A Pulsed, Precessing Jet in Cepheus A
aa r X i v : . [ a s t r o - ph . S R ] F e b D RAFT VERSION N OVEMBER
6, 2018
Preprint typeset using L A TEX style emulateapj v. 08/22/09
A PULSED, PRECESSING JET IN CEPHEUS A N ATHANIEL
J. C
UNNINGHAM , N ICKOLAS M OECKEL , J
OHN B ALLY Department of Physics and Astronomy, University of Nebraska-Lincoln,116 Brace Laboratory, Lincoln, NE 68588-0111 Center for Astrophysics and Space Astronomy,University of Colorado, 389 UCB, Boulder, CO 80309-0389 and School of Physics & Astronomy, University of St Andrews, St Andrews KY16 9SS, Scotland
Draft version November 6, 2018
ABSTRACTWe present near-infrared H , radio CO, and thermal infrared observations of the nearby massive star-formingregion Cepheus A (Cep A). From H bow shocks arranged along four distinct jet axes, we infer that the massiveprotostellar source HW2 drives a pulsed, precessing jet that has changed its orientation by about 45 ◦ in roughly10 years. The current HW2 radio jet represents the most recent event in this time series of eruptions. This sce-nario is consistent with the recent discovery of a disk around HW2, perpendicular to the current jet orientation,and with the presence of companions at projected distances comparable to the disk radius. We propose that theCep A system formed by the disk-assisted capture of a sibling star by HW2. We present a numerical model of a15 M ⊙ star with a circumstellar disk, orbited by a companion in an inclined, eccentric orbit. Close passages ofthe companion through or near the disk result in periods of enhanced accretion and mass loss, as well as forcedprecession of the disk and associated orientation changes in the jet. The observations reveal a second powerfuloutflow that emerges from radio source HW3c or HW3d. This flow is associated with blueshifted CO emissionand a faint H bow shock to the east, and with HH 168 to the west. A collision between the flows from HW2and HW3c/d may be responsible for X-ray and radio continuum emission in Cep A West. Subject headings:
ISM: jets and outflows — ISM: Herbig-Haro objects — ISM: kinematics and dynamics —ISM: individual: CepA — stars: formation INTRODUCTION
Collimated, bipolar outflows accompany the birth of youngstars from the earliest stages of star formation to the end oftheir accretion phase (e.g. Reipurth & Bally 2001; Bally et al.2006). The structure and kinematics of these flows provide afossil record of the mass-loss histories of the associated youngstellar objects (YSOs); the most distant shocks trace the old-est major mass-loss events while inner jets and shocks provideclues about the recent processes. The structure and symme-tries of outflows record orientation changes of the underlyingaccretion disk and motion of the outflow sources relative tothe surrounding interstellar medium (ISM). For example, C-shaped outflows may indicate deflection by a side-wind (e.g.,within expanding plasma of the Orion Nebula, Bally et al.2006); other C-shaped bends may result from motion of thesource YSO through the medium, possibly ejected by dynam-ical processes in a young cluster (e.g., HH 498 in NGC 1333,Bally & Reipurth 2001; HH 366 in Barnard 5, Yu et al. 1999).Point-symmetric S- or Z-shaped bends can be produced by theprecession of a circumstellar disk. Examples of this kind ofsymmetry include HH 199 in the L1228 cloud (Bally et al.1995) and 20126+4104 (Su et al. 2007). Such outflow orien-tation changes may provide indirect evidence for a companionstar in an orbit inclined relative to the disk, inducing forcedprecession.In this paper, we present images of shocked outflows in theCep A star-forming region, together with radial velocity mapsof CO emission. We interpret the outflow morphology as ev-idence that the massive protostar HW2 drives a pulsed, pre-cessing jet whose orientation changes may be induced by theperiastron passage of a moderate-mass companion in an ec-
Electronic address: [email protected] address: [email protected] address: [email protected] centric, non-coplanar orbit. We present a numerical modelthat demonstrates the plausibility of jet precession forced bya companion for the case of HW2.
Overview of Cep A
The Cep A star-forming complex contains the second near-est region of massive star formation, after the Orion com-plex. Located at a distance of 725 pc (Blaauw et al. 1959;Crawford & Barnes 1970), the Cepheus OB3 association con-tains a 20 by 60 pc molecular cloud that houses six localizedpeaks of CO emission, designated Cep A through F (Sargent1977, 1979). Cep A contains dense molecular clumps(Torrelles et al. 1993), molecular outflows (Rodríguez et al.1980a; Narayanan & Walker 1996; Gómez et al. 1999), H Oand OH masers (Cohen et al. 1984), hyper-compact H II regions (Hughes & Wouterloot 1984), variable radio con-tinuum sources (Garay et al. 1996), Herbig–Haro objects(Hartigan et al. 1986), bright shock-excited H emission(Hartigan et al. 2000), a cluster of far-infrared (FIR) sourceswith a luminosity of 2 . × L ⊙ (Koppenaal et al. 1979),and a cluster of Class I and Class II YSOs (Gutermuth et al.2005). The bulk of the region’s luminosity likely arisesfrom radio sources HW2 and HW3c/d (Hughes & Wouterloot1984) that are associated with bright H O masers.Cep A contains a massive bipolar molecular outflowaligned primarily east–west (Rodríguez et al. 1980b),but with additional components aligned northeast–southwest (Bally & Lane 1990; Torrelles et al. 1993;Narayanan & Walker 1996; Gómez et al. 1999) . The central2 ′ region contains high velocity (HV) as well as morecompact extremely high velocity (EHV) CO componentswith radial velocities ranging from -50 to 70 km s - relativeto the CO centroid (Narayanan & Walker 1996). The axis ofthe EHV outflow is rotated roughly 40 ◦ clockwise relative Cunningham, Moeckel, & Ballyto the HV outflow on the plane of the sky. The smallerspatial extent together with the higher velocity suggests thatthe EHV flow traces a younger outflow component. Severalself-absorption dips follow trends seen in the low-velocityline wings, with regions east of HW2 blueshifted and west ofHW2 redshifted. Thus cooler, self-absorbing gas traces thelow-velocity bipolar outflow rather than the quasi-stationaryambient medium. At low velocities, there are additionalblue- and redshifted components centered on HW2 that areoriented northeast–southwest.The Cep A outflow complex contains several Herbig–Haroobjects, including the extremely bright HH 168 located about90 ′′ due west of HW2, and several fainter bow shocks lo-cated to the east (Hartigan et al. 2000). Fainter HH objects(HH 169 and 174) are located in the eastern, blueshifted lobe.Near-inrared (NIR) images show an extremely bright reflec-tion nebula centered on HW2 with an illumination cone thatopens toward the northeast. The 2.12 µ m H line exhibits acomplex, filamentary structure.There are multiple luminous sources in the complex of HWradio sources at the core of the Cep A region, led by HW2,and including HW3c, HW3d, and perhaps HW3a, HW8, andHW9. To the west, it has been proposed that radio source W-2may also be internally heated. Because some of these (HW2,HW3d, W-2) have the radio signatures of jets, these are likelycandidates for driving the outflows observed in CO and othermolecules as well as the shocks traced by H , Fe II , and opticalHH emission. The complexity of the outflows, their multipleorientations, ages, and velocities, and the unclear morphologyof some of the shock features make Cep A challenging to in-terpret. Are deflections involved, as suggested by Goetz et al.(1998)? Are there additional outflow sources we have not yetdetected? Are we confusing externally shocked or jet-heatedfeatures with self-luminous sources?HW2 is the strongest radio continuum source in the re-gion with a flux density of 15.8 ± L ≈ L ⊙ , implying a mass of 15–20 M ⊙ .HW2 is located near the center of a small ( < ′′ ), denseclump apparent in the NH maps of Torrelles et al. (1993)at the tip of one of the larger NH concentrations. Multi-frequency observations by Rodríguez et al. (1994) using theVLA show that HW2 is elongated, and that both its spectralindex and size versus frequency relation match those expectedfor a biconical thermal jet. Higher angular resolution obser-vations (Hughes et al. 1995; Hoare & Garrington 1995) showthat HW2 contains a clumpy radio continuum jet whose knotshave proper motions of order 500 km s - (Curiel et al. 2006).Many models of the region postulate that HW2 is the majoror sole source of outflows in the region.The recent interferometric observations of the dust con-tinuum, free–free emission, and several molecular tracerswith arcsecond and subarcsecond angular resolution haveshown that HW2 is surrounded by a complex dust distribu-tion and several close-by protostars. A hot core that mustcontain a moderate-mass YSO is located about 400 AU eastof HW2 (Martín-Pintado et al. 2005). VLA observations re-veal a supposedly low-mass protostar that emits radio wavesand is located at the center of the H O maser arc detectedby Torrelles et al. (2001b). HW2 is surrounded by a cir-cumstellar disk at least several hundred AU in radius andlikely by several low-mass protostars (Rodríguez et al. 2005;Patel et al. 2005; Brogan et al. 2007; Comito et al. 2007;
Table 1
NIC-FPS
OBSERVATIONS OF C EPHEUS
ADate Filter Number of Exposure time (s)exposures each total2005 Dec 11 H -2.12 µ m 10 240 2400K s -2.12 µ m 5 240 1200K s
20 12 2402007 Jan 29 J 15 20 300H 15 20 300K s
15 20 300
Jiménez-Serra et al. 2007; Torrelles et al. 2007).The radio sources Cep A HW3d and HW3b are also associ-ated with a clusters of H O masers. HW3c and/or HW3d aresuspected to be moderate-mass protostars. They are located5 ′′ south of HW2 and close to the western end of a nearlycontinuous chain of radio sources along the southern rim ofthe Cep A East radio source complex. The eastern part of thischain may in part trace a radio continuum jet from HW3c/d atabout P.A. ≈ ◦ . The counter-jet direction points directlyat the radio source W-2 at the eastern end of HH 168 in Cep AWest.Submillimeter wavelength continuum interferometry at 875 µ m (Brogan et al. 2007) shows strong dust emission fromHW2, HW3c, and extended emission to the southwest nearHW3a. No submm continuum was detected from HW3d,leading these authors to conclude that HW3c is most likelyto harbor the second most luminous and massive YSO in theCep A core. In this interpretation, radio source 3d may tracepart of a thermal jet from HW3c. The associated maser emis-sion may be an indicator of shocks in a dense molecular gas.For the rest of this paper, we will assume that the dominantenergy source is embedded in HW3c, but our interpretationdoes not depend critically on which peak, HW3c or 3d, con-tains the source.The extinction to the HW sources is extremely large with A V ≈
500 to 1000 magnitudes. Thus, none of these radiocontinuum sources is visible in the NIR. An extremely brightinfrared (IR) reflection nebula emerges from the vicinity ofHW2 toward P.A. ≈ ◦ . The IR continuum polarization vec-tors indicate that the illumination is coming from the directionof HW2, HW3, and HW8 (Colome & Harvey 1995). How-ever, the accuracy of polarization measurements is insufficientto determine the source unequivocally. OBSERVATIONS
NIR Imaging
We observed Cep A on three nights from 2005 to 2007,using the Near Infrared Camera/Fabry Perot Spectrograph(NIC–FPS) on the Astrophysical Research Consortium 3.5mtelescope at Apache Point Observatory, New Mexico. TheNIR emission in the Cep A region spans 9 ′ east to west, andrequires at least two telescope pointings for complete cover-age with the 4 . ′ × . ′ field-of-view of NIC–FPS. On eachoccasion, the total exposure time was evenly split between awestern field and an eastern field. We obtained broadband im-ages in J, H, and K s filters of the Mauna Kea filter set, and ina narrowband (0.4% bandpass) filter centered on the H µ m line. The dates and details of observations taken on eachnight are listed in Table 1.Images were taken in dithered sets of 5 or 15 frames; indi-vidual frames were reduced by performing dark and sky sub-recessing Jet in Cep A 3 Figure 1.
The 2.12 µ m H emission (green) superimposed on emission from high-velocity CO with - < V LSR < -
18 km s - (blue) and - < V LSR < - (red). In this (and the next two figures) the purple circle shows a beam-sized spot at the location of HW2. Figure 2.
The 2.12 µ m H emission (green), and emission from moderate-velocity CO with - < V LSR < -
13 km s - (blue) and - < V LSR < - - (red). Cunningham, Moeckel, & Ballytraction, flat fielding, and correction for instrumental geomet-ric distortion. Frames in each set were then co-aligned andmedian combined to eliminate bad pixels and to reduce noise.Registration to absolute coordinates was achieved by match-ing locations of many point sources in each combined imageto stellar J2000 coordinates for the region from the 2MASSproject. Once properly registered to this common frame, allthe images taken in a single filter were co-added; each regionof the combined image was scaled to properly account for avarying number of overlapping images and differing exposuretimes.As the H µ m narrowband filter falls within the band-pass of the K s filter, the latter was used as a proxy for the con-tinuum that contaminates the H emission in the narrowbandimage. We removed the continuum component by subtractinga scaled version of the final K s image from the same nightsas the H observations. The scale factor was chosen basedon the best continuum removal as judged by eye in the sub-tracted image. Variations in seeing between the narrowbandand broadband exposures result in the incomplete removal ofsome stars. Optical spectroscopy
Visual wavelength long-slit spectra of the shocks locatedeast and northeast of Cep A HW2 were obtained with theDouble Imaging Spectrograph (DIS) at the f/10 Nasmyth fo-cus of the 3.5 meter reflector at the Apache Point Observatoryon 2007 January 9. A 1.5 ′′ wide by 5 ′ long slit was used witha grating providing a spectral resolution R = λ/ ∆ λ ≈ Radio observations
The Nobeyama Radio Observatory (NRO) 45 m radio tele-scope was used in February 1987 to map the distribution ofJ=1–0 CO from the Cep A outflow complex with 15 ′′ res-olution at 115 GHz. Nobeyama spectra were collected on auniform grid with a 10 ′′ spacing over a 5 by 9 arcmin re-gion containing the Cep A outflow complex. Over 1900 in-dividual positions were observed for an average integrationtime of about 1 min per position, with 20 seconds on sourceand 20 seconds off source plus 10 seconds of dead time usedfor reading the spectrometer, and moving the antenna aftereach 20 second integration period. Approximately 14 sepa-rate 10-hr periods were devoted to observing Cep A. A cooledSchottky receiver with a single sideband receiver temperature T SSB ≈ T SSB ≈ ′ to the northwest.The data were reduced using the COMB package developedby R. W. Wilson at AT&T Bell Laboratories. Thermal IR Imaging
Observations of the Cep A region were made using theKeck Observatory facility mid-IR camera Long Wave Spec-trometer (LWS) on 2002 November 16 November. LWS isa mid-IR imaging and spectroscopy instrument mounted onthe forward Cassegrain focus of Keck I, employing a Boe-ing 128 ×
128 As:Si BIB array with a 10 . ′′ × . ′′ field ofview (Jones & Puetter 1993). Weather conditions were poor;the thermal background and atmospheric transmission variedby 50% throughout the first half of the night. Approximately10 individual observations (frames) were obtained at slightlyoverlapping positions using the 12.5 µ m filter (12 to 13 µ mbandpass with >
80% transmission). The chopping secondarymirror was driven at 2 Hz with a 30 ′′ east–west throw. Eachframe was observed using the standard mid-IR chop-nod tech-nique with two chopping positions plus (on-source) and mi-nus (off-source); after chopping with the source in chop-beamplus, the telescope was nodded along the chop axis so that theobject would sit in chop-beam minus, and chopping wouldcontinue. Each frame was observed for one complete chop-nod cycle, yielding a total on-source integration time of 27.6s per mosaic frame. The standard stars β Peg, β And, and β Gem were observed for points-spread function (PSF) determi-nation and flux calibration. The images are nearly diffractionlimited at 12.5 µ m with PSF FWHM = 0.38 ′′ and a Strehl ra-tio of 35%. Data reduction details are given in Shuping et al.(2004). The image coordinates are accurate to about 1 ′′ . Observational results and interpretation
The CO Outflow Complex
NRO single-dish observations of accelerated CO emissionprovide constraints on the past evolution of the Cep A out-flow complex. While radio jets and HH objects near theirsources tend to trace the highest velocity and recently ejectedoutflow components with velocities higher than a few hun-dred km s - , HH objects located far from their sources andshock-excited H emission tend to trace moderate speeds of10 to about 200 km s - . CO and other molecules trace the gasswept up by secondary interactions with the ambient mediumand have velocities of a few to tens of km s - . Thus, CO andsimilar tracers tend to act as calorimeters of the total amountof momentum injected into the parent cloud by the faster flowcomponents.The NRO CO data (Figures 1, 2, and 3) show a prominentoutflow complex with complicated structure. In general, theCep A CO emission is collimated along an east–west direc-tion at the highest velocities, while at the lowest velocities,the CO emission is very poorly collimated and bipolar along anortheast–southwest direction. At velocities more than about8 km s - away from the rest velocity of the Cep A core ( - - relative to the local standard of rest), the strongestoutflow component consists of an east–west bipolar flow thatis blueshifted toward the east (blue component in Figure 1)and redshifted toward the west. Additionally, an anomalousblueshifted lobe appears about 6 ′ west by northwest of HW2(upper-right corner of Figure 1). A fainter bipolar compo-nent, blueshifted toward the northeast and redshifted towardthe southwest, emerges from the HW2 region at P.A. ≈ ◦ ;this is most apparent at moderate velocities (diagonal bipolarfeature in Figure 2). At low velocities of about 2–5 km s - with respect to the centroid velocity, a prominent blueshiftedlobe emerges from near HW3c (located about 5 ′′ south ofHW2) and extends about 2 ′ east at P.A. ≈ ◦ . Below, itis argued that this feature traces an outflow from HW3c.recessing Jet in Cep A 5 Figure 3.
The 2.12 µ m H emission (green), and emission from low-velocity CO with - < V LSR < -
13 km s - (blue) and - < V LSR < - - (red). Rather than tracing two independent flows, the brightereast–west outflow from HW2 and the secondary, weaker out-flow component oriented northeast–southwest at P.A. ≈ ◦ may trace the walls of a large-scale bipolar cavity that openstoward the northeast and southwest. In this interpretation, thenortheast cavity walls are at P.A. ≈ ◦ and 40 ◦ and thesouthwest cavity walls are at P.A. ≈ ◦ and 200 ◦ . The ori-entation of the IR reflection nebula emerging from the HW2region (Figure 5) provides support for this scenario; its axislies at P.A. ≈ ◦ , which is aligned with the opening of thenortheast cavity. The Core in the Mid-IR
Figure 4 shows the Keck 12.5 µ m image of the Cep A core.The brightest feature is the double-lobed nebulosity surround-ing the position of HW2. The two peaks of mid-IR emissionare separated by about 1.4 ′′ along an axis oriented at P.A.= 45 ◦ , which is similar to the orientation of the HW2 radiocontinuum jet. The dark lane that separates the two emis-sion peaks has an orientation similar to the circumstellar disksurrounding HW2. However, it is unclear whether the darklane is an IR disk shadow produced by a more compact disksurrounding HW2, or the gap is actually the disk seen in sil-houette. Multicolor imaging is needed to determine the natureof the gap. No prominent mid-IR sources are seen at the loca-tions of radio sources HW3c and d. However, an extended IRnebula is visible about 2 ′′ west of HW3c. This feature mightbe illuminated by either HW2, 3c, or 3d. IR polarization mea-surements are needed to distinguish these possibilities.A faint point source is located at J(2000) = 22:56:18.2,+62:01:44. Although there are no NIR sources at this loca-tion, there is an H O maser spot (H O-E) located 0.7 ′′ from the nominal mid-IR source position. Radial Velocities of Optical HH Objects
In contrast to the spectacular 300 km s - Doppler shiftsseen toward HH 168 in Cep A West, the long-slit spectra ofthe faint HH objects HH 169 and 174 located east and north-east of HW2 show only small Doppler shifts not larger thanabout 60 km s - in the core of the line profile. All compo-nents to the east are blueshifted. The radio proper motionsof the continuum jet emerging from HW2 indicate veloci-ties of order 500 to 300 km s - within a few seconds of thesource. Thus, it may seem surprising that shocks located atdistances of 1–5 ′ away show much lower velocities. The lowvelocities either indicate that the flow east of HW2 is mostlyalong the plane of the sky, or that the ejecta have been de-celerated significantly. Future proper motion measurementsare needed to distinguish these possibilities. However, mostHH objects exhibit rapid decreases in their velocities with in-creasing distance from their sources (Reipurth & Bally 2001;Devine et al. 1997). Such deceleration is probably caused bythe interaction of the ejecta with the ambient medium. Thestructure of the Cep A cloud is such that the outflows emerg-ing from the Cep A core propagate relatively freely toward thewest, but impact dense molecular gas toward the east. Thus,it is not surprising that HH 168 has high velocities while theeastern shocks associated with HH 169 and 174 have low ve-locities. For the rest of the paper, we assume that the spacevelocities of the ejecta east of the Cep A core have a speed ofabout 100 km s - , typical of HH objects located at comparabledistance from their sources. The H Outflow Complex
Cunningham, Moeckel, & Bally
Figure 4.
The Cep A core at 12.5 µ m observed with LWS on the Keck 10 m telescope. The locations of the three suspected most massive protostars are shownin black circles. HW2 is the unlabeled circle. Figure 5 shows a three-color composite image built fromthe J, H, and K s broadband filter images. The continuum-subtracted narrowband image of the H emission in the CepA complex is displayed in Figure 6, and is shown in rela-tion to the CO emission in Figures 1–3. The narrowband im-age exhibits three main components: shock emission east ofthe central protostellar cores containing HW2 and HW3c/d,bright arcs of shock-excited emission located to the west ofthese cores that is associated with HH 168, and a faint rosetteof emission to the northwest. The rosette is associated with aneast-facing globule or small pillar, located about 1.5 ′ north ofHH 168, which exhibits slightly blueshifted CO emission inFigures 2 and 3. The H emission is likely to be excited by UVradiation from the Cep OB3 association and is thus probablyfluorescent in nature. The globule contains an m B = 16 mag-nitude star that is also called Cep A IRS2. Because there is noredshifted emission in this region, the blueshifted CO emis-sion is unlikely to be a tracer of an outflow. The CO velocityfield in the globule may reflect random motions within theCep A cloud, or be the result of UV-photo-heating-induced ablation of the globule surface that can accelerate gas to a ve-locity of a few km s - . Bipolar Outflow from HW3c: HH 168 and its Counterflow
The series of bright arcs and H bow shocks located 1–3 ′ west of HW2 are associated with HH 168. The axes ofsymmetry of most of these shocks indicate a point of ori-gin about 10 ′′ south of HW2. The submillimeter and radiocontinuum source HW3c and the bright H O maser and ra-dio source HW3d lie on or near this axis. A dim, 40 ′′ di-ameter bow shock is located on this axis about 2 ′ east ofHW3c, directly opposite HH 168. This feature and the visual-wavelength part of HH 168 are symmetrically placed aboutHW3c. Thus, HW3c is likely to be the source of these shocksand the associated CO outflow components. High-resolutioncm-wavelength VLA observations (Garay et al. 1996) reveala chain of radio sources approximately aligned with the axisof this flow; these may either trace an ionized jet, or the inter-face between a jet and surrounding dense cloud material.Figures 2 and 3 show that the bipolar outflow from HW3crecessing Jet in Cep A 7 Figure 5.
Color-combined broadband image of Cep A. J-band emission is shown in blue, H in green, and K s in red. Figure 6.
Continuum-subtracted 2.12 µ m H emission in the Cep A outflow complex. The successive orientations of a suspected precessing jet from HW2 areindicated by red arrows; the position of HW2 is marked with a cross. The oldest (eastward) ejection appears to power HH 174 to the east. The next two ejectionsmay be responsible for HH 169. The current orientation of the radio jet emerging from HW2 is at position angle (P.A.) = 45 ◦ , indicated by the dashed magentaline. Bright HH 168, and the H bows at far right, may result from an outflow from HW3c or HW3d (also marked with crosses) along the green axis. A faint H bow marks the opposite lobe of this flow. is associated with low-velocity to intermediate-velocity red-and blueshifted gas. The bright jet-like H feature located 30 ′′ east of HW3c is associated with the strongest blueshifted COemission. However, about 1 ′ east, the blueshifted CO lobeshifts about 15 ′′ south of the flow axis and terminates justbelow (south) of the H bow shock. Evidently, the HW3c flowinteracts strongly with ambient cloud material along its southwall. However, there appears to be less interaction along thenorth rim of this cavity, perhaps because ambient material inthis region has been cleared away by the HW2 flow.The western lobe of this flow is associated with redshiftedCO co-located with HH 168. Three bow shocks bright in H lie on the southern side of HH 168; their orientations are con-sistent with a flow powered by HW3c. The bow shock mor-phologies and low CO velocities indicate that this flow liesclose to the plane of the sky. It is somewhat unusual that atvisual and NIR wavelengths, the redshifted outflow lobe isbrighter than the blueshifted lobe. This feature may be a con-sequence of the morphology of the Cep A cloud core that ap-pears to be mostly located east of the region containing HW2and HW3c. The large range of radial velocities observed inHH 168 in visual-wavelength spectra may indicate sidewayssplashing of the high velocity ejecta moving mostly in theplane of the sky that has encountered slower-moving mate- Cunningham, Moeckel, & Ballyrial. Non-detection of Source W-2 in HH 168 in the IR
Radio source W is located at the eastern end of the HH 168shock complex in Cep A West. The radio emission fromthis object is extended with an elongation similar to that ofthe shocked H emission. Based on radio spectral indices,Garay et al. (1996) infer that subcomponent W-2 houses anultracompact internal source, shielded from view in the IR bythe surrounding molecular cloud, and that the elongated radioemission is likely due to a jet launched by this source. Ourimages exhibit co-located, intense NIR H emission that co-incides with bright optical HH features at the eastern end ofHH 168, making it unlikely that this region is highly obscured.Yet the broadband NIR images do not reveal any point sourcesin this region; thus it seems unlikely that a stellar source ex-ists at this location. Rather, radio source W may trace a hardshock, with speeds above 400 km s - . The detection of softX-ray emission from this region (Pravdo & Tsuboi 2005) pro-vides support for this interpretation. Source W is locatedwhere two outflows, one emerging from HW2 and anotherfrom HW3c, intersect on the plane of the sky. Thus it is pos-sible that these flows are actually colliding. We suggest thatthe radio and X-ray emission from HW may arise from col-liding flows emerging from the two most massive protostarsin Cep A East. This interpretation is generally in line withthe westward proper motions measured for this radio source(Rodríguez et al. 2005). A Pulsed, Precessing Outflow from HW2
The blueshifted, eastern lobe of the Cep A outflow containsfour distinct chains of H emitting features, each of whichterminates in a well-formed bow shock. The four chains ap-pear to emerge from the immediate vicinity of HW2. Theiraxes, defined by lines connecting HW2 to the bow shocks atthe eastern and north-eastern ends of the chains, shift sys-tematically clockwise from nearly east–west to northeast–southwest. The longest chain, which terminates 4.8 ′ East ofHW2 at HH 174, has a position angle P.A. ≈ ◦ . The sec-ond chain, which terminates about 3.4 ′ from HW2 at the east-ern component of HH 169, has P.A. ≈ ◦ . The third chainterminates at the western component of HH 169, about 2.2 ′ northeast of HW2 and has P.A. ≈ ◦ . The fourth chain endsin a bright but compact H bow at P.A. ≈ ◦ about 1 ′ fromHW2. The current orientation of the HW2 radio jet contin-ues this clockwise migration of outflow orientations and hasP.A. ≈ ◦ . The chains of H knots get progressively shorteras the axes rotate clockwise toward decreasing P.A.. This re-markable progression may be an indication that HW2 pow-ers a pulsed and precessing jet. Between each major outflowejection event, the jet orientation changes by about 10 ◦ –15 ◦ as seen in projection on the plane of the sky.A rough dynamical age for each chain can be estimated bydividing the length of each by the velocity of its tip. The ra-dial velocities of the HH objects located at the ends of the firstthree chains are low ( v <
60 km s - ), indicating that they aremost likely moving close to the plane of the sky. Fabry–Perotimaging of the H emission (Hiriart et al. 2004) also indicateslow, but chaotic, radial velocities. On the other hand, the ex-citation of the visual wavelength H α and S II emission of theeastern HH objects and comparison with the speeds of otherHH objects located at similar distances from their sources,indicate that shock speeds of order 100 km s - are not unrea-sonable. Future proper motion measurements are needed to determine the velocities of the ejecta. Assuming 100 km s - (reasonable for similar shocks exhibiting NIR H emission,e.g. Reipurth & Bally 2001), the dynamical ages of the foureastern H chains are 9900, 7000, 4500, and 2100 years, re-spectively, indicating that an eruption/ejection event occursapproximately every 2500 years. Furthermore, the presenceof the HW2 radio jet indicates that there is currently an erup-tion underway.This periodic, geometric progression suggests an underly-ing temporal sequence. We propose a model for Cep A Eastin which the accretion disk surrounding HW2, responsible forlaunching a bipolar jet along the disk rotation axis, has period-ically changed its orientation. In this picture, the easternmostlobe (position angle 90 ◦ ) constitutes the oldest tracer of pastjet activity. Since its launch, the accretion disk and associatedbipolar jet at HW2 have been periodically torqued through aseries of new orientations.Periastron passages of a companion in an eccentric orbitthat is not coplanar with the disk can readily explain thetorques required to explain both the impulsive nature of theHW2 outflow, and the periodic changes in disk orientation. Acluster member in an elliptical orbit around HW2, with a peri-astron distance of order the disk radius, will exchange angularmomentum with the disk. If the binary and disk axes are notaligned, the result is a progressive tilting of the disk at eachperiastron passage. The close passage of the orbiting compan-ion will also disturb the disk during periastron passage, trig-gering an increase in the accretion rate onto the central star,and consequent mass loss in the form of a collimated jet. Theperiodic reorientation of the jet axis and disk, and the episodiccharacter of the outflow, may be the signatures of a crowdedand dynamically active environment surrounding HW2.The angular extent of the longest shock lobe (position angle90 ◦ ) is 4.8 arcmin, yielding a projected length of about 1 pcgiven the adopted distance of 725 pc (Blaauw et al. 1959). Asestimated above, the time between disk reorientations (and inthis model, the orbital period) is ≈ M ⊙ (Patel et al. 2005, and referencestherein), so a less massive companion must have an orbitalsemimajor axis of ≈
400 AU. The periastron distance must besomewhat smaller than this value. As indicated in the Intro-duction, Cep A HW2 has at least two companions within therequired distance, including the “hot core" that may be heatedby a moderate-mass protostar (Martín-Pintado et al. 2005)and the source of the expanding maser ring (Torrelles et al.2001b; Curiel et al. 2002). NUMERICAL MODELING
The interaction of a binary companion and a circumstellardisk can cause the disk to precess (e.g. Papaloizou & Terquem1995; Terquem et al. 1999). This precession may be traced bythe orientation of the jet, and has been implicated in the ap-pearance of several precessing outflows (e.g. Eisloffel et al.1996; Davis et al. 1997). During repeated encounters of acaptured companion on an eccentric orbit, the disk orientationmoves through several angles impulsively (Moeckel & Bally2006), rather than smoothly varying as would occur in anearly circular binary. Inspired by the observations presentedabove, we searched the parameter space of encounters simu-lated in Moeckel & Bally (2007a) and selected three combi-nations of parameters for further investigation.
Method and initial conditions recessing Jet in Cep A 9
Table 2 S IMULATION PARAMETERS . M primary M disk r disk M impactor i r peri M ⊙ M ⊙ AU M ⊙ degrees AU15 1.5 350 5 30 15015 1.5 350 5 45 8015 1.5 350 5 135 80 We used a modified version of the SPH/ N -body codeGADGET-2 (Springel 2005) to model interactions between amassive protostar and a captured companion. The models andthe method are similar to those of Moeckel & Bally (2006),and we summarize them here. A Keplerian disk is set up withsurface density Σ ( r ) ∝ r - , and temperature T ( r ) ∝ r - / . Thisdisk is allowed to evolve in isolation until the system has re-laxed from its initial conditions. At this time an impactor staris introduced on a slightly hyperbolic orbit ∼ M ⊙ star with a 2 M ⊙ , 500 AU diskand impactors of varying masses, periastron, and inclinationangle. By inspecting these simulations for disk orientationchanges of roughly 10 ◦ , we selected three impactor parame-ters to study the HW2 system, shown in Table 2. We scaledthe primary and disk masses down to values perhaps moresuitable for HW2, a 15 M ⊙ primary with 350 AU, 1.5 M ⊙ disk. The mass of the impactor in all cases is 5 M ⊙ ; afairly massive companion is needed to torque the disk through ∼ ◦ during a periastron passage. We modeled the disk us-ing ∼ . × particles, and followed each system throughthree encounters. Numerical results
We consider the orientation of the inner disk ˆ L disk to be de-fined by the summed angular momenta of the gas particleswithin 30 AU of the primary. In order to determine what theoutflow appearance would be, we construct a crude jet modelas follows. We assume that the orientation of the outflow isthe same as the inner disk orientation, ˆ V jet ≡ ˆ L disk . By assum-ing a constant-velocity jet launched in this direction, we canconstruct a model of what the jet from this interacting systemwill look like. Of the three cases considered, the two pro-grade simulations ( i = 30 ◦ and i = 45 ◦ ) produced jets that are similar in appearance to the observed system. The retrogradeencounter, i = 135 ◦ , showed a shift in orientation similar to theother two for the first passage, but the subsequent passages didnot resemble the actual jet. We believe this is in part due tothe rapid disk destruction associated with retrograde passageshaving small r peri / r disk (Moeckel & Bally 2006).Plots of the jets through the first three passages from thetwo prograde simulations are shown in Figure 7. They havebeen rotated about the initial jet axis in order to most closelymatch the observed appearance of the HW2 system, and theinitial jet axis lies on the x-axis to match the orientation of thejet lobes in Figure 6. The simulations yield the elapsed timeof the encounters, and at an assumed distance of 725 pc, thedistance to the oldest bowshock is approximately 1 pc. Bychoosing the velocity of the jet to match this distance, we findjet velocities of 88 km s - for the 30 ◦ case, and 139 km s - for the 45 ◦ case. The boxes on the figure indicate where bow-shocks would be expected, and are labeled with the age of theflow at that point. While both simulations capture the qual-itative orientation of the jet, the case with i = 45 ◦ appears tomatch the observed jet orientation slightly better.We stress that we are not claiming that the parameters stud-ied here reflect the reality of the HW2 region’s dynamics. Theuncertainties in the masses, radii, and even number of objectsimpede attempts to model the actual system, and the rangeof parameters that produce similar outflows is large. Rather,these simulations should be viewed as a proof-of-concept ex-periment. Using plausible values for the masses of the com-ponents, we can generate outflow orientations similar to thoseobserved, via a series of postcapture interactions between amassive protostellar system and a moderate-mass companion. DISCUSSION
The two most massive protostellar objects in Cep A, HW2and HW3c, each power separate outflows. HW2 appears tohave been responsible for at least four quasi-periodic ejec-tions that produced collimated flows visible in NIR H emis-sion in the eastern lobe of the Cep A outflow complex. Whilethe first eruption launched a flow nearly due east, subsequentejections indicate that the source orientation rotated north byabout 10 ◦ between events. The current orientation of the radiojet, rotated clockwise by 45 ◦ compared to the ejection axis ofthe first event, continues this trend. Radial velocity measure-ments of the visual wavelength emission associated with HH169 and 174 located at the tips of three of the outflow axes in-dicate low radial velocities and dispersions, an indication thatthe outflow axis probably lies close to the plane of the sky.The bright NIR reflection nebula emerging from the vicinityof HW2 has an axis close to the last two ejection events. Mil-limeter wavelength CO emission toward the eastern lobe ofthe Cep A outflow complex reveals two ridges of blueshiftedemission at P.A. ∼ ◦ and 35 ◦ , approximately symmetri-cally placed about the axis of the reflection nebula. A pair ofredshifted ridges of CO emission are located on the oppositeside of HW2 toward the southwest. These CO features maydefine the walls of an outflow cavity created by the outflowover time. The CO outflow from HW3c is superimposed onthe east–west rim of this suspected cavity.The data presented above are interpreted in terms of apulsed, precessing jet emerging from Cep A HW2. A numer-ical model is used to show that a pulsed, precessing jet can bethe result of disk perturbations produced by a companion starin an eccentric orbit that is not co-planar with the disk. Thisscenario is the most likely outcome of the capture-formed bi-0 Cunningham, Moeckel, & Bally Figure 7.
Locations of jet ejecta from two simulations, each as viewed at a single time. The solid lines mark projected distance from outflow source, witholder material at far left and recent ejecta at far right. Dotted lines indicate successive jet orientations; in the absence of disk precession, the jet would remainhorizontal. The jet velocities have been scaled to give a 1 pc distance to the oldest shock, and the initial jet orientation has been chosen for comparison to Figure6. Boxes indicate where shocks at the ends of the outflow lobes would be expected, and are labeled with the age of the flow at that point. nary proposed by (Moeckel & Bally 2007a,b). We note thatNarayanan & Walker (1996) also suggested periodic outflowfrom Cep A, although the dynamical timescale they deter-mined for the launch of successive CO shells is ∼ . × yr,two orders of magnitude greater than the ∼ gas presentedhere.The second outflow from the Cep A core region appears toemerge from slightly south of HW2. This flow is associatedwith a low-velocity blueshifted lobe of CO emission that ter-minates about 1.5 to 2 ′ to the east in a faint but large H bowshock. HW3d is associated with strong H O masers that arelocated at the base of a chain of radio sources at position angleP.A. ≈ ◦ (Torrelles et al. 1998, 2001a) that has the spectralindex and morphology of a thermal radio jet that has the sameorientation as the suspected large-scale outflow from HW3c.A third young star located southwest of HW2 may alsopower an outflow that produced the radio sources HW1b,3a, and 3b. HW3b is the third source in Cep A to con-tain masers (Torrelles et al. 1998, 2001a). The submillimetersource SMA4 (Brogan et al. 2007) is located on the axis ofthis chain close to HW3b. The chain of radio sources has aposition angle P.A. ≈ ◦ (or 240 to 245 ◦ ), close to theorientation of the H jet associated with the third suspectederuption of HW2. Thus, an alternative to our model is thatthis collimated component east of the Cep A core is a steadyjet associated with a protostar located in the HW1b, 3a, and 3bchain. If SMA4 is the source, then the chain of radio sourcesis the counter-jet associated with the prominent H feature.However, this model does not explain the orderly progressionof terminal bow-shock orientations and distances discussedabove. Furthermore, the chain of radio sources HW1b, 3a, and 3b lies in the "line-of-fire" of the current HW2 jet. Thus,it is possible to interpret these features as shocks where theHW2 jet rams part of the Cep A molecular cloud, or the out-flow from HW3c.No NIR continuum source is detected at the location of ra-dio source W at the eastern end of HH 168. The presenceof radio continuum and X-ray emission from this portion ofHH 168 may indicate that this region is being impacted byshocks with speeds of at least 400 km s - . We propose thatHH 168 marks the location where the HW3c outflow collideswith the outflow from HW2. The location of HH 168 indi-cates that the currently shocked gas is mostly excited by thedebris from the third and fourth eruptions of HW2 with theflow form HW3c. Ejecta from the fifth eruption, which istraced by the HW2 radio continuum jet, has not yet reachedthe interaction region. If HW3b also drives its own jet withan orientation similar to the third eruption from HW2, it mayalso contribute to the excitation of HH 168.The complex shock morphology may be the result of tur-bulent mixing of the colliding flows. The north-to-south den-sity gradient in the ambient molecular cloud may contributeto the bending of this shock complex toward the north. TheCO emission, which is both red- and blueshifted at the loca-tion of HH 168, may also be deflected along our line of sightby this gradient to produce the blueshifted CO lobe fartherdownstream.Collisions between outflows and clouds have been in-voked to explain the complex structure of some Herbig–Haro objects such as HH 110 in Orion where the HH 270jet slams into a small molecular cloud in the L1617 com-plex (López et al. 2005). An embedded IR source in thiscloud, IRAS 05487+0255, drives a jet and molecular out-recessing Jet in Cep A 11flow (Reipurth & Olberg 1991; Lee et al. 2000). Kajdic etal. (2009, in preparation) propose that HH 110 is formedby the collision of the HH 270 flow with the flow fromIRAS 05487+0255. Theoretical models of jet–cloud interac-tions were presented by Canto & Raga (1996); Raga & Canto(1996); de Gouveia Dal Pino (1999); Raga et al. (2002). Jet–jet collisions were modeled by Cunningham et al. (2006). InCep A, HH 168 may be excited by the collision between apair of outflows and be additionally deflected by the densitygradient in the cloud. Observations show that the sources ofthe HW2 and 3c flows are physically close. Furthermore, bothsources appear to drive flows that are close to the plane of thesky. These factors make a collision of flows reasonably prob-able. The absence of obvious shocks southwest of the CepA core supports the view that ejecta from HW2 are deflected.Still, the presence of low-velocity CO emission southwest ofHW2 and HW3c indicates that this part of the cloud has beenimpacted by outflow activity. A possible explanation for thisdichotomy is that, in the distant past, only the outflow fromHW2 was active, and that the deflection of this flow by theflow from HW3c is a relatively recent phenomenon. CONCLUSIONS
We present new narrowband NIR images of the the Cep Aoutflow complex in the 2.12 µ m S(1) line of H . These imagesprovide evidence for two major outflows from the two mostluminous members of the Cep A outflow complex.The radio source HW3c appears to drive a collimated east–west outflow with a redshifted lobe emerging toward P.A. ≈ ◦ that may be responsible for much of the emission associ-ated with HH 168 in Cep A West. The blueshifted lobe of thisoutflow appears to terminate in a large but faint H bow shocklocated about 1.5–2 ′ east of HW3c. The H bow and the thebrightest portion of HH 168 are placed symmetrically aboutHW3c. The three H bow shocks located along the south-ern portion of HH 168 have orientations consistent with beingdriven by this source.The radio source HW2, the most luminous and massive pro-tostar in the Cep A complex, appears to drive a precessing andpulsed jet. The H images reveal four distinct jet axes. Thelongest and oldest outflow lobe terminates in HH 174 about5 ′ due east of HW2. The second and third lobes are pro-gressively shorter, have rotated clockwise by about 10–15 ◦ between ejections, and terminate in the western and easterncomponents of HH 169. The fourth ejection is the shortestlobe and was ejected toward the northeast, but has no visualwavelength HH objects associated with it, probably due to thehigh obscuration. The radio continuum jet may trace a cur-rent period of activity and continues the trend of clockwiserotation of the outflow axes. Thus, over the past 10 years,the HW2 outflow appears to have precessed by nearly 45 ◦ ina clockwise direction as seen on the plane of the sky. Assum- ing an average flow speed of 100 km s - , HW2 undergoes aneruption about every 2500 years.Most massive stars are born in clusters, and Cep A is noexception. In an environment with a high stellar density, in-teractions between a massive star and its siblings become rel-atively common. Cep A HW2 is surrounded by a disk atleast several hundred AU in radius and with a mass of order1 M ⊙ . This disk can serve as a dissipative environment, andany star passing within about a disk radius has a high prob-ability of being captured into a highly eccentric orbit. Themore massive the intruder, the higher its probability of cap-ture (Moeckel & Bally 2007a,b). This mechanism provides anatural explanation for the presence of a moderate-mass com-panion, indicated in the HW2 system by a hot core 0.6 ′′ (400AU in projection) east of HW2 (Martín-Pintado et al. 2005).It is proposed that Cep A HW2 has a moderate-mass com-panion in an eccentric orbit whose orbital plane is inclinedwith respect to the circumstellar disk surrounding HW2. Pe-riastron passages of the companion may perturb the disk,and drive accretion onto the central star, and produce quasi-periodic episodes of collimated mass ejection. The non-coplanar eccentric orbit also applies a torque to the disk,changing its orientation. We investigate this hypothesis us-ing an SPH code and demonstrate that such interactions canresult in the type of disk orientation change proposed to haveoccurred in Cep A. Modeling indicates that a companion ina prograde orbit inclined with respect to the HW2 disk byabout 40 ◦ fits the observations best. Because an impactormass greater than the disk mass is needed to torque the diskthrough an appreciable angle, the presence of one or morelow-mass sources in the vicinity do not affect this model.The bright shock complex HH 168 may be excited by thecollision of two or more outflows emerging from the Cep Acloud core located about 90 ′′ to the east. The western end ofHH 168 may trace the chaotic mixing resulting from the col-lision of the steady (in orientation) flow from HW3c and thethird eruption of HW2; the eastern end of HH 168 may tracethe collision of the flow from HW3c with the fourth eruptionof HW2. If SMA4 also drives a flow, it may also participatein this interaction.This work was supported by NSF grant AST0407356 andthe CU Center for Astrobiology funded by NASA under co-operative agreement no. NNA04CC11A issued by the Officeof Space Science. Some of the observations presented wereobtained with the Apache Point Observatory 3.5-meter tele-scope, which is owned and operated by the Astrophysical Re-search Consortium. We thank Drs. Mark Morris and RalphShuping for assistance with the acquisition and reduction ofthe Keck data. REFERENCESBally, J., Devine, D., Fesen, R. A., & Lane, A. P. 1995, ApJ, 454, 345Bally, J., & Lane, A. P. 1990, in ASP Conf. Ser. 14: Astrophysics withInfrared Arrays, ed. R. Elston (San Francisco: ASP), 273Bally, J., Licht, D., Smith, N., & Walawender, J. 2006, AJ, 131, 473Bally, J., & Reipurth, B. 2001, ApJ, 546, 299Blaauw, A., Hiltner, W. A., & Johnson, H. L. 1959, ApJ, 130, 69Brogan, C. L., Chandler, C. J., Hunter, T. R., Shirley, Y. L., & Sarma, A. P.2007, ApJ, 660, L133Canto, J., & Raga, A. C. 1996, MNRAS, 280, 559Cohen, R. J., Rowland, P. R., & Blair, M. M. 1984, MNRAS, 210, 425Colome, C., & Harvey, P. M. 1995, ApJ, 449, 656 Comito, C., Schilke, P., Endesfelder, U., Jiménez-Serra, I., &Martín-Pintado, J. 2007, A&A, 469, 207Crawford, D. L., & Barnes, J. V. 1970, AJ, 75, 952Cunningham, A. J., Frank, A., & Blackman, E. G. 2006, ApJ, 646, 1059Curiel, S., Ho, P. T. P., Patel, N. A., Torrelles, J. M., Rodríguez, L. F.,Trinidad, M. A., Cantó, J., Hernández, L., Gómez, J. F., Garay, G., &Anglada, G. 2006, ApJ, 638, 878Curiel, S., Trinidad, M. A., Cantó, J., Rodríguez, L. F., Torrelles, J. M., Ho,P. T. P., Patel, N. A., Greenhill, L., Gómez, J. F., Garay, G., Hernández,L., Contreras, M. E., & Anglada, G. 2002, ApJ, 564, L35Davis, C. J., Eisloeffel, J., Ray, T. P., & Jenness, T. 1997, A&A, 324, 10132 Cunningham, Moeckel, & Bally