FIRST OPTICAL AND NEAR-INFRARED POLARIMETRY OF A MOLECULAR CLOUD FORMING A PROTO-BROWN DWARF CANDIDATE

The question of what role magnetic field (B-field) plays in either facilitating or hindering the formation of a rotationally supported disk is not fully answered. MHD simulations (e.g., Machida et al. 2005; Banerjee & Pudritz 2006) have shown that the efficient transport of angular momentum from the inner to the outer regions of the cloud through magnetic breaking can suppress the formation of a centrifugally supported disk. This is true even for relatively modest values of the magnetic intensity. In reality, though, disks are often observed around Class I and II young stellar objects (e.g., Jørgensen et al. 2009; Takakuwa et al. 2012) and recently also around Class 0 sources (e.g., Tobin et al. 2012). However, Maury et al. (2010, 2014) have not found any disks around Class 0 sources. Exactly how nature overcomes the magnetic breaking to form disks is still an open issue.

Recent simulations of magnetized core collapse suggest that any misalignment that may exist between the B-field and the rotation axis significantly reduces the magnetic breaking, thus allowing the formation of rotationally supported disks (Joos et al. 2012, 2013). By considering outflow axis as a proxy for the rotation axis, Hull et al. (2013) found that rotation axes are not aligned with the ∼1000 AU scale B-field. Based on the observations of seven, more carefully selected low-mass sources, Davidson et al. (2011) and Chapman et al. (2013) found a positive correlation between core B-field and outflow directions. However, the outflow from low-mass sources possess sufficient energy density to distort the B-field structure in the cores (Davidson et al. 2011). Therefore, it is essential to examine the relative orientation of outflow and B-field in sources that have the lowest energy density possible.

LDN 328 (hereafter L328) is a molecular cloud containing a dark opaque region of ∼2' × 2' and several long curved structures extending to the south–west direction. The cloud is located at a distance of ∼220 pc (Maheswar et al. 2011). L328 contains three sub-cores, namely, S1, S2, and S3 (Lee et al. 2013). Previously classified as "starless" (Lee & Myers 1999), a very low luminosity object (VeLLO) was discovered, based on the Spitzer Space Telescope (SST) and the ground-based observations (from near-infrared, NIR, to millimeter wavelengths), in the least massive ($\sim 0.09\;{{M}_{\odot }}$) of the three sub-cores, S2 (Lee et al. 2009). The broadest line widths observed in ¹³CO and N₂H⁺ toward sub-core S2 is shown as the direct evidence for the physical association of L328-IRS with it (Lee et al. 2013). The VeLLO, L328-IRS, is classified as a Class 0 object (Lee et al. 2009). The internal luminosity of $\sim 0.05\;{{L}_{\odot }}$ of L328-IRS is a factor of ∼30 fainter than the lowest-mass protostar of $\sim 0.08\;{{M}_{\odot }}$ that could be formed through typical accretion rate of $\sim 2.0\times {{10}^{-6}}\;{{M}_{\odot }}\;{\rm y}{{{\rm r}}^{-1}}$. A sub-parsec-scale CO outflow associated with L328-IRS was detected by Lee et al. (2013). The mass accretion rate of $\sim 3.6\times {{10}^{-7}}\;{{M}_{\odot }}\;{\rm y}{{{\rm r}}^{-1}}$ inferred from the analysis of the CO outflow is an order of magnitude less than the typical value of low-mass protostars (Lee et al. 2013). With such a low accretion rate and a small envelope mass of $\sim 0.09\;{{M}_{\odot }}$ of sub-core S2 available, L328-IRS may only accrete the mass of a brown dwarf and not that of a star (assuming the main accretion phase only lasts during the Class 0 stage). Hence, L328-IRS is considered a good candidate for a proto-brown dwarf. The momentum flux estimated for L328-IRS (Lee et al. 2013) is found to be the lowest among the VeLLOs, implying that the distortion to the B-field is expected to be the lowest in its vicinity.

In this Letter, we report the plane-of-the-sky magnetic field (B_pos-field) mapping observations of a core that harbors a proto-brown dwarf candidate using optical and NIR polarimetry. The main objective of the study is to map the B_pos-field geometry around L328-IRS and to correlate it with the orientation of the outflow associated with this proto-brown dwarf candidate. While passing through aspherical interstellar dust grains, the starlight gets polarized due to selective absorption (Lazarian 2003). While the optical polarization vectors trace the orientation of the ambient B_pos-field at the periphery of the clouds where ${{A}_{V}}\;\sim$ 1–2 mag (Goodman et al. 1995), those in the NIR trace the B_pos-field structure of the inner parts of the clouds where ${{A}_{V}}\;\sim$ 10–20 mag (Tamura 1999; Tamura et al. 2011). Therefore, by combining the results obtained in both wavelength regimes, one can trace the B_pos-field geometry at different physical scales of the clouds.

Optical (R-band) polarimetric observations were carried out on 2013 May 15 and 16 using the ARIES Imaging Polarimeter (AIMPOL; Rautela et al. 2004) mounted at the cassegrain focus of the 104 cm Sampurnanand telescope, India. A 1024 × 1024 pixel² CCD chip (Tektronix TK1024) was used. When combined with the camera, the 8' field of view (FOV) falls on the central $325\;\times \;325\;{\rm pixel}{{{\rm s}}^{2}}$. The mean exposure time per observed field is ∼250 s. The plate scale of the CCD used is 148 pixel⁻¹. AIMPOL consists of an achromatic half-wave plate (HWP) modulator and a Wollaston prism beam-splitter. The observations were carried out using a standard Johnson ${{R}_{{\rm KC}}}$ filter. The standard aperture photometry in the IRAF package was used to extract the fluxes of ordinary and extraordinary beams for all the observed sources with a good signal-to-noise ratio. Four Stokes parameters were calculated keeping the fast axis of the HWP at 0°, 225, 45°, and 675 positions. The typical instrumental polarization, estimated by observing zero polarization standard stars, is found to be 0.1% (Rautela et al. 2004; Eswaraiah et al. 2011; Soam et al. 2013). We observed polarized standards from Schmidt et al. (1992) for the calibration of the polarization angles (for details, see Soam et al. 2013).

NIR polarimetric observations were carried out using the SIRIUS camera and the SIRPOL polarimeter on the IRSF 1.4 m telescope in South Africa, on the night of 2013 July 7. IRSF/SIRPOL can provide 1.25 (J), 1.63 (H), and 2.14 (K_s) μm images, simultaneously, with a FOV of 77 × 77 (the NIR data used in this study is from the ∼4' × 4' region around L328-IRS to see the magnetic fields in the immediate vicinity). The pixel scale is 045 pixel⁻¹. The camera is equipped with three 1024 × 1024 HgCdTe (HAWAII) arrays, three broad-band filters, and two beam-splitters for simultaneous imaging. The polarimeter is composed of an achromatic (0.5–2.5 μm) wave-plate rotator unit and a high-extinction-ratio polarizer, both of which are located on the upstream side of the camera at room temperature. Details of the SIRIUS camera are presented in Nagashima et al. (1999) and Nagayama et al. (2003), and those of the SIRPOL polarimeter in Kandori et al. (2006). The instrumental polarization of SIRPOL is found to be negligible. The polarization efficiencies are 95.5, 96.3, and 98.5% in the J, H, and K_s bands, respectively. During observations, dithered exposures were taken at four wave-plate angles (0°, 225, 45°, and 675). After image calibrations in the standard manner using IRAF (dark subtraction, flat-fielding with twilight flats, bad-pixel substitution, and sky subtraction), the Stokes parameters, the degree of polarization, and position angles were calculated (see Kwon et al. 2011).

The degree of polarization (${{P}_{R}}$) and position angles (${{\theta }_{R}}$) of 215 sources (with $P/\sigma P\gt 2$) observed in the R band toward L328 are shown in Figure 1(a). Figure 2(a) shows the R-band polarization vectors drawn on the continuum-subtracted Hα image (obtained from the SuperCosmos survey). L328-IRS is identified with a star symbol. The double-headed arrow shows the direction of outflow associated with it. Figure 1(a) shows the histogram of ${{\theta }_{R}}$ values and the ${{P}_{R}}$ versus ${{\theta }_{R}}$ plot. The histogram of the ${{\theta }_{R}}$ shows a bimodal distribution with no clear peaks. The mean value of ${{P}_{R}}$ is found to be 1.8%. The mean values of the ${{\theta }_{R}}$ for the two dominant components are found to be 87° and 176°. The 87° and 176° components are dominant toward the head and the tail parts of the L328, respectively (see Figure 2(a)).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The region over which we present the NIR polarimetric data (∼4' × 4') is identified using a cyan box in Figure 2(a). The results of NIR polarization with $P/\sigma P\gt 3$ in J (109 sources), H (159 sources), and K_s (87 sources) bands are shown in Figures 1(b)–(d), respectively. The measurements are available from ∼0.02 to 0.2 pc. Figure 2(b) shows the polarization vectors obtained in J (blue), H (green), and K_s (red) bands overplotted on the H-band image of the region. The histograms of the position angles (${{\theta }_{J}}$, ${{\theta }_{H}}$, and ${{\theta }_{K}}$) and the degree of polarization (${{P}_{J}}$, ${{P}_{H}}$, and ${{P}_{K}}$) versus the corresponding position angles in the J, H, and K_s bands are shown in Figures 1(b)–(d), respectively. Mean values of ${{P}_{J}}$, ${{P}_{H}}$, and ${{P}_{K}}$ are found to be 5.6 ± 4.5%, 3.2 ± 1.8%, and 2.1 ± 1.0%, respectively. The mean values of ${{\theta }_{J}}$, ${{\theta }_{H}}$, and ${{\theta }_{K}}$ are found to be 64 ± 28°, 64 ± 33°, and 52 ± 26°, respectively. The histogram of ${{\theta }_{H}}$ shows a two-population distribution with a secondary peak at ∼90°, which is not very conspicuous in the J and K_s bands.

The cloud as a whole shows a head–tail morphology, with the tail extending to ∼15' in projection on the sky. The cloud morphology is similar to that of cometary globules that are typically found near OB associations in H ii regions (e.g., Reipurth 1983). The parsec-scale B_pos-field traced by the R-band polarimetry seems to follow the curved structure of the cloud. The curvature in the B-field geometry would result in a range of ${{\theta }_{R}}$ values. The large dispersion found in the histogram of ${{\theta }_{R}}$ in Figure 1(a) with no clear peaks in the distribution could be due to this effect. The large dispersion in ${{\theta }_{R}}$ could also occur if the B-field in the region is highly distorted or tangled.

The B_pos-field geometry mapped in NIR wavelengths traces the field toward the inner parts (∼0.02–0.2 pc) of the cloud. The histograms presented in Figures 1(b)–(d) show that the dispersion in the position angles decreases toward longer wavelengths. This happens when the B-field becomes more and more ordered as we probe deeper into the cloud. However, a systematic decrease in the degree of polarization is also found toward longer wavelengths in the NIR. This occurs when the population of dust grains responsible for the polarization of the background starlight declines as a result of the grain growth toward inner parts of the cloud. A dominant component of position angle at ∼60° is evident in all the three (J, H, and K_s) wavelengths. The degree of polarization is found to peak around this position angle and shows a systematic drop for other position angles, implying that the dust grains responsible for producing polarization are sufficiently aligned with the ambient magnetic field. It is worth noting that the region to the south–west of the yellow line drawn in Figure 2(b) shows a more ordered B_pos-field structure. This region belongs to the tail part of L328. The ordered field structure could be due to the shielding of the tail part from the external influence (responsible for causing the head–tail structure) by the head of L328. We, therefore, may expect undisturbed pre-existing B-fields to be preserved here. The histograms of the data from two regions are shown separately in the top panels of Figure 3. The striped and shaded histograms represent the data from the north–east and the south–west regions, respectively. The shaded histograms show peak position angles at ∼60° with relatively less dispersion. In Figure 2(b), it is quite apparent that even in NIR wavelengths, the B_pos-field of the densest parts of the core could not be traced. However, since sub-core S2 is located at the southern edge of the L328 core, to trace the B_pos-field in the vicinity of L328-IRS, we selected all the sources observed in a circular region of 1' (∼0.07 pc) radius around L328-IRS. The results are plotted in the lower panels of Figure 3. Gaussian-fitted mean values of ${{\theta }_{J}}$, ${{\theta }_{H}}$, and ${{\theta }_{K}}$ with corresponding 1σ uncertainty are found to be 60 ± 19°, 60 ± 15°, and 61 ± 13°, respectively. Relatively low dispersion values found in ${{\theta }_{K}}$ show that the B_pos-field structure around L328-IRS is well ordered, as found in a number of cores where low-mass protostars are forming (Holland et al. 1996; Davidson et al. 2011; Chapman et al. 2013; Soam et al. 2015). However, because of the low number statistics, estimating a single B-field orientation based on the Gaussian fit may not be accurate. Therefore, we considered the range of ${{\theta }_{K}}$ values (which is from 40° to 70°) where more than one polarization measurement is available in the histogram and calculated the mean value of ${{\theta }_{K}}$. The estimated B_pos-field orientation and the uncertainty is found to be ∼55 ± 15°.

Zoom In Zoom Out Reset image size

The sub-parsec (∼0.08 pc) scale bipolar outflow detected toward L328-IRS shows both blue- and redshifted components (Lee et al. 2013) in the northeast–southwest direction. The extent of outflow is comparable to the circular region chosen by us around L328-IRS to determine the B_pos-field orientation. Based on the SST observations, only one point source (i.e., L328-IRS) has been detected so far in L328. Therefore, the outflow detected toward L328 is most likely associated with L328-IRS. Apart from IRAM 04191-IRS and L673-7-IRS, L328-IRS is the third VeLLO where an outflow signature has been detected based on the single-dish observations. The presence of outflow is more pronounced in CO(J = 3–2) than in CO(J = 2–1). The redshifted component is found to be more noisy, weak, and unclear than the blueshifted component. By considering the spatial extent of both the components of the outflows, the possible outflow position angle could be anywhere between 0° and 60°. Thus, the most plausible outflow position angle is ∼30 ± 30°. Contemplating the uncertainties in the determination of position angles, the projected angular offset between the outflow axis and the B $_{{\rm pos}}$-field direction is found to be in the range of 0°–70°. Among the cores with VeLLOs (Soam et al. 2015), IRAM 04191+1522 has the inner magnetic fields inferred based on the SCUBA-POL⁷ observations by Matthews et al. (2009). The angular offset between the inner magnetic field direction (32°; Soam et al. 2015) and the the orientation of the outflow axis (28°; Belloche et al. 2002) in IRAM 04191-IRS is found to be 4°. If one assumes that the outflows emerge perpendicular to the circumstellar disks (or along the rotation axes), as is expected from the magnetocentrifugally launched wind model (e.g., Shu et al. 1994), then the results toward L328-IRS and IRAM 04191-IRS imply that the rotation axes in these systems are preferably parallel to the B_pos-field.

Using a modified Chandrasekhar–Fermi (CF) relation (${{B}_{{\rm pos}}}=9.3\sqrt{n({{H}_{2}})}\delta v/\delta \theta$; Chandrasekhar & Fermi 1953; Crutcher 2005), we estimated the B_pos-field strength toward L328-IRS. The velocity dispersion ($\delta v$) was calculated using the N₂H⁺ line width (${\Delta}v=0.61\pm 0.03\;{\rm km}\;{{{\rm s}}^{-1}}$) measured by Lee et al. (2013). The dispersion ($\delta \theta$) in ${{\theta }_{K}}$ for sources from within a 1' region surrounding L328-IRS, corrected for the position angle uncertainty (Lai et al. 2001; Franco et al. 2010), was used in the CF relation. Using a N₂H⁺ line, Crapsi et al. (2005) estimated the value of the central volume density toward L328 as ${{n}_{c}}=1.8\pm 0.7\times {{10}^{5}}$ cm⁻³ (the same value of n_c was estimated by Bacmann et al. 2000). Bacmann et al. (2000) has categorized L328 as a spherical core. We estimated an average volume density ($1.2\pm 0.5\times {{10}^{4}}$ cm⁻³) at scales ranging from 0.02 to 0.07 pc (a range over which the NIR polarization data has been used to estimate the magnetic field strength) using analytical approximation for a Bonnor–Ebert sphere given by Tafalla et al. (2002) and estimated a B_pos-field strength of 22 μG. Using the uncertainties in $n({{H}_{2}})$, $\delta v$, and $\delta \theta$ measurements, the uncertainty in the B_pos-field strength is estimated to be ∼10 μG. This is in agreement with the typical uncertainty ($\sim 0.5{{B}_{{\rm pos}}}$) cited for the B_pos-field strength estimations in earlier studies (e.g., Crutcher et al. 2004). The B-field strength in L328, forming a proto-brown dwarf source, is comparable to those estimated toward the other low-mass star-forming cores (e.g., Crutcher 2012). The N₂H⁺ line width toward sub-core S2, which L328-IRS is associated with, is higher by a factor of 1.2 as compared to the line widths found toward sub-cores S1 and S3 (∼0.5 km s⁻¹). If we attribute the higher value of the line width toward S2 to the presence of an outflow component and assume a line width of 0.5 km s⁻¹ similar to those in S1 and S3 (both classified as starless), the B_pos-field strength becomes ∼17 μG, which is well within the error of our B_pos-field strength estimation.

The relative importance of gravity to magnetic fields is given by the mass-to-flux ratio that can be quantified by defining a parameter λ (Crutcher et al. 2004). This parameter is calculated using the hydrogen column density and the B-field strength. Based on the absorption at mid-IR wavelengths, Bacmann et al. (2000) have estimated the column density as a function of angular radius from the center of L328 (the center coinciding with L328-IRS). The advantages of using absorption in the mid-IR are that it is more sensitive to the outer parts as compared to the (sub)millimeter observations and the measurement of column density derived does not assume the dust temperature distribution in the core. The column density profile of L328 is found to steepen beyond a radius of ∼0.03 pc with a slope of 1.7, until it merges with the ambient molecular cloud (Bacmann et al. 2000). Using this as the slope, we estimated the column density as $1.2\times {{10}^{22}}$ cm⁻² at a radial extend of 0.045 pc (mid-point of 0.02–0.07 pc). The value of λ thus estimated is found to be 1.3 ± 0.6, suggesting that the region is marginally supercritical.

Given the uncertainties, especially in the determination of the position angle of the outflow axis from L328-IRS, it is hard to make any conclusive comparisons with the results from the simulations. However, the results obtained on L328-IRS and IRAM 04191-IRS tend to suggest that a possible alignment may exist between the B-field and the outflow directions in cores harboring VeLLOs. More systematic studies of the cores with VeLLOs are required to understand the role played by the B-field in their formation.

We present results from the first B_pos-field mapping observations (using optical and NIR polarimetry) of cloud L328, which is forming a proto-brown dwarf candidate (L328-IRS) in one of its three sub-cores. L328-IRS, a VeLLO, is found to be associated with a sub-parsec-scale molecular outflow. On a parsec scale, the projected B_pos-field is found to follow the curved structure of the cloud that shows a head–tail morphology. At scales of 0.02–0.2 pc, the B_pos-field is found to be smoothly oriented as inferred toward a number of cores with very low mass protostars. The field lines are found to be smoother in shadowed regions of the L328 head part. The projected angular offset between the B $_{{\rm pos}}$-field direction and the outflow orientation is found to be in the range of 0°–70°. The values of the B_pos-field strength and mass-to-flux ratio estimated here are the first such calculations made toward a core forming a proto-brown dwarf candidate. The results obtained in this study tend to suggest that the processes involved in the formation of proto-brown dwarfs are similar to those involved in the formation of low-mass stars.

The authors thank the referee for a very constructive report that has resulted in a significant improvement in the manuscript. We acknowledge the use of NASA's SkyView facility located at NASA Goddard Space Flight Center. C.W.L. was supported by Basic Science Research Program though the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science, and Technology (NRF-2013R1A1A2A10005125) and also by the global research collaboration of Korea Research Council of Fundamental Science & Technology (KRCF). J.K. was supported by Grant-in-Aid for JSPS Fellows (24·110, 26·04023). M.T. was supported by MEXT KAKENHI grant No. 19204018, 22000005. The authors are thankful to P. Bhardwaj for help with writing.