HIGH-CONTRAST NEAR-INFRARED POLARIZATION IMAGING OF MWC480

For the vast majority of protoplanetary disks, modeling of the IR spectral energy distribution (SED) has been the principal tool for characterizing the properties of the dust disk. Models for disks derived from such data are non-unique since the viewing geometry (inclination), the assumed grain opacity model, and the presence or absence of structure in the disk can combine to produce similar SEDs. Moreover, disks are not static: ISO and Spitzer studies of T Tauri stars have shown that variability in the 3–4.5 μm range is ubiquitous (Rebull et al. 2011; Kóspál et al. 2012). At least some higher mass analogs to T Tauri stars, the Herbig Ae stars, also exhibit striking variability in the IR SED, which has been interpreted as reflecting changes in the scale height of the dust disk at or near the sublimation radius (Sitko et al. 2008).

For protoplanetary disks where there has been significant dust grain growth and settling, the dust disk may have flattened sufficiently to be partially or largely shadowed (Meeus et al. 2001; Dullemond & Dominik 2004a, 2004b) by material at or near the dust sublimation radius. Under these circumstances, the amount of light reaching the outer disk is expected to be anti-correlated with the IR excess in the 1–10 μm range, originating from material at or near the dust sublimation radius (Acke et al. 2009; Carciofi et al. 2004; Robitaille et al. 2007). Hubble Space Telescope (HST) and ground-based NIR single-epoch studies of Herbig Ae disks associated with IR SEDs interpreted as arising from shadowed disks (Meeus group II) resulted in a high non-detection rate, even for mm-bright, large protoplanetary disks (Grady et al. 2005; Fukagawa et al. 2010), but also in some surprising detections (Grady et al. 2000). Multi-epoch imaging of one of these stars, HD 163296, demonstrated that the illuminated portion of the disk varied with time (Wisniewski et al. 2008). These observations typically lacked simultaneous measurements of the IR SED, which would be needed to test whether the disk illumination is indeed anti-correlated with the near- and mid-IR light.

Another Herbig Ae star with documented near- and mid-IR excess variability is MWC 480 (HD 31648), located at d = 137⁺³¹_{− 21} pc (van Leeuwen 2007). This star is one of the brightest Herbig Ae stars at millimeter wavelengths (Mannings et al. 1997) and has been the subject of detailed studies from the X-ray through the radio regime (Grady et al. 2010; Simon et al. 2000; Piétu et al. 2006, 2007), as well as repeated high-contrast imaging with HST (Augereau et al. 2001; Grady et al. 2001, 2005, 2010) and NIR imaging with Subaru (Fukagawa et al. 2010). HST imaging at the epoch of ISO SWS observations resulted in a marginal scattered-light detection of this disk (Grady et al. 2010) at a time when the IR SED suggested that the IR excess was weaker than average. However, the disk detection rate in scattered light is known to increase as further steps are taken to increase the disk-to-star contrast, such as the use of polarimetric differential imaging, and as the inner working angle of the observations decreases. We present the first firm scattered light detection of the disk of MWC 480 with an inner working angle of 02 (27 AU).

2.1. Subaru/HiCIAO

We carried out H-band (1.6 μm) linear polarimetric observations of MWC 480 using the high-contrast imaging instrument HiCIAO (Tamura et al. 2006) with a dual-beam polarimeter at the Subaru 8.2 m Telescope on 2010 January 24. These observations were part of the Strategic Explorations of Exoplanets and Disks with Subaru (SEEDS; Tamura et al. 2009). The polarimetric observation mode acquires o-ray and e-ray simultaneously, and images a field of view of 10'' × 20'' with a pixel scale of 9.53 mas pixel⁻¹. MWC 480 was observed using a circular occulting mask with 03 diameter. The exposures were performed at four position angles (P.A.s) of the half-wave plate, with a sequence of P.A. = 0°, 45°, 225, and 675 to measure the Stokes parameters. The integration time per wave plate position was 19.5 s and we obtained 24 sets by repeating the cycle of the wave plate. The adaptive optics system (AO188; Hayano et al. 2008) provided a diffraction limited and mostly stable stellar point-spread function (PSF) with an FWHM of 007 in the H band. Low quality images were removed prior to production of the final images, resulting in 18 good data sets with a total integration time for the polarization intensity (hereafter PI) image of 351.0 s. Note that the data sets are taken with Angular Differential Imaging mode, and the field rotation is 32°.

The polarimetric data reduction is described in Hashimoto et al. (2011) and Muto et al. (2012), using the standard approach for polarimetric differential imaging (Hinkley et al. 2009). Instrumental polarization of HiCIAO at the Nasmyth instrument was corrected by following Joos et al. (2008). We calculated PI from $\sqrt{Q^2+U^2}$. This approach mitigates two of the limitations of high-contrast imagery: the PSF observations are obtained simultaneously with the disk data, eliminating the effects of variable seeing, and also providing a PSF measurement which is a perfect color match to the star. This is particularly important for systems where the J − H or H − K colors are redder than any main-sequence candidates. While a PSF reference target was observed immediately after MWC 480, the seeing was gradually changing and the reference target data did not provide a good match to the PSF of the MWC 480 data. As a consequence, we focus on the PI image rather than the PSF subtraction intensity (I) image. The PI image is shown in Figure 1.

Zoom In Zoom Out Reset image size

Non-saturated images of MWC 480 in the H band were obtained immediately after the coronagraphic sequence, using an ND filter that transmits 10% of the light. Similar observations of the photometric calibration star HD 18881 (A0V, J − H = 0.0^m = 0.055, H − K = 0^m) were obtained on 2010 January 23. HD 18881 was selected from the UKIRT bright standard catalog (Hawarden et al. 2001) for the Mauna Kea filter set described by Simons & Tokunaga (2002), Tokunaga et al. (2002), and Tokunaga & Vacca (2005). We derived the zero-point magnitude as 24.8 ± 0.01 from the photometry of the calibration star.

2.2. BASS Spectrophotometry

2.3. SpeX Cross-dispersed Observations

2.4. SpeX Prism Observations

2.5. Near-IR Photometry

3.1. Disk Polarization

The disk of MWC 480 is detected out to a radius of ∼1'' (137 AU) (Figure 1). The central black circle indicates the 015 radius of the occulting mask. We measure an integrated PI from 02 to 20 of 1.2 ± 0.1 mJy, with 0.6 ± 0.1 mJy exterior to 06 in radius. On the same night, the star was at H = 6.74 ± 0.08 mag, resulting in a polarized fractional intensity of 0.08%. In contrast to the NICMOS marginal detection (Grady et al. 2010), the disk is detected at all position angles, except for a dark lane along P.A. ∼ 60 ± 10°. This P.A. is in reasonable agreement with the average disk semiminor axis position angle of 57° ± 22 derived from millimeter interferometry for an average of five CO and HCO⁺ transitions (Piétu et al. 2007) and the P.A. of the jet (Grady et al. 2010). We therefore interpret the dark band as a feature caused by depolarization due to the grain scattering phase function, with no indication of material deficits in the disk (Min et al. 2012; Tanii et al. 2012).

The vector map of the PI disk is shown in Figure 2, superposed on the bright PI disk. The vectors show only polarization angle information, since we have no contemporary I data. The bright inner disk (0.2 ⩽ r ⩽ 0.4) polarization vectors show centro-symmetric features (e.g., a butterfly pattern), indicating that the PI disk is really scattered light from the central star.

Zoom In Zoom Out Reset image size

3.2. Polarization Intensity Radial Profile

In common with other SEEDS studies of young stellar object disks, we measure the disk radial polarized surface brightness (SB) profile along the disk major axis, or the axis of maximum elongation in the HiCIAO data. We extracted a swath 013 wide (2× the PSF FWHM) along the major axis from the center of the coronagraphic spot out to 3'' on both sides of the star. PI brightness was measured in 14 pixel × 1 pixel, 013 × 001 (Figure 3). Both radial profiles are shown in Figure 4. In the profile, r = 015 is the occulting mask edge.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

From 02 to 06 (27–84 AU) we find that the PI radial SB of P.A. = 150° and 30° drops as r^{−1.7 ± 0.1} and r^{−2.0 ± 0.1}, respectively. From 06 to 10 (84–137 AU), the radial SB profile is steeper, r^{−3.0 ± 0.3}. The outer disk radial SB profile is similar to that of HD 169142 from 123 to 200 AU (Fukagawa et al. 2010; Grady et al. 2007), HD 100546 from 10 to 140 AU (Quanz et al. 2011), and SAO 206462 (Muto et al. 2012). The shallow radial SB profile from 02 to 06 is similar to that seen in SAO 206462 in the region of the spiral arms where the polarization fractional intensity is 0.6% (Muto et al. 2012), although no resolved spiral features are detected in the disk of MWC 480 at the resolution of our data.

3.3. Placing the HiCIAO Detection in Context

MWC 480 has a history of correlated near- and mid-IR variability (Sitko et al. 2008) with changes in the amplitude of the IR excess, rather than changes in stellar temperature, which has continued into the period covered by this paper. Individual measurements obtained from 2006 to 2011 are shown, together with the envelope of NIR variations from the literature (shaded) in Figure 5. The HiCIAO data were obtained at H = 6.74 ± 0.08, a historic minimum of NIR brightness for MWC 480. The constant temperature of the IR excess permits us to estimate the dust disk scale from the NIR excess. At the time of the HiCIAO observation, the disk scale height at the sublimation radius was only ∼60% of that seen in the 1998 data, which yielded a marginal detection of the disk. By contrast, Spitzer IRS and BASS observations made in 2004 corresponded to a maximum excess state. The HiCIAO data support an anti-correlation between disk detectability and the NIR excess, consistent with predictions for dust disks where grains have grown and settled to the disk midplane (Sitko et al. 2008; Wisniewski et al. 2008).

Zoom In Zoom Out Reset image size

Despite being the brightest Herbig Ae star at millimeter wavelengths (Mannings et al. 1997), the disk of MWC 480 has proven elusive to imaging in scattered light. Despite multiple attempts with HST (Augereau et al. 2001; Grady et al. 2010) and Subaru/CIAO (Fukagawa et al. 2010), the disk has previously been imaged in scattered light only once, and then with only a marginal detection and incomplete recovery of the disk at all position angles (Grady et al. 2010). The HiCIAO PI image of this disk represents the first detection of the disk in polarized light, and a detection which moreover probes the disk at all position angles, except along the system semiminor axis. The successful detection of the disk relies on two factors. First, the HiCIAO observation was made at a time when the NIR excess was at a historic minimum, providing the most favorable illumination conditions for the outer disk. The second factor is polarimetric differential imaging with simultaneous measurement of the stellar PSF from the unpolarized stellar signal. The PI image was made by combining the Stokes Q and U images. The Stokes Q and U parameters were created by subtracting two split images on the HiCIAO image, which can subtract unpolarized light (PSF) of MWC 480 itself. This approach provides not only a perfect seeing match, but also an exact color match to the star.

Nevertheless, the disk is faint, with a fractional polarized intensity of only ∼0.1%. Despite its relatively old age (7 Myr; Simon et al. 2000), the disk has yet to develop distinct structural features, including divots, rings, or a cavity. Our data do reveal what appears to be a break in the radial polarized intensity power law at 06 (84 AU), with a shallower power-law interior to that point. A similarly shallow power law is seen in the region of the spiral arms for SAO 206462 (Muto et al. 2012). However, the absence of distinct structure in the disk of MWC 480 constrains any spiral features in the disk to have a relative scale height of (H/R) < 0.05, in order for them not to be detectable at the spatial resolution of the HiCIAO data. In tandem with the presence of ongoing accretion, which manifests itself through a soft X-ray spectrum, FUV excess light and mass loss via a jet (Grady et al. 2010), the disk of MWC 480 has a distinctly primordial character.

As befits a dynamically cold disk, MWC 480's dust disk is also geometrically quite thin. Our detection of the disk, in tandem with the previous NICMOS marginal detection, allows us to better constrain where the optically thick surface of the disk lies in the H band relative to the gas disk. Piétu et al. (2007) found from modeling of CO interferometric data a 10 ± 1.1 AU gas disk scale height at 100 AU, corresponding to a disk opening half angle of 6° and a disk inclination of 375. Grady et al. (2010) estimated that the angle subtended by the dust disk rim at the sublimation radius was 22 for the 1998 NICMOS marginal detection of the disk in scattered light. For the HiCIAO data, the star was 0.5 mag fainter at H. If we assume that variation in the NIR excess light scales with changes in the scale height of the dust disk inner rim, then the rim subtended only ∼60% of the scale height seen in 1998 data. This constrains the optically thick dust disk surface to lie above 13. In tandem with the disk scattered light detection, this constrains the optically thick dust disk surface to lie above of 13. The available data therefore suggest that the surface of the dust disk lies between 1.3 ⩽θ ⩽ 22. Adopting the average of this range, the dust disk has a scale height at 100 AU only ∼30% as high as the gas disk (H/R ∼ 0.03), demonstrating that this is a stratified disk. The synoptic data for MWC 480 further suggest that the outer dust disk is only rarely illuminated, in contrast to HD 163296 (Wisniewski et al. 2008; Fukagawa et al. 2010), although quantifying the duty cycle for disk illumination will require a denser grid of photometric and spectroscopic data than are currently available.

A further advantage of the availability of synoptic data in tandem with more specialized observations is that we can start to extract the vertical structure of the disk. By combining the imagery (NICMOS 1.1 μm non-detection data; Grady et al. 2010) with the synoptic data, we find that the 2004 Spitzer IRS detection of polycyclic aromatic hydrocarbon (PAH) emission in MWC 480 (Keller et al. 2008; Acke et al. 2010) corresponded to a high excess state (BASS data), a factor of three above the ISO SWS data (Figure 5). In this configuration, the entire dust disk surface would have been shadowed, indicating that PAH emission either originates from the inner rim of the dust disk or that it must be associated with material in the gas disk at altitude, rather than in close proximity to the outer dust disk (Figure 6). If extended to other data sets, this technique can allow us to understand which disk constituents are routinely exposed to UV radiation and which have less irradiation.

Zoom In Zoom Out Reset image size

We can compare the outer disk properties of MWC 480 with the roughly coeval SAO 206462 and the younger, solar analog LkCa 15, both of which have similar PI measurements using HiCIAO (J. P. Wisniewski 2011, private communication; Thalmann et al. 2010; Muto et al. 2012) and prior HST imagery. Both MWC 480 and SAO 206462 have radial polarized intensity profiles beyond 06 proportional to r⁻³ if the polarization efficiency is constant, suggesting that the outer disks are not highly flared. Model fits to the SED and submillimeter visibility for SAO 206462 indicate a disk scale height at the inner edge of the outer disk of 9.2 AU at 46 AU (Andrews et al. 2011). To be illuminated at 100 AU, the dust disk in that system must lie above 20 AU, a factor of 6.7 times higher than that for MWC 480. As for LkCa 15, the NICMOS data demonstrate that the disk is detected out to 12 (G. Schneider 2011, private communication), with no evidence of shadowing of the outer disk along the system major axis. To meet this illumination condition, assuming a wall height of 7 AU at 50 AU in radius (Andrews et al. 2011), the dust disk of LkCa 15 must extend more than 14 AU above the midplane at 100 AU in radius. The combination of a planet candidate in the disk of LkCa 15 (Kraus & Ireland 2012), and spiral arms which may also be associated with Jovian-mass planets in SAO 206462 (Muto et al. 2012), suggests that the outer disks of these transitional disk systems are dynamically excited by massive planets. The extent to which this process is present may therefore constrain the masses of planets, even for systems where the inner cavity is inaccessible to current imaging technology. However, the primordial character of MWC 480's disk compared to coeval or even younger systems demonstrates that the process of giant planet formation may not be ubiquitous or proceed at the same rate in all systems. Addressing how the planet hosting disks differ from those with no evidence of planet formation will be a major task both for ALMA and for the next generation of high-contrast imagers.

We report the detection of the disk of the Herbig Ae star, MWC 480, in scattered, polarized light from 02 to 10 (28–137 AU) in data taken with HiCIAO at the Subaru 8.2 m telescope in early 2010. At the epoch of the detection, the NIR excess was at a historic low. Given a history of excess amplitude but not temperature changes, this implies that the vertical extent of the inner rim of the dust disk, and by inference the accretion rate onto the star, was at a historic low. An anti-correlation between the scale height of the dust disk near the sublimation radius and the degree of illumination of the outer disk is expected only when the dust disk has settled toward the disk midplane. The data presented here, in tandem with earlier synoptic photometry and spectroscopy and a history of high-contrast imaging observations, provide direct confirmation that Herbig Ae stars with IR excesses characterized by power laws of their H/R are disks where dust grains have grown, settled, and where the outer disk is largely shadowed. Higher cadence synoptic observations are required to quantify the illumination/shadowing duty cycle for MWC 480 and to obtain sufficient data for other Meeus group II disks to establish whether MWC 480 is typical of the shadowed disks. The first steps in comparing the disk geometry of systems like MWC 480 to older, transitional disks have demonstrated that the transitional disks considered have not only wide gaps, but outer disks which are much more vertically extended than seen for MWC 480. As the SEEDS survey continues, we will have the opportunity to obtain data not only for other transitional disks, but also for additional primordial disks, and should be able to test the hypothesis that transitional disks have structural changes that extend over the entire disk, reflecting the presence of giant planets within the gapped region.

This work is partially supported by KAKENHI 22000005 (M.T.), KAKENHI 23103004, 23740151 (M.F.), NSF AST 1008440 (C.A.G.), NSF AST 1009314 (J.P.W.), NASA NNX09AC73G (C.A.G. and M.L.S.), and the IR&D program at The Aerospace Corporation (R.W.R.).