The Magnetic Field in the Class 0 Protostellar Disk of L1527
Dominique M. Segura-Cox, Leslie W. Looney, Ian W. Stephens, Manuel Fernandez-Lopez, Woojin Kwon, John J. Tobin, Zhi-Yun Li, Richard Crutcher
aa r X i v : . [ a s t r o - ph . S R ] D ec Accepted by ApJL
Preprint typeset using L A TEX style emulateapj v. 5/2/11
THE MAGNETIC FIELD IN THE CLASS 0 PROTOSTELLAR DISK OF L1527
Dominique M. Segura-Cox , Leslie W. Looney , Ian W. Stephens , Manuel Fern´andez-L´opez , Woojin Kwon ,John J. Tobin , Zhi-Yun Li , Richard Crutcher Accepted by ApJL
ABSTRACTWe present subarcsecond ( ∼ ′′ ) resolved observations of the 1.3 mm dust polarization from theedge-on circumstellar disk around the Class 0 protostar L1527. The inferred magnetic field is consistentwith a dominantly toroidal morphology; there is no significantly detected vertical poloidal componentto which observations of an edge-on disk are most sensitive. This suggests that angular momentumtransport in Class 0 protostars (when large amounts of material are fed down to the disk from theenvelope and accreted onto the protostar) is driven mainly by magnetorotational instability ratherthan magnetocentrifugal winds at 50 AU scales. In addition, with the data to date there is an early,tentative trend that R >
30 AU disks have so far been found in Class 0 systems with average magneticfields on the 1000 AU scale strongly misaligned with the rotation axis. The absence of such a diskin the aligned case could be due to efficient magnetic braking that disrupts disk formation. If this isthe case, this implies that candidate Class 0 disk systems could be identified by the average magneticfield direction at ∼ Subject headings:
ISM: individual objects (L1527) — ISM: magnetic fields — polarization — stars:protostars INTRODUCTION
Circumstellar disks are a key component of the starformation process and are fundamental for accretion andangular momentum distribution during the early phasesof star formation. Class 0 objects are the youngestand most embedded protostars, and circumstellar disksform at this earliest stage of star formation if angu-lar momentum is conserved during cloud collapse (e.g.,Cassen & Moosman 1981). Class 0 disks are extremelyobscured by envelopes, which contribute &
90% of thetotal emission (Looney et al. 2000), making the searchfor Class 0 disks challenging. However, observations ofClass 0 disks and their properties are essential to pro-vide the initial conditions for mass accretion onto thecentral protostar and planet formation. To date, only afew Class 0 systems have observed disks with clear Keple-rian rotation (e.g., Tobin et al. 2012; Murillo et al. 2013;Codella et al. 2014); L1527, VLA 1623, and HH212 haveKeplerian disks with R >
30 AU, sizes larger than mag-netic braking models predict.In addition to the properties of young disks and en-velopes, the morphology and strength of the magneticfield in these systems also play an important role in starformation (e.g., Crutcher 2012). For example, the mor-phology of the magnetic field in the young disk pro-vides important clues into angular momentum trans- Department of Astronomy, University of Illinois, Urbana, IL61801, USA; [email protected] Institute for Astrophysical Research, Boston University,Boston, MA 02215, USA Instituto Argentino de Radioastronom´ıa, CCT-La Plata(CONICET), C.C.5, 1894, Villa Elisa, Argentina SRON Netherlands Institute for Space Research, Landleven12, 9747 AD Groningen, The Netherlands National Radio Astronomy Observatory, Charlottesville, VA22903, USA Astronomy Department, University of Virginia, Char-lottesville, VA 22904, USA port: disk accretion driven by magnetorotational in-stabilities (MRI, Balbus & Hawley 1998) favor toroidalfields while angular momentum removal via magnetocen-trifugal winds arising from the disk favor poloidal fields(e.g., Blandford & Payne 1982). The magnetic field mor-phology in the envelope and disk can be inferred; dustgrains preferentially align with their long-axis perpen-dicular to the magnetic field, causing the dust emissionto be polarized (e.g., Lazarian 2007). An interferomet-ric survey of dust polarization around 26 low-mass Class0/I protostars has been recently conducted (TADPOL,Hull et al. 2013, 2014). For these sources, the averagemagnetic field axes are generally misaligned with the ro-tation axes of the systems (as proxied by the outflow). Inthe case of L1527 and VLA 1623, TADPOL observationsshow that the magnetic field lines are perpendicular tothe outflows. However, the TADPOL results only probeenvelope size scales and do not approach disk scales. Themagnetic field morphology on smaller scales has been ob-served in the Class 0 protostars L1157 and IRAS 16293-2422 B. L1157, whose disk is yet to be resolved and hasR <
20 AU (Tobin et al. 2013a), has vertical poloidal com-ponent magnetic fields aligned with the rotation axis ofthe system in the inner envelope (Stephens et al. 2013).IRAS 16293-2422 B—which is thought to have a face-ondisk and hence no clear Keplerian motion—has resolvedobservations of the candidate disk with a polarizationpattern indicative of a toroidal magnetic field compo-nent (Rao et al. 2014), although the face-on geometrymakes the detection of any vertical poloidal componentimpossible.In this Letter, we present high-resolution CARMA1.3 mm dust polarimetric observations of the Class 0 pro-tostar L1527. Lower-resolution CARMA 1.3 mm polari-metric observations of L1527 were previously conductedas a part of the TADPOL survey, probing the magneticfield morphology on ∼ ∼
50 AU disk Segura-Cox et al.size scales. The data presented are the first detection ofpolarized dust emission emanating directly from a Class0 Keplerian disk. OBSERVATIONS
CARMA 1 mm full-Stokes observations of L1527 wereobtained in 6 tracks of the B array ( ∼ ′′ resolution)from 2013 December 9–13 and 15 for a total of 21 hourson-source. The correlator was set up with a local oscil-lator frequency of 233.731 GHz and four 500 MHz-widebands centered at intermediate frequency values of 2.187,2.740, 4.716, and 5.544 GHz. We used the MIRIAD soft-ware package (Sault et al. 1995) to reduce the data. Thepolarization calibration followed the standard process forCARMA (Hull et al. 2014). The phase and polarizationleakage calibrator for all tracks was 0510+180. The pre-ferred bandpass calibrator was 3C84, and 3C454.3 wasused when 3C84 was unavailable. For most tracks, theflux calibrator was MWC349 with a calibration accuracyestimation of ∼ RESULTS
The 1.3 mm dust emission map of L1527 is presentedin Figure 1 and is consistent with the known edge-ondisk (Tobin et al. 2013b). The disk has been resolvedpreviously at 3.4 mm and 870 µ m (Tobin et al. 2012)and also has been shown to have Keplerian motion anda radius of 54 AU (Ohashi et al. 2014, in press). Ourobservations are the first resolved detection of the diskat 1.3 mm. At a resolution of ∼ ′′ and a distanceof 140 pc (Loinard et al. 2007), our interferometric ob-servations probe L1527 on size scales of ∼
50 AU. Anelliptical Gaussian fit to the high-resolution 1.3 mmStokes I data measures a deconvolved size of 0.53 ′′ × ′′ with position angle of 5.2 ◦ (PA, measured counterclock-wise), consistent with the deconvolved sizes at 3.4 mm,870 µ m (Tobin et al. 2013b), and 1.3 mm (Ohashi et al.2014, in press). In addition, our flux density at 1.3 mm(139 ± µ m and the derived β =0 fromTobin et al. (2013b), we can estimate the expected diskemission at 1.3 mm (116 mJy and 96 mJy, respectively),which is congruent with our measured 1.3 mm fluxeswhen taking account a 20% extrapolating and amplitudeuncertainty. Based on this evidence, our observations aredominated by disk emission of the L1527 system, with lit-tle contamination from the large-scale envelope emission.We detect dust polarization of the young disk over2 synthesized beams with an average polarization of2.5% ± ◦ ± ◦ measuredcounterclockwise from north, aligning well with theStokes I elliptical Gaussian fitted position angle of5.2 ◦ ± ◦ . The inferred magnetic field (with polarizationvectors rotated by 90 ◦ ) is shown in Figure 1. The mor- RA (J2000) +26°03'08.5"09.0"09.5"10.0"10.5" D e c ( J ) F l u x ( J y / b e a m ) L1527
Figure 1.
Polarimetric map (polarization vectors rotated by 90 ◦ to show inferred magnetic field orientation) of the L1527 disk fromCARMA data with a 0.39 ′′ × ′′ beam. Fractional polarizationvectors ≥ σ displayed. Contours are Stokes I data with levels of[-6, -4, -3, 3, 4, 6, 10, 20, 40, 60, 80, 100] × σ , σ =0.45 mJy beam − .Grayscale shows the polarized intensity ≥ σ . Outflows in the planeof the sky are marked by red and blue arrows. phology of the inferred field is parallel to the disk axis,as is expected from an edge-on toroidal field—uniformand aligned with the disk. We compare a uniform fieldat a 5 ◦ PA (the same as the dust emission) to the dataand find a reduced χ <
1. The polarization fraction ofthe circumstellar disk of L1527 is larger than the 1.4%polarization fraction found in the face-on candidate diskof IRAS 16293-2422 B (Rao et al. 2014), although thelower polarization fraction of IRAS 16293-2422 B maybe due in part to beam-averaging; due to orientation,an edge on toroidal field is less beam-averaged as thevectors are more uniform. Our polarization percent-age is similar to the theoretically predicted 2-3% po-larization fraction found in simulations of magnetizeddisks (Cho & Lazarian 2007). On the other hand, ob-servations of the disks of older T Tauri systems havemuch lower polarization percentages <
1% (Hughes et al.2009, 2013), which may be an outcome of dust process-ing or de-alignment mechanisms during disk evolution(Stephens et al. 2014).A uniform field in the plane of the disk is physicallyunlikely for a rapidly-rotating, Keplerian disk system.One expects either poloidal, toroidal, or a combinationof the two in such a disk (e.g. Balbus & Hawley 1998;K¨onigl & Pudritz 2000). For an edge-on disk, observa-tions are most sensitive to vertical poloidal field com-ponents because they are expected to vertically threadthe disk and thus lay roughly in the plane of the sky.However, our data does not exhibit any obvious poloidalmorphology which would be perpendicular to the disk.To show that a toroidally dominant morphology is con-sistent with our observations, we compare to a purelytoroidal, a purely vertical poloidal, and combinations oftoroidal and vertically poloidal toy models (see captionof Figure 2 for details). We used the best-fit disk pa-rameters of Tobin et al. (2013b): disk inclination angleof 85 ◦ , 0 ◦ PA, M disk = 0 . ⊙ , R inner = 0 . outer = 125 AU, and stellar and accretion luminosityof 2.75 L ⊙ . The temperature distribution was calcu-lated using the Monte-Carlo radiative code RADMC-3Dhe Magnetic Field in the Class 0 Protostellar Disk of L1527 3 (a) (b) (c) (d) (e) (f) Figure 2.
Synthetic maps of the L1527 disk magnetic field morphology. Contours, grayscale, and vectors are the same as Figure 1. (a)Toroidal field only, (b) 70% toroidal/30% vertical poloidal field, (c) 60%/40%, (d) 50%/50%, (e) 40%/60%, (f) vertical poloidal field only. (Dullemond & Dominik 2004). The Stokes I, Q, and Umaps were numerically solved using dust radiative trans-fer in the disk along the line of sight with an assumedconstant polarization fraction. The magnetic field vec-tors at each integral element have been tilted and rotatedbased on the inclination and position angles of the diskmodel. Finally, all maps are convolved with the synthe-sized beam from the polarization observations (for moredetails see Stephens et al. 2014). As shown in Figure2, the detected polarization is characterized well by ourtoroidally dominant toy models with up to 40% poloidalfield (Figure 2a, 2b, 2c), with a reduced χ ∼ DISCUSSION
In quiescent (non-turbulent) systems with alignedmagnetic field and disk rotation axes, magnetic brak-ing can have a significant effect on the infalling materialin the ideal MHD limit, removing angular momentum(Mellon & Li 2008; Hennebelle & Fromang 2008), andsuppressing growth of the early circumstellar disk by al-lowing larger accretion rates (Li et al. 2011). Magneticbraking can be so effective in Class 0 sources that rota-tionally supported disks are limited to R <
10 AU (e.g., Dapp & Basu 2010), only reaching R ∼
100 AU at the endof the main mass accretion phase when the envelope isless massive and magnetic braking becomes inefficient(e.g., Dapp et al. 2012; Mellon & Li 2009; Machida et al.2011). Conversely to this prediction, Keplerian diskshave been detected in Class 0 sources with sizes largerthan expected from magnetic braking models. L1527 andVLA 1623 have disk sizes of R ∼
54 AU and R ∼
189 AUrespectively (Ohashi et al. 2014, in press; Murillo et al.2013), and HH212 has a disk of R >
30 AU (Codella et al.2014).There are large disks in some young systems, suggest-ing that significant magnetic braking has not happened,has already occurred, or the magnetic field has diffusedto the point where a R >
10 AU disk could form. On theother hand, similar high resolution observations of theClass 0 protostar L1157 have not detected a circumstellardisk down to spatial resolutions of ∼
15 AU (Tobin et al.2013a). This result suggests that magnetic braking mayhave been more significant in L1157 than L1527. Ofcourse, there are differences in age since L1527 is an oldersource and could have been classified as a Class I source,were it not viewed edge-on (Tobin et al. 2008). What isclear is that some Class 0 sources have R >
10 AU cir-cumstellar disks and others do not. Such differences indisk size could be a consequence of misalignment betweenthe magnetic field and rotation axis, which modifies thestrength of magnetic braking (Hennebelle & Ciardi 2009;Joos et al. 2012; Li et al. 2013; Krumholz et al. 2013).To better understand the role of the magnetic field inthe early disk and envelope, we can compare the mag-netic field of L1527 presented here with the larger-scale Segura-Cox et al.
Table 1
B array and TADPOL ResultsData Set Stokes I Flux Polarized Flux ¯ P % PA Beam(mJy) (mJy) (%) ( ◦ ) ( ′′ )B array 139 ± ± ± ± × ± ± ± ± × Note . — Uncertainties are statistical. Results were found using data > σ . Position angles are measured counterclockwise. Stokes I fluxis measured across the entire disk or inner envelope; polarized flux and ¯ P % are measured in the polarized region only. magnetic field detected with TADPOL (Figure 3 and Ta-ble 1). When comparing the polarization at 1000 AUand 50 AU, the two scales have the same average fieldangle: perpendicular to the outflow and well aligned withthe disk plane. The higher resolution observations haveless than half of the polarized emission, which suggestswe are resolving out large-scale emission. The projectedfield morphology on the 1000 AU scale is consistent withthe view that the initial magnetic field on this scale isgreatly misaligned with respect to the rotation axis, al-though it is also possible that the field on this scaleis already modified by the collapsing and rotating mo-tions in the envelope. The magnetic fields on even largerscales are expected to be affected less by rotation andcollapse, and are more likely to keep their initial config-uration. The fields are better traced by single dish ob-servations using SHARP on CSO Davidson et al. (2011)and SCUPOL on JCMT (Figure 17 of Hull et al. 2014;Matthews et al. 2009). These single dish data are mod-eled by Davidson et al. (2014, in press) together withCARMA data; we refer the reader to that paper for adetailed discussion of the magnetic field on large-scales.In any case, on the small scale of the disk, the avail-able data are consistent with the field being predom-inantly toroidal, and such a toroidally dominant diskmagnetic field is also consistent with the magnetorota-tional instability (e.g., Balbus & Hawley 1998) drivingaccretion during the main accretion phase. We do not de-tect significant vertical poloidal component fields that areneeded to launch magnetocentrifugal winds; such windsare probably not the dominant driver of angular momen-tum transport during the main accretion phase at the ∼
50 AU size scale. On the other hand, a disk wind wouldlikely be launched from the disk upper layers that are notwell traced by our observations, which are most sensitiveto the dust in the midplane. If the poloidal field is some-how limited to the surface layers, then our observationswould be less constraining on the existence or absence ofa disk wind.With dust polarization observations and high-resolution searches for disks, we can compare the mag-netic field orientations and morphologies with disk prop-erties. L1527, VLA 1623, and L1157 have all been ob-served with CARMA dust polarization at 500 AU or bet-ter resolution (Hull et al. 2014). L1527 and VLA 1623,the first two Class 0 systems with known Keplerian disks,have average magnetic fields perpendicular to the rota-tion axes (inferred from the outflow direction). In con-trast, L1157, a system with a disk R <
20 AU in size, hasan inferred average magnetic field parallel to the rota-tion axis. Although we have few examples so far (Table2), this observational, tentative trend is intriguing andsuggested from theory (e.g., Joos et al. 2012); the mag-netic field morphology at the earliest stages of collapse
RA (J2000) +26°03'06.0"08.0"10.0"12.0"14.0" D e c ( J ) F l u x ( J y / b e a m ) L1527
Figure 3.
Polarimetric map of the L1527 inner envelope from theCARMA TADPOL data with a 2.63 ′′ × ′′ beam. Contours areStokes I data with levels of [-6, -4, -3, 3, 4, 6, 10, 20, 40, 60, 80,100] × σ , σ =2.34 mJy beam − . Grayscale and vectors are the sameas Figure 1. may play an important role in the formation of the ear-liest disk, with strongly misaligned magnetic fields androtation axes producing R >
10 AU disks at early times.Clearly more objects are necessary to better establishthis relationship.We therefore suggest that the morphology of the mag-netic field in the inner envelope ( ∼ >
10 AU. Sources withR >
10 AU disks may have a projected magnetic field onthe 1000 AU scale perpendicular to the rotation axis, andthe magnetic field would appear uniform for edge-on diskcases like L1527. At more oblique viewing angles at highresolution, a purely toroidal field would be observed astwo maxima of fractional polarization along the axis ofrotation on either side of the protostar, with the magneticfield oriented perpendicular to the outflow axis withinthe maxima regions (e.g., see Figure 1 in Hughes et al.2013). A purely toroidal disk field observed directly downthe rotation axis will appear as a pattern of concentriccircles (e.g., Padovani et al. 2012). Sources with R < Table 2
Class 0 Magnetic Field Morphologies and Candidate DisksSource α δ
Mag-Rot-Axis a Candidate Known References(J2000) (J2000) Disk? b Keplerian?L1527 04:39:53.9 26:03:09.6 Perpendicular Yes Yes 1,2IRAS 16293-2422 B 16:23:22.9 -24:28:35.7 Perpendicular Yes No 3,4VLA 1623 16:26:26.4 -24:24:30.5 Perpendicular Yes Yes 5,2L1157 20:39:06.2 68:02:15.8 Parallel No No 6,2
Note . — a Orientation of the magnetic field compared to the rotation axis (estimated from the outflow for L1157). b Does the sourcehave a candidate disk of R >
30 AU?
References . — (1) Tobin et al. (2012); (2) Hull et al. (2014); (3) Zapata et al. (2013); (4) Rao et al. (2014); (5) Murillo & Lai (2013);(6) Tobin et al. (2013a) phology. In the extreme cases of purely toroidal andpurely poloidal magnetic field components, the observedmorphology alone can be used to distinguish between thetwo cases and point towards R >
10 AU candidate Class0 disks. CONCLUSIONS
L1527 is the first Class 0 protostar with a known Kep-lerian disk and direct detection of linearly polarized dustemission from the circumstellar disk, indicating magneticfields are aligned perpendicular to the rotation axis ofthe disk. The magnetic field is consistent with toroidallydominant field lines. It may be that L1527’s large diskarises from the strongly misaligned rotation axis andmagnetic field on large scales, while aligned rotation axesand magnetic fields inhibit disk formation on R >
10 AUscales. The toroidally dominant field morphology fa-vors the magnetorotational instability (Balbus & Hawley1998) as the dominant angular momentum transport pro-cess in Class 0 circumstellar disks.L1527 is one of two Class 0 sources (with VLA 1623)where both magnetic fields and Keplerian disks havebeen detected. Both of these sources have perpendicularmagnetic fields and rotation axes (Murillo & Lai 2013;Tobin et al. 2012; Hull et al. 2013) on 1000 AU scales.The alternative case is where the magnetic field and ro-tation axes are parallel on envelope scales, such as theClass 0 source L1157 with no disk detected down to 20AU (Tobin et al. 2013a). It is possible that aligned mag-netic fields may have braked rotation so efficiently as toinhibit the disk formation and growth at early times.The tentative trend of misaligned magnetic field and ro-tation axes in Class 0 systems with disks is suggestiveand expected from theory, requiring follow-up to makehard conclusions about Class 0 disk formation.We thank Chat Hull and Dick Plambeck for assistancewith data reduction. This research made use of APLpy,an open-source plotting package for Python hosted athttp://aplpy.github.com.Support for CARMA construction was derived fromthe states of California, Illinois, and Maryland, the JamesS. McDonnell Foundation, the Gordon and Betty MooreFoundation, the Kenneth T. and Eileen L. Norris Foun-dation, the University of Chicago, the Associates of theCalifornia Institute of Technology, and the NSF. Ongo-ing CARMA development and operations are supportedby the NSF under a cooperative agreement and by theCARMA partner universities.J. Tobin acknowledges support provided by NASA through Hubble Fellowship grant
REFERENCESBalbus, S. A., & Hawley, J. F. 1998, Rev. Mod. Phys., 70, 1Blandford, R. D. & Payne, D. G. 1982, MNRAS, 199, 883Cassen, P., & Moosman, A. 1981, Icar, 48, 353Cho, J., & Lazarian, A. 2007, ApJ, 669, 1085Codella, C., Cabrit, S., Gueth, F., et al. 2014, A&A, 568, 5Crutcher, R. M. 2012, ARA&A, 50, 29Dapp, W. B., & Basu, S. 2010, A&A, 521, L56Dapp, W. B., Basu, S., & Kunz, M. W., 2012, A&A, 541, A35Davidson, J.A., Li, Z.-Y., Kwon, W., et al. 2014, ApJ, in pressDavidson, J.A., Novak, G., Matthews, T. G., et al. 2011, ApJ,732, 97Dullemond, C. P., & Dominik, C. 2004, A&A, 417, 159Hennebelle, P., & Ciardi, A. 2009, A&A, 506, 29Hennebelle, P., & Fromang, S. 2008, A&A, 477, 9Hughes, A. M., Andrews, S. M., Espaillat, C., et al. 2009, ApJ,698, 131Hughes, A. M., Hull, C. L. H., Wilner, D. J, et al. 2013, AJ, 145,115Hull, C. L. H., Plambeck, R. L., Bolatto, A. D., et al. 2013, ApJ,768, 159Hull, C. L. H., Plambeck, R. L., Kwon, W., et al. 2014, ApJS,213, 13Joos, M., Hennebelle, P., & Ciardi, A. 2012, A&A, 543, A128Kataoka, A., Machida, M. N., & Tomisaka, K. 2012, ApJ, 761, 40K¨onigl, A., & Pudritz, R. E. 2000, in Protostars and Planets IV,ed. V. Mannings, A. P. Boss, & S. S. Russell (Tucson, AZ:Univ. Arizona Press), 759Krumholz, M. R., Crutcher, R. M., & Hull, C. L. H. 2013, ApJ,767, L11Lazarian, A. 2007, JQSRT, 106, 225Li, Z.-Y., Krasnopolsky, R., & Shang, H. 2011, ApJ, 738, 180Li, Z.-Y., Krasnopolsky, R., & Shang, H. 2013, ApJ, 774, 82Loinard, L., Chandler, C. J., Rodr´ıguez, L. F., et al. 2007, ApJ,670, 1353Looney, L. W., Mundy, L. G., & Welch, W. J. 2000, ApJ, 529, 477Machida, M. N., Inutsuka, S.-I., & Matsumoto, T. 2007, ApJ,670, 1198Machida, M. N., Inutsuka, S.-I., & Matsumoto, T. 2011, PASJ,63, 555Matthews, B. C., McPhee, C. A., Fissel, L. M., & Curran, R. L.2009, ApJS, 182, 143Mellon, R. R., & Li, Z.-Y. 2008, ApJ, 681, 1356Mellon, R. R., & Li, Z.-Y. 2009, ApJ, 698, 922Murillo, N. M., & Lai, S.-P. 2013, ApJL, 764, L15Murillo, N. M., Lai, S.-P., Bruderer, S., et al. 2013, A&A, 560, 103Ohashi, N., Saigo, K., Aso, Y., et al., 2014, ApJ, in pressPadovani, M., Brinch, C., Girart, J. M., et al. 2012, A&A, 543,A16Rao, R., Girart, J. M., Lai, S.-P., et al. 2014, ApJL, 780, L6