Discovery of extreme asymmetry in the debris disk surrounding HD 15115

TL;DR

This paper reports the first optical and near-infrared scattered light detection of the HD 15115 debris disk, revealing extreme asymmetry likely caused by dynamical perturbations from a nearby star, supporting its membership in the Beta Pic group.

Contribution

First detection of the HD 15115 debris disk in scattered light, revealing extreme asymmetry and suggesting dynamical influence from a nearby star.

Findings

East side of disk extends to ~315 AU

West side of disk extends beyond 550 AU

Disk has blue scattering color similar to AU Mic

Abstract

We report the first scattered light detection of a dusty debris disk surrounding the F2V star HD 15115 using the Hubble Space Telescope in the optical, and Keck adaptive optics in the near-infrared. The most remarkable property of the HD 15115 disk relative to other debris disks is its extreme length asymmetry. The east side of the disk is detected to ~315 AU radius, whereas the west side of the disk has radius >550 AU. We find a blue optical to near-infrared scattered light color relative to the star that indicates grain scattering properties similar to the AU Mic debris disk. The existence of a large debris disk surrounding HD 15115 adds further evidence for membership in the Beta Pic moving group, which was previously argued based on kinematics alone. Here we hypothesize that the extreme disk asymmetry is due to dynamical perturbations from HIP 12545, an M star 0.5 degrees (0.38 pc) east of HD 15115 that shares a common proper motion vector, heliocentric distance, galactic space velocity, and age.

Figures

Tables

Table 1. Table 1 Stellar Properties

	HD 15115	HIP 12545	Ref.
Spectral Type	F2	M0	Hipparcos
$m_V$ (mag)	6.79	10.28	Hipparcos
Mass (M⊙)	1.6	0.5	Astrophys. Quant.
Distance (pc)	${44.78}_{-2.01}^{+2.22}$	${40.54}_{-3.61}^{+4.38}$	Hipparcos
RA (ICRS)	02 26 16.2447	02 41 25.89	Hipparcos
DEC (ICRS)	+06 17 33.188	+05 59 18.41	Hipparcos
${\mu }_{\alpha }$(mas/yr)	$86.09\pm 1.09$	$82.32\pm 4.46$	Hipparcos
${\mu }_{\delta }$ (mas/yr)	$-50.13\pm 0.71$	$-55.13\pm 2.45$	Hipparcos
${\mu }_{\alpha }$(mas/yr)	$87.1\pm 1.2$	$82.3\pm 4.3$	Tycho-2
${\mu }_{\delta }$ (mas/yr)	$-50.9\pm 1.2$	$-55.1\pm 2.7$	Tycho-2
U (km / s)	$-13.2\pm 1.9$	$-14.0\pm 0.5$	a
V (km / s)	$-17.8\pm 1.2$	$-16.7\pm 0.9$	a
W (km / s)	$-6.0\pm 2.3$	$-10.0\pm 0.5$	a

Full text

††slugcomment: Submitted to ApJ Letters December 18, 2006; Accepted April 04, 2007

1. Introduction

Volume-limited, far-infrared surveys of the solar neighborhood suggest that $\sim$15% of main sequence stars have excess thermal emission indicative of circumstellar grains (Aumann, 1985; Backman & Paresce, 1993; Meyer et al., 2007). Direct imaging of dust scattered light reveals the geometry of the grain population relative to the star, which further elucidates the origin of dust. In some cases, a circumstellar nebulosity may be amorphous with asymmetric striated features produced when stellar radiation pressure deflects interstellar dust (Kalas et al., 2002). In other cases, such as $\beta$ Pictoris and Fomalhaut, the geometry of dust is consistent with a circumstellar disk or belt related to the formation of planetesimals (Smith & Terrile, 1984; Kalas et al., 2005). Though larger bodies such as comets and asteroids are not directly observed, they most likely exist as a reservoir for injecting fresh debris into the system over the lifetime of the star. Furthermore, circumstellar debris disks display significant structure and asymmetry that may be linked, in principle, to dynamical perturbations from a planetary system (Roques et al., 1994; Liou & Zook, 1999; Moro-Martín & Malhotra, 2002). Unfortunately, only $\sim$10% of stars with excess thermal emission have detected scattered light disks due to the high contrast between the host star and the low surface brightness nebulosity at optical and near-infrared wavelengths. Fortunately, the observational capabilities have improved in recent years due to instrument upgrades on the Hubble Space Telescope (HST) and the implementation of adaptive optics (AO) on large, ground-based telescopes.

Here we show new scattered light images of a debris disk surrounding HD 15115, an F2 star at 45 pc (Table 1), first reported as a source of thermal excess emission by Silverstone (2000). The spectral energy distribution is consistent with a single temperature dust belt at $\sim$35 AU radius with an estimated dust mass of 0.047 M*⊕* (Zuckerman & Song, 2004; Williams & Andrews, 2006). Recently, Moór et al. (2006) identified HD 15115 as a candidate for membership in the 12 Myr-old $\beta$ Pic moving group (BPMG), based on new radial velocity measurements that resulted in galactic kinematics similar to those of the BPMG.

2. Observations & Data Analysis

We first detected the HD 15115 disk in scattered light using the HST ACS High Resolution Camera (HRC) on 2006 July 17. We used the F606W broadband filter and the 1.8$\arcsec$ diameter occulting spot to artificially eclipse the star. Three flatfielded frames of 700 seconds each from standard pipeline processing of the HST data archive were median combined for cosmic ray rejection. The point spread function was then subtracted iteratively using five other stars of similar spectral type obtained from the HST archive. The relative intensity scaling between images was iteratively adjusted until the residual image showed a mean radial profile equal to zero intensity perpendicular to the circumstellar disk. The images were then corrected for geometric distortion, giving a 25 mas pixel*-1* scale.

The resulting optical images revealed a needle-like feature projecting westward from the star to the edge of the field, but with almost no counterpart to the east. Given the high degree of asymmetry that could conceivably arise from instrumental scattering, we endeavored to confirm the disk using the Keck II telescope with AO on 2006 October 07 and 2007 January 26. Utilizing the near-infrared camera NIRC2, a 0.4$\arcsec$ radius occulting spot and a 10 mas pixel scale, we confirmed the existence of the disk in J (1.2 $\mu$m), H (1.6 $\mu$m), and K′ (2.2 $\mu$m). PSF subtraction is accomplished by allowing the sky to rotate relative to the detector, thereby separating the stellar PSF from the disk. The observing procedure and data reduction procedure are fully described in Fitzgerald et al. (2007). Due to poor observing conditions in October, including intermittent cirrus clouds, we used only the best fraction of data by visually selecting frames of relatively constant intensity and PSF sharpness. The resulting effective integration times are 450 s, 980 s, and 600 s for J, H, and K′, respectively. Standard star observations were obtained under similar, non-photometric conditions and processed in a similar manner. In January 2007 we re-observed HD 15115 (1930 s cumulative integration time) and two standard stars under photometric conditions from Keck using the same instrumentation with the H broadband filter. However, the observations were made after meridian transit and the limited rotation of the sky relative to the instrumental PSF causes disk emission to be included in the PSF estimate, resulting in disk self-subtraction at small radii. Our analysis of the 2007 January data therefore yields a detection of the west ansa in the region 1.3$\arcsec-3.3\arcsec$ radius. The photometry in this second epoch agrees well with that of the first epoch (on average, the 2007 January disk photometry is 0.13 mag arcsec*-2* fainter than 2006 October), suggesting that our frame selection technique for the first epoch of cloudy conditions effectively filtered out non-photometric data.

3. Results

Fig. 1 shows the PSF-subtracted images of HD 15115 with HST and with Keck. The west side of the disk in the optical HST data has PA $=278\arcdeg.5\pm 0.5$ and is detected from the edge of the occulting spot at 1.5$\arcsec$ (67 AU) radius to the edge of the field at 12.38$\arcsec$ (554 AU) radius. The east midplane is detected as far as $\sim$7$\arcsec$ (315 AU) radius. At this radius the east midplane begins to intersect the outer portion of the coronagraph’s 3.0$\arcsec$ occulting spot. Further east, past the spot and to the edge of the field, no nebulosity is detected 9.0$\arcsec-14.9\arcsec$ radius. The appearance of the disk is more symmetric in the 2006 October Keck data, which show the disk between 0.7$\arcsec$ (31 AU) - 2.5$\arcsec$ (112 AU).

Optical surface brightness contours (Fig. 2) reveal a sharp midplane morphology for the west extension that indicates an edge-on orientation to the line of sight. The west midplane is qualitatively similar to that of $\beta$ Pic’s northeast midplane, including a characteristic width asymmetry (Kalas & Jewitt, 1995; Golimowski et al., 2006). The northern side of the west midplane is more vertically extended than the southern side. For example, the full-width at half-maximum across the disk midplane at 2$\arcsec$ radius is 0.19$\arcsec\pm 0.10\arcsec$ in both the optical and NIR data. However, the vertical cuts are not symmetric about the midplane when measuring the half-width at quarter-maximum (HWQM). The HWQM north of the west midplane is $1.6\pm 0.1$ times greater than that of the HWQM south of the west midplane. This width asymmetry is confirmed in the Keck data. If the width asymmetry is found to be in the opposite direction in the opposite midplane, then Kalas & Jewitt (1995) refer to such a feature as the butterfly asymmetry. The butterfly asymmetry is evident in the morphology of $\beta$ Pic, that Golimowski et al. (2006) recently related to the presence of a second disk midplane tilted relative to the main disk midplane. However, our detection of HD 15115’s east midplane has insufficient signal to noise to confirm the presence of a width asymmetry here.

We note that none of the surface brightness profiles show evidence for significant flattening inward toward the star (Fig. 3). All four surface brightness profiles are well-represented by a single power law decrease with radius. If there is an inner dust depletion, then it resides within 40 AU radius. This constraint is consistent with model fits of the spectral energy distribution that place the dominant emitting dust component at $\sim$35 AU radius (Zuckerman & Song, 2004; Williams & Andrews, 2006).

The color of the disk may be estimated in the $2.0\arcsec-3.3\arcsec$ region where the H-band and V-band data overlap (Fig. 3). At face value, $\Sigma_{V}-\Sigma_{H}\approx-0.6$ mag arcsec*-2* at 2$\arcsec$ radius, increasing to $-1.9$ mag arcsec*-2* at 3.3$\arcsec$ radius for the West disk extension. The east ansa has similar blue scattering at 2$\arcsec$ radius, but the V-band surface brightness profile is steeper in the east than in the west, giving a roughly constant blue color with increasing radius in the east.

In a future paper we will present a detailed model of dust scattering and thermal emission properties, which requires a more complicated treatment of the obvious disk asymmetry. However, for isotropically scattering grains in an edge-on disk, an analytic approach shows that the grain number density distribution as a function of radius within the disk midplane follows a power-law with index equivalent to one minus the sky-projected radial midplane power-law index. From the Keck data in Fig. 3, we estimate that the disk number density distribution decreases with disk radius as $r^{-3}$ in the inner region up to $\sim$3.3$\arcsec$ radius for both sides of the disk. At $>3.3\arcsec$ radius, the optical data show that this profile continues for the east extension, but that the disk number density profiles flattens for the west extension, as described in Fig. 3. A precise measurement of the color and polarization of the disk scattered light is necessary to further constrain the grain size distribution, the corresponding scattering phase function and albedo, and the effect on the disk number density profile.

4. Discussion

Asymmetric disk structure is evident in the majority of debris disks, and most authors invoke planetary perturbations as the likely origin. Secular perturbations may offset the center of global disk symmetry from the location of the star, though this effect may also be produced by an external perturber (Wyatt et al., 1999). The edge-on debris disk surrounding $\beta$ Pic displays a variety of radial and vertical asymmetries on large scales (Kalas & Jewitt, 1995) that may be most relevant to the study of HD 15115. In the deepest optical images of the $\beta$ Pic disk, the northeast and southwest disk midplanes are traced to 1835 AU and 1450 AU, respectively, giving a ratio of 1.27 (Larwood & Kalas, 2001). In the case of HD 15115 the corresponding ratio is $>$1.75. This ratio is a lower limit given that the 550 AU extent of the west midplane is limited only by our field of view.

A single stellar flyby, or a periodic flyby by a bound companion on an eccentric orbit, has been studied as a potential mechanism for producing $\beta$ Pic’s large-scale asymmetry (Larwood & Kalas, 2001). However, in a kinematic study of $Hipparcos$-detected stars with published radial velocities, Kalas et al. (2001) did not find any perturbers that approached closer than 0.6 pc of $\beta$ Pic, though the sample was estimated as only 20% complete. In the case of HD 15115, Moór et al. (2006) noted that another $\beta$ Pic moving group member, HIP 12545, is located relatively nearby in sky position.

Table 1 summarizes the observed properties of both stars. Their projected separation is 0.51$\arcdeg$, which translates to 0.38 pc at a mean heliocentric distance of 43 pc. Within the uncertainties, the proper motion vectors, the ($U,V$) galactic space motions, and heliocentric distance are identical. Furthermore, the eastward location of HIP 12545 is in the direction of the truncated side of the HD 15115 debris disk. This geometry is consistent with the dynamical simulation of a disk disrupted by a stellar flyby in Larwood & Kalas (2001). Specifically, in their Fig. 18, the long end of a highly perturbed disk is located in the direction of periastron. The perturber follows a parabolic trajectory such that in a later epoch it is located in the direction opposite of periastron, or in the direction of the truncated side of the disk. Periastron in these models is $\sim$700 AU, with an initial disk radius of $\sim$500 AU. Overall, the ensemble of evidence favors further consideration of HD 15115 and HIP 12545 as a possible wide-separation multiple system with a highly eccentric orbit ($e>0.95$).

If the heliocentric distances are in fact nearly equivalent, then the projected sky separation approximates the true separation. Kalas et al. (2001) discuss the Roche radius, $a_{t}$, of a star as containing the volume within which the stellar potential dominates the Galactic tidal field. Using their Eq. 2 and the stellar mass estimates in our Table 1, we find $a_{t}$ = 1.1 pc and $a_{t}$= 0.7 pc for HD 15115 and HIP 12545, respectively. Therefore, for a small body gravitationally bound to HD 15115, the potential well of HIP 12545 exerts a more significant perturbing force than the Galactic tidal field at the current epoch. This is not the case if we take the $Hipparcos$ parallaxes at face value. These give a line-of-sight separation between the stars of $\sim$4 pc, and we derive a 3-D separation of 5.1$\pm$2.8 pc. We further calculate that closest approach will occur $\sim$1 Myr in the $future$. Therefore, improving the parallax measurements for both stars is a critical task for future work that would examine their possible physical association.

A prediction of the Larwood & Kalas (2001) model is that the perturber may capture disk material, and display a tenuous and highly asymmetric tail of escaping material pointing away from the mother disk. To test the hypothesis that HD 15115 suffered a close encounter with HIP 12545, high-contrast observations of HIP 12545 should reveal circumstellar nebulosity due to captured material. Since this is captured material, the nebulosity may not resemble a disk, and any tail should point away from the mother disk (eastward).

To futher investigate this hypothesis, we examined ACS/HRC coronagraphic observations of HIP 12545 obtained by program GO-10487 (Principal Investigator David Ardila). The observing technique is similar to that described here for HD 15115. After PSF subtraction, we do not detect nebulosity in the vicinity of HIP 12545. Therefore the possibility that the extreme disk asymmetry of HD 15115 is created by dynamical interactions with HIP 12545 does not have further supporting evidence at the present time.

Finally, we note that among the four debris disks imaged in scattered light in the BPMG, the dust appears depleted for HD 15115. The values of dust optical depth in 10*-4* units are given as 24.3$\pm$1.1, 4.9$\pm$0.4, 29.3$\pm$1.6 and 4.0$\pm$0.3 for $\beta$ Pic (A5V), HD 15115 (F2V), HD 181327 (F5.5V) and AU Mic (M2V), respectively (Moór et al., 2006). The factor of $\sim$five smaller optical depth for HD 15115 compared to $\beta$ Pic and HD 181327 suggests a different evolutionary path for the disk. Though a stellar flyby is one possibility, migration and dynamical instabilities within a hypothetical planetary system may also play a role in the rapid diminution of dust parent bodies around HD 15115 (e.g. Morbidelli & Valsecchi, 1997).

5. Summary

Optical and near-infrared coronagraphic images of the F2 star HD 15115 reveal a highly asymmetric debris disk with an edge-on orientation. We describe the morphological and photometric properties of the disk, deferring a detailed model of scattering and thermal emission of grains to future work. The blue scattered light color may indicate grain properties most similar to those of the AU Mic debris disk, where $\Sigma_{V}-\Sigma_{H}\approx-1$ mag arcsec*-2* relative to the star (Fitzgerald et al., 2007), and less like those of $\beta$ Pic, which is predominantly red scattering (Golimowski et al., 2006). A key follow-up measurement would be polarization, which in the case of AU Mic revealed highly porous macroscopic grains (Graham et al., 2007).

With outer optical radius $>$550 AU, HD 15115 possesses the second largest debris disk next to $\beta$ Pic. However, the length asymmetry between its west and east midplanes greatly exceeds that of $\beta$ Pic and other disks. HD 15115 is now the fourth debris disk discovered in scattered light among the $\beta$ Pic moving group members. Future work should test our hypothesis that extreme asymmetries are due to dynamical perturbations from the nearby M star HIP 12545.

Acknowledgements: Support for GO-10896 was provided by NASA through a grant from STScI under NASA contract NAS5-26555.