are summarized in the following:1. The envelope mass based on the dust continuum flux is estimated to be ∼ . ± . M (cid:12) . We also found that the dust opacity spectral index β changes along radius in theimage, which is also shown in the visibility data.2. Among CO isotopes, CO traces the CO peaks of the bipolar outflow, while C Otraces the edges. C O shows a structure elongated and perpendicular to the bipolar outflowon the scale of 100 AU (i.e., it may be tracing the inner envelope). We also detected a velocitygradient in C O and C O, which is consistent with Keplerian rotation around a protostarwith a mass of ∼ . M (cid:12) (approximately 1/14 of the envelope mass). The envelope isrotating clockwise when looking down from the north.3. N H + presents double peaks in east-west along the known flattened envelope directionwith a hole at the center. Each of the double peaks shows a line profile understood by twovelocity components, which suggests infall motion in the outer envelope of 1000 AU scales.4. From the bipolar outflow mosaic image in CO, we detected several ballistic ejectionsin the redshifted (northern) outflow that support multiple jets. We find that the bipolaroutflow could be interpreted as two jets for a reliable example, and we constrain this scenarioby modeling the data cube. The idea of two jets is supported by the following observationalevidence. 23 –i. The morphology of the CO emission agrees with the two jet scenario. High-angularresolution CO data show two apparent collimated molecular jets in an X-shape distributionclose to the protostar. In addition, the curved features at large-scale are well reproduced bythe precessing two jet model.ii. The CO brightness distribution is not well explained by single outflow cavity modelsbut can be explained by the two jet scenario. The two jet scenario can explain the brightnessasymmetry between the NW-SE ridge and the NE-SW ridge. The two jet scenario also canexplain the large brightness contrast between the edges and inner part of the lobes.iii. The radial velocity of the NW-SE jet gas is 2 − − higher than the NE-SWjet gas, which is unexpected for a single jet. Our model of two precessing jets can explainthe different radial velocities, as well as positions, of the NW-SE and NE-SW jets.iv. The spatial distribution of the SiO emission, which has been observed in the NW-SEjet but not in the NE-SW jet (Bachiller et al. 2001), can be explained if these two jets havedifferent ages.We are grateful to CARMA staff and observers for their dedicated work and anonymousreferee for valuable comments that allowed us to improve the paper significantly. Support forCARMA construction was derived from the states of Illinois, California, and Maryland, theJames S. McDonnell Foundation, the Gordon and Betty Moore Foundation, the Kenneth T.and Eileen L. Norris Foundation, the University of Chicago, the Associates of the CaliforniaInstitute of Technology, and the National Science Foundation (NSF). Ongoing CARMAdevelopment and operations are supported by NSF under a cooperative agreement, and bythe CARMA partner universities. L.W.L. acknowledges NSF AST-1139950.Facilities: CARMA, SST(IRAC). REFERENCES
Andre, P., Ward-Thompson, D., & Barsony, M. 1993, Astrophysical Journal, 406, 122Arce, H. G., & Sargent, A. I. 2004, ApJ, 612, 342Arce, H. G., Shepherd, D., Gueth, F., et al. 2007, Protostars and Planets V, 245Artymowicz, P. 1994, ApJ, 423, 581Bachiller, R., & P´erez Guti´errez, M. 1997, The Astrophysical Journal, 487, L93 24 –Bachiller, R., & P´erez Guti´errez, M. 1997, ApJ, 487, L93Bachiller, R., P´erez Guti´errez, M., Kumar, M. S. N., & Tafalla, M. 2001, Astronomy &Astrophysics, 372, 899Bachiller, R., P´erez Guti´errez, M., Kumar, M. S. N., & Tafalla, M. 2001, A&A, 372, 899Blandford, R. D., & Payne, D. G. 1982, MNRAS, 199, 883Caselli, P., Myers, P. C., & Thaddeus, P. 1995, Astrophysical Journal Letters v.455, 455,L77Cerqueira, A. H., Vel´azquez, P. F., Raga, A. C., Vasconcelos, M. J., & de Colle, F. 2006,A&A, 448, 231Chiang, H.-F., Looney, L. W., & Tobin, J. J. 2012, The Astrophysical Journal, 756, 168Chiang, H.-F., Looney, L. W., Tobin, J. J., & Hartmann, L. 2010, The Astrophysical Journal,709, 470Choi, M., Kang, M., & Tatematsu, K. 2011, ApJ, 728, L34Codella, C., Bachiller, R., & Reipurth, B. 1999, A&A, 343, 585Draine, B. T. 2006, The Astrophysical Journal, 636, 1114Dunhill, A., Alexander, R., Nixon, C., & King, A. 2014, arXiv.org, 3842Fern´andez-L´opez, M., Girart, J. M., Curiel, S., et al. 2013, ApJ, 778, 72Ferreira, J. 1997, A&A, 319, 340Ferreira, J., Dougados, C., & Cabrit, S. 2006, A&A, 453, 785Frank, A., Ray, T. P., Cabrit, S., et al. 2014, arXiv.org, 3553Fridlund, C. V. M., & Liseau, R. 1998, ApJ, 499, L75G´omez-Ruiz, A. I., Hirano, N., Leurini, S., & Liu, S. Y. 2013, Astronomy & Astrophysics,558, 94Gueth, F., Guilloteau, S., & Bachiller, R. 1996, A&A, 307, 891Gueth, F., Guilloteau, S., & Bachiller, R. 1998, Astronomy & Astrophysics, 333, 287Gueth, F., Guilloteau, S., & Bachiller, R. 1998, A&A, 333, 287 25 –Gueth, F., Guilloteau, S., Dutrey, A., & Bachiller, R. 1997, Astronomy & Astrophysics, 323,943Guilloteau, S., Dutrey, A., Pety, J., & Gueth, F. 2008, A&A, 478, L31Hildebrand, R. H. 1983, Quarterly Journal of the Royal Astronomical Society, 24, 267Itoh, Y., Kaifu, N., Hayashi, M., et al. 2000, PASJ, 52, 81Konigl, A., & Pudritz, R. E. 2000, Protostars and Planets IV, 759Kwon, W., Looney, L. W., Mundy, L. G., Chiang, H.-F., & Kemball, A. J. 2009, TheAstrophysical Journal, 696, 841Launhardt, R., Pavlyuchenkov, Y., Gueth, F., et al. 2009, A&A, 494, 147Lee, C.-F., Hasegawa, T. I., Hirano, N., et al. 2010, ApJ, 713, 731Li, G., Naoz, S., Kocsis, B., & Loeb, A. 2014, ApJ, 785, 116Lim, J., & Takakuwa, S. 2006, ApJ, 653, 425Looney, L. W., Mundy, L. G., & Welch, W. J. 2000, The Astrophysical Journal, 529, 477Looney, L. W., Tobin, J. J., & Kwon, W. 2007, The Astrophysical Journal, 670, L131Masciadri, E., & Raga, A. C. 2002, ApJ, 568, 733Mathis, J. S., Rumpl, W., & Nordsieck, K. H. 1977, Astrophysical Journal, 217, 425Miettinen, O., Harju, J., Haikala, L. K., & Pomr´en, C. 2006, A&A, 460, 721Montgomery, M. M. 2009, ApJ, 705, 603M¨uller, H. S. P., Thorwirth, S., Roth, D. A., & Winnewisser, G. 2001, Astronomy & Astro-physics, 370, L49Murphy, G. C., Lery, T., O’Sullivan, S., et al. 2008, A&A, 478, 453Nisini, B., Benedettini, M., Codella, C., et al. 2010, Astronomy & Astrophysics, 518, L120Ossenkopf, V., & Henning, T. 1994, Astronomy and Astrophysics (ISSN 0004-6361), 291,943Pety, J., Gueth, F., Guilloteau, S., & Dutrey, A. 2006, A&A, 458, 841 26 –Pyo, T.-S., Hayashi, M., Kobayashi, N., Tokunaga, A. T., & Terada, H. 2009, in AmericanAstronomical Society Meeting Abstracts, Vol. 213, American Astronomical SocietyMeeting Abstracts
This preprint was prepared with the AAS L A TEX macros v5.2.
28 – T a b l e . C A R M A o b s e r v a t i o n s t o w a r d L D a t e s ( U T ) C o nfi g . C a li b r a t o r s ( flu x /ga i n ) P o i n t i n g s Sp ec i e s J u l E - a rr a y M W C , + u t fl o w ( m o s a i c ) mm C o n t ., C O , C O , C N , C H O H , H O J u l E - a rr a y M W C , + u t fl o w ( m o s a i c ) mm C o n t ., C O , C O , C N , C H O H , H O A u g04 D - a rr a y M W C , + e n v e l o p e mm C o n t ., C O , C O , C N , C H O H , H O A u g05 D - a rr a y U r a nu s , + e n v e l o p e mm C o n t ., C O , C O , C N , C H O H , H O O c t C - a rr a y M W C , + e n v e l o p e mm C o n t ., ( C O , C O , N H + , C H O H , S O , S O ) O c t C - a rr a y M W C , + e n v e l o p e mm C o n t ., C O , C O , N H + , C H O H , S O , S O O c t C - a rr a y M W C , + e n v e l o p e mm C o n t ., C O , C O , N H + , C H O H , S O , S O J u l a E - a rr a y M W C , + e n v e l o p e mm C o n t ., C O , C S a P o l a r i m e t r i c d a t a p r e s e n t e d i nS t e ph e n s e t a l. ( )
29 – T a b l e . O b s e r v a t i o n a l r e s u l t s o f m o l ec u l a r li n e s t o w a r d L M o l ec u l a r li n e s F r e q u e n c i e s B e a m s V e l o c i t y R e s . R M S N o t e s G H z (cid:48)(cid:48) × (cid:48)(cid:48) ( P A ◦ ) k m s − J y b m − ( c h − ) mm c o n t i nuu m . . × . ( - ) . C O J = . . × . ( ) . . d a t a m a pp i n g t h e o u t fl o w C O J = . . × . ( - ) . . d a t a t o w a r d t h ee n v e l o p e C O J = . . × . ( - ) . . d e t ec t e d i n t h ee n v e l o p e C S J = . . × . ( - ) . . d e t ec t e dn o r t h w e s t o f e n v e l o p e C NN , J , F = , , , , . . × . ( - ) . . d e t ec t e d i n t h ee n v e l o p e C H O H ( - , ) –7 ( , ) . . × . ( - ) . . n o t d e t ec t e d H O v = ( , ) –6 ( , ) . . × . ( - ) . . n o t d e t ec t e d mm c o n t i nuu m . . × . ( - ) . N H + - F = F = - a . . × . ( - ) . ( . ) . d e t ec t e d i n t h ee n v e l o p e C O J = . . × . ( - ) . . d e t ec t e d i n t h ee n v e l o p e C O J = . . × . ( - ) . ( . ) . d e t ec t e d i n t h ee n v e l o p e C H O H ( , ) –4 ( , ) A ++ . . × . ( - ) . . n o t d e t ec t e d S O N , J = , , . . × . ( - ) . . n o t d e t ec t e d S O ( , ) –2 ( , ) . . × . ( - ) . . n o t d e t ec t e d a N o t e t h a t a ll t h e h y p e r fin e s t r u c t u r e li n e s w e r e o b s e r v e d i n a b a nd a nd t h i s li n e i s u s e d a s t h e r e f e r e n ce f r e q u e n c y f o r t h e v e l o c i t y . T h ec o n t i nuu m f r e q u e n c i e s a r e a r e p r e s e n t a t i v e v a l u e o f a ll c o n t i nuu m d a t a t a k e nb y m u l t i p l e w i d e s p ec t r a l w i nd o w s . T h e v a l u e s i np a r e n t h e s e s a r e v e l o c i t y r e s o l u t i o n s o f t h e o r i g i n a l d a t a , w h i c hh a v e b ee n r e - b i nn e db y . k m s − f o r s t ud i e s i n t h i s p a p e r .
30 –Table 3. Model fitting parameters
Jet 1 Jet 2 s prec +1 +1 v j ( km s − ) a –38 ±
10 / 59 ±
12 –32 ±
12 / 85 ± α ( ◦ ) 11 ± ± T (yr) b ± ± ◦ ) 6 ± ± ◦ ) 26 ± ± φ ( ◦ ) 340 ± ± a Jet velocities for the southern (negative) andnorthern (positive) lobes b Large uncertainties are expected for the precess-ing period T and the jet velocity since these param-eters cannot be constrained separately by the model(see Raga et al. 2009). 31 –Fig. 1.— L1157 maps in continuum and the opacity spectral index β . The contour levels are ±
3, 9, 15, 21, 27, 33, 39, 45, 51, 57, 63, and 69 times 4.5 and 0.5 mJy beam − for λ = 1 . β is in the range of 0.33 to 2.15. The crosses indicate the phasecenter of our observations: [R.A., dec.] (J2000) = [20:39:06.20, +68:02:15.90]. 32 – -2 -1 A m p li t u d e [ J y ] uv distance [kλ] β Fig. 2.— Amplitudes and β along uv distances. We obtained annulus averages of bothmillimeter wavelength data in logarithmic bins and calculated β in the optically thin casedescribed in the text. The error bars indicate the β ranges varying by the statistical standarderrors of annulus mean fluxes at λ = 1 . λ = 1 . ± .
35 in β (Kwon et al. 2009). 33 –Fig. 3.— Integrated intensity maps of various molecular lines toward the L1157 envelope.The blue and red contours represent CO in velocity ranges of –22 to 2 and 4 to 30 km s − ,respectively. All the contour levels are ±
3, 4, 5, 6, 8, 10, 13, 16, 20, and 24 times σ : σ = 2 . O), 0.30 (CS), 0.05 (CN), 0.045 (C O), 0.055 ( CO), and 0.050 (N H + )Jy beam − km s − . The synthesized beams are also marked at the top-right for CO data andthe bottom-right for the others. The cross of each panel is the phase center. The gray scalesindicate λ = 1 . − . The velocity ranges integrated inindividual lines are: 0.83 to 4.47 km s − for C O, 1.38 to 4.17 km s − for CS, 1.25 to 3.25km s − for CN, 1.35 to 3.95 km s − for C O, 0.35 to 5.22 km s − for CO, and 0.59 to3.80 km s − for N H + . 34 –Fig. 4.— Line profiles averaged in a beam at the peak positions of the integrated intensitymaps presented in Figure 3. The three CO isotope profiles have been taken at the center, asthe line peak is consistent with the continuum peak. For N H + the integrated intensities ofa small area at the east and west blobs are shown: a 3 (cid:48)(cid:48) × (cid:48)(cid:48) box for the east and a 2 (cid:48)(cid:48) × (cid:48)(cid:48) box for the west blob. The yellow shaded regions indicate the velocity regions summed upfor the integrated intensity maps. For a better signal-to-noise, the C O and N H + datahave been regridded in a velocity width of 0.1 km s − . The dotted vertical lines indicatethe V LSR of 2.6 km s − . The short solid vertical lines in the middle of the N H + panel markthe 7 hyperfine structures. 35 – −0.20.00.20.40.6 N H + east blob −10 −5 0 5 10Velocity [km/s]−0.20.00.20.40.6 F l u x [ J y / b e a m ] N H + west blob Fig. 5.— N H + line fitting using two Gaussian profiles toward individual east and westblobs. The leftmost component is the J F F = 101 →
012 line. 36 –Fig. 6.— Position-velocity diagrams in C O and C O along the line perpendicular to thebipolar outflow (P.A. = 70 ◦ ) through the phase center. The plus in position of the verticalaxis is eastward. The lowest contours are ± .
18 and ± .
15 Jy beam − , and the intervalsare 0.06 and 0.05 Jy beam − for C O and C O, respectively. The vertical solid line of theC O diagram marks V LSR = 2 . − and the horizontal line indicates the location ofthe continuum peak, which is offset from the phase center by 0 . (cid:48)(cid:48)
1. The green lines of thetwo panels connect the same locations in position and velocity. The blue and orange curvesindicate the Keplerian rotations around a central protostar of 0 . M (cid:12) and 0 . M (cid:12) at adistance of 250 pc, respectively, and the dashed lines are for the cases of a 60 ◦ inclination. 37 –Fig. 7.— Left:
CO CARMA E-array configuration image (black contours) on top of the8 µ m Spitzer image of the L1157 outflow system. The contours are 3, 8, 13, 18, 23, 28, 33 and38 × − km s − (the rms noise level is ∼ − km s − ). The positions ofknots R, R0, R1, R2, B0, B1, and B2 are marked, as well as the central millimeter continuumsource. The synthesized beam is plotted in the bottom right corner. The white solid linemarks the area covered by the half-power primary beams of the 25 pointing mosaic. Topright:
Zoom-in of the central part of the L1157 system. The contours show the combinedCO CARMA E+D-array configuration image on top of the 8 µ m Spitzer image. Contoursare 3, 15, 35, 55, 75 and 85 × − km s − , the rms noise level. The synthesizedbeam is shown in the bottom right corner. Bottom right:
Same as above but showingthe velocity centroid (moment 1) map of the CO emission (color scale) overlaid with theintegrated intensity moment 0 image (black contours). 38 – crossingarches crossingarches
Fig. 8.— CO emission velocity map (contours) overlapped with the integrated emission (greyscale). The contours are 3, 6, 9, 12, 15, 18, 21, 24, 27 and 30 × the rms noise level of eachchannel (0.7, 0.4, 0.3, 0.3, 0.4 and 0.7 Jy beam − km s − , for a-f panels respectively). Thesynthesized beam is shown in the bottom right panel and the velocity interval over whicheach channel was integrated is indicated in each panel. In panels c) and e) the path of somearches of emission are marked with dashed orange lines. 39 –Fig. 9.— Zoom-in of the northern lobe of the L1157 outflow system. The contours show themost redshifted CO emission and the color scale is the Spitzer µ m. The contours are 15,30, 45, 60, 75, and 90% of the peak, 15.4 Jy beam − km s − . The blue lines indicate theballistic trajectories of three possible molecular bullets ejected at different position angles. 40 –Fig. 10.— Cartoon showing the new interpretation for the L1157 outflow system as includingmultiple precessing molecular jets launched at different directions. Dashed lines show thepath of the jets on the rear part of the cones, while solid lines show the path on the frontpart of the cones. The pointed arrows show the axis of symmetry of the two bipolar outflows. 41 –Fig. 11.— Two jet model fit (colored lines) on top of the jet points used in the fits (coloredcircles) and the contours representing the CARMA CO moment 0 image. The colors of thelines and points represent the velocity with respect to the cloud velocity. 42 –Fig. 12.— Left:
Jet 1 fit (colored line) on top of the jet points used in the fit (coloredpoints). The colors represent the velocity with respect to the cloud velocity.
Right:
Theplot shows the distance from the ejecting source position against the velocity of each pointused to make the
Jet 1 fit (black dots). The red line shows the best fit found for this jet. 43 –Fig. 13.— The caption is the same as Figure 12, except now is for the fit of