A Major Asymmetric Dust Trap in a Transition Disk

Science  07 Jun 2013:
Vol. 340, Issue 6137, pp. 1199-1202
DOI: 10.1126/science.1236770

From Dust Grains to Planets

Almost 900 extrasolar planets have been identified, but we still struggle to understand exactly how planets form. Using data from the Atacama Large Millimeter Array, van der Marel et al. (p. 1199; see the Perspective by Armitage) report a highly asymmetric distribution of millimeter-sized grains surrounding a young star. Modeling suggests that these particles—the material from which planets form—are being trapped within a protoplanetary disk by an anticyclonic vortex. Localized concentration of large grains within a protoplanetary disk is thought to be a step in planet formation.


The statistics of discovered exoplanets suggest that planets form efficiently. However, there are fundamental unsolved problems, such as excessive inward drift of particles in protoplanetary disks during planet formation. Recent theories invoke dust traps to overcome this problem. We report the detection of a dust trap in the disk around the star Oph IRS 48 using observations from the Atacama Large Millimeter/submillimeter Array (ALMA). The 0.44-millimeter–wavelength continuum map shows high-contrast crescent-shaped emission on one side of the star, originating from millimeter-sized grains, whereas both the mid-infrared image (micrometer-sized dust) and the gas traced by the carbon monoxide 6-5 rotational line suggest rings centered on the star. The difference in distribution of big grains versus small grains/gas can be modeled with a vortex-shaped dust trap triggered by a companion.

Although the ubiquity of planets is confirmed almost daily by detections of new exoplanets (1), the exact formation mechanism of planetary systems in disks of gas and dust around young stars remains a long-standing problem in astrophysics (2). In the standard core-accretion picture, dust grains must grow from submicrometer sizes to rocky cores ~10 times the mass of Earth (MEarth) within the ~10-million-year lifetime of the circumstellar disk. However, this growth process is stymied by what is usually called the radial drift and fragmentation barrier: Particles of intermediate size [~1 m at 1 astronomical unit (AU) (1 AU = 1.5 × 108 km = distance from Earth to the Sun), or ~1 mm at 50 AU from the star] acquire high drift velocities toward the star with respect to the gas (3, 4). This leads to two major problems for further growth (5): First, high-velocity collisions between particles with different drift velocities cause fragmentation. Second, even if particles avoid this fragmentation, they will rapidly drift inward and thus be lost into the star before they have time to grow to planetesimal size. This radial drift barrier is one of the most persistent issues in planet formation theories. A possible solution is dust trapping in so-called pressure bumps: local pressure maxima where the dust piles up. One example of such a pressure bump is an anticyclonic vortex, which can trap dust particles in the azimuthal direction (610).

Using the Atacama Large Millimeter/submillimeter Array (ALMA), we report a highly asymmetric concentration of millimeter-sized dust grains on one side of the disk of the star Oph IRS 48 in the 0.44-mm (685 GHz) continuum emission (Fig. 1). We argue that this can be understood in the framework of dust trapping in a large anticyclonic vortex in the disk.

Fig. 1 IRS 48 dust and gas observations.

The inclined disk around IRS 48 as observed with ALMA Band 9 observations, centered on the star (white star symbol). The ALMA beam during the observations is 0.32′′ × 0.21′′ and is indicated with a white ellipse in the lower left corner. (A) The 0.44-mm (685 GHz) continuum emission expressed both in flux density and relative to the root mean square (rms) level (σ = 0.82 mJy per beam). The 63 AU radius is indicated by a dashed ellipse. (B) The integrated CO 6-5 emission over the highest velocities in contours (6,12,...,60σCO levels, σCO = 0.34 Jy km s−1): integrated over –3 to 0.8 km s−1 (blue) and 8.3 to 12 km s−1 (red), showing a symmetric gas disk with Keplerian rotation at an inclination i = 50°. The green background shows the 0.44-mm continuum. The position angle is indicated in the upper right corner. (C) The Very Large Telescope Imager and Spectrometer for the mid-infrared (VISIR) 18.7-μm emission in orange contours (36 to 120σVISIR levels in steps of 12σVISIR, σVISIR = 0.2 Jy arc sec−2) and orange colors, overlayed on the 0.44-mm continuum in green colors and the 5σ contour line in green. The VISIR beam size is 0.48′′ in diameter and is indicated with an orange circle in the bottom right corner.

The young A-type star Oph IRS 48 [distance from Earth ~120 parsecs (pc), 1 pc = 3.1 × 1013 km] has a well-studied disk with a large inner cavity (a deficit of dust in the inner disk region), a so-called transition disk. Mid-infrared imaging at 18.7 μm reveals a disk ring in the small dust grain (size ~50 μm) emission at an inclination of ~50°, peaking at 55 AU radius or 0.46 arc sec from the star (11). Spatially resolved observations of the 4.7-μm CO line, tracing 200 to 1000 K gas, show a ring of emission at 30 AU radius and no warm gas in the central cavity (12). This led to the proposal of a large planet clearing its orbital path as a potential cause of the central cavity. Although these observations provide information about the inner disk dynamics, they do not address the bulk cold disk material accessible in the millimeter regime.

The highly asymmetric crescent-shaped dust structure revealed by the 0.44-mm ALMA continuum (Fig. 1) traces emission from millimeter-sized dust grains and is located between 45 and 80 AU (±9 AU) from the star. The azimuthal extent is less than one-third of the ring, with no detected flux at a 3σ level (2.4 mJy per beam) in the northern part (fig. S1). The peak emission has a very high signal-to-noise ratio of ~390, and the contrast with the upper limit on the opposite side of the ring is at least a factor of 130. The complete absence of dust emission in the north of IRS 48 and resulting high contrast make the crescent-shaped feature more extreme than earlier dust asymmetries (10, 13). The spectral slope α of the millimeter fluxes Fν [0.44 mm combined with fluxes at lower frequencies ν (14)] is only 2.67 ± 0.25 (Fν να), suggesting that millimeter-sized grains (15) dominate the 0.44-mm continuum emission. However, the gas traced by the 12CO 6-5 line from the same ALMA data set indicates a Keplerian disk profile characteristic of a gas disk with an inner cavity around the central star (Fig. 1B). 12CO 6-5 emission is detected down to a 20 AU radius, which is consistent with the hot CO ring at 30 AU (14). This indicates that there is indeed still some CO inside the dust hole, with a significant drop of the gas surface density inside of ~25 ± 5 AU. The simultaneous ALMA line and continuum observations leave no doubt about the relative position of gas and dust.

The observations thus indicate that large millimeter-sized grains are distributed in an asymmetric structure, but that the small micrometer-sized grains are spread throughout the ring. To our knowledge, the only known mechanism that could generate this separation in the distribution of the large and small grains is a long-lived gas pressure bump in the radial and azimuthal direction. The reason that dust particles get trapped in pressure bumps is their drift with respect to the gas in the direction of the gas pressure gradient: Embedded Image (3, 4), where Embedded Image and Embedded Image are the dust and gas velocities and p is the pressure. In protoplanetary disks without vortices, this gradient typically points inward, so dust particles experience the above-mentioned rapid radial drift issue. If, however, there exists (for whatever reason) a local maximum of the gas pressure in the disk (i.e., where Embedded Image and Embedded Image), then particles would converge toward this point and remain trapped there (3, 5), avoiding both inward drift and destructive collisions (14). Because small dust particles are strongly coupled to the gas, they will be substantially less concentrated toward the pressure maximum along the azimuthal direction than large particles. Various mechanisms have been proposed that could produce a local pressure maximum in disks; for instance, when there is a "dead zone" (16) or a substellar companion or planet (14, 17) in the disk, hindering accretion. Until recently, however, the presence of such dust pressure traps was purely speculative, because pre-ALMA observations did not have the spatial resolution and sensitivity necessary to constrain the distribution of gas and dust required for testing dust-trapping models (18).

Although vortices in models have an azimuthal gas contrast up to only a factor of a few (16, 19, 20), models predict that even a very minor pressure variation in the gas ring will trap the dust efficiently, leading to a strong lopsided azimuthal asymmetry in the dust ring if the vortex is long-lived. A gas contrast of only 10% can create a dust contrast of 100 for large dust particles in ~105 years (9), so a long-lived azimuthal pressure bump can readily explain the observed high asymmetry in the dust structure of IRS 48. Vortices created by planets have been shown to survive over at least 105 years (8). Even though these vortices tend to diffuse at longer time scales, ~105 years is enough time to create strong dust concentrations that remain even if the vortex disappears. It takes millions of years to even out these dust concentrations completely (9). More generally, vortices are expected to be long-lived if they have an elongation (an aspect ratio of arc length over width) of at least 4 (21). The accumulated dust in IRS 48 has an elongation of ~3.1 (±0.6).

We present a detailed numerical model (14), showing the feasibility of our proposed scenario (Fig. 2). Given the central cavity in the Oph IRS 48 disk, we propose a substellar companion as the cause of the inner cavity, which also creates a long-lived ring of enhanced pressure outside the planetary orbit. The gas densities inside the cavity are decreased, with the level depending on the companion's mass and the disk viscosity (22, 23). The drop of the gas surface density at ~25 ± 5 AU, in combination with the shape and steep radial drop of the millimeter dust emission at 45 AU, suggests that this substellar companion is located between the star and dust trap around ~20 AU and has a mass of at least 10 MJupiter (17). The presented model with these parameters shows that the radial overdensity at the edge of the cavity is Rossby-unstable, leading to the production of an anticyclonic vortex (14). Dust accumulation in this pressure bump results in the spatial separation between the gas and the millimeter dust emission (17, 24). Other than the hole, the high gas velocities are symmetric in the east and west and consistent with Keplerian motion around a 2 MSun star (Fig. 1). Any gas density variation along the azimuthal direction cannot be observed directly in the CO 6-5 observations, because of high line optical depth within the disk and foreground absorption, but C17O data are not inconsistent with a full gas ring (14).

Fig. 2

Model of the dust asymmetry. (A) Simulated ALMA Band 9 dust continuum images at 0.44 mm (685 GHz) for our model using the dust-trapping scenario (9). Details of the model and model parameters are derived based on the IRS 48 properties (14). The high azimuthal contrast is similar to that found in the data. (B) Normalized logarithmic azimuthal cut at the peak emission radius through the dust ring of the observations (black) and the ALMA-simulated model (red), showing the large contrast in the millimeter dust density between the maximum and the opposite side of the ring. The (normalized) 3σ upper limit of the ALMA data is indicated with a dashed line. The data points below the normalized 3σ level have been removed for clarity.

Regardless of the formation mechanism, the ALMA observations clearly show a concentration of dust grains within a small region of the disk. The total measured dust emission corresponds to 9 MEarth, assuming a dust temperature of 60 K. The millimeter observations confirm dust growth up to a maximum grain size of amax = 4 mm. Including larger grains, the dust mass could be a factor of Embedded Image larger (14). The mass is similar to the full-disk dust masses found in other young disks (25). This amount of large dust in a small area will favor grain growth up to ~1 m size until the fragmentation barrier (5). Further growth to planetesimal sizes is possible when additional mechanisms, such as the sweeping-up of small particles by larger seeds and bouncing effects including mass transfer, are considered (26, 27). Because these closely formed planetesimals will scatter and disperse along the ring on short time scales, it is not possible to continue growth and form a planetary core with regular orderly growth models within ~10 million years at this large distance from the star. On the other hand, the dust trap could initiate the formation of a Kuiper Belt around IRS 48, such as that found in our own solar system at comparable radii (28). The dust trap as a "Kuiper Belt object factory" is analogous to a "planet factory" at smaller radii around other stars, where both the fragmentation barrier is higher and the collisional growth is faster (5). Thus, the possibility of dust trapping as the start of core formation could help to explain the observations of massive planets at smaller radii around A-stars such as found in HR 8799 (29) and beta Pictoris (30).

Dust asymmetries have been hinted at in other disks by SubMillimeter Array (SMA) observations (13) and are clearly seen in earlier ALMA observations (10). The low image fidelity of the SMA data (low sensitivity and spatial sampling) and the lower contrast leave room for multiple interpretations, although a relation to vortices has been hinted at (10). In contrast, the ALMA observations of IRS 48, with their unprecedented spatial resolution and sensitivity, show a contrast of at least 130 in the continuum along the ring, with no indications of a highly asymmetric small dust/gas distribution. Alternative scenarios are discussed to be less likely (14). A long-lived azimuthal gas pressure bump triggered by a companion, followed by particle trapping, appears to be the most viable scenario that could produce this. The key feature is the observed separation between big and small dust grains/gas, which is a direct consequence of the dust-trapping model.

Supplementary Materials

Materials and Methods

Figs. S1 to S5

References (3154)

References and Notes

  1. Materials and methods are available as supplementary materials on Science Online.
  2. Acknowledgments: We thank M. Benisty and W. Lyra for useful discussions. This paper makes use of the following ALMA data: ADS/JAO.ALMA no. 2011.0.00635.SSB. ALMA is a partnership of the European Southern Observatory (ESO) (representing its member states), NSF (USA), and National Institutes of Natural Sciences (Japan), together with the National Research Council (Canada) and National Science Council and Academia Sinica Institute of Astronomy and Astrophysics (Taiwan), in cooperation with the Republic of Chile. The Joint ALMA Observatory is operated by the ESO, Associated Universities Inc./National Radio Astronomy Observatory, and National Astronomical Observatory of Japan. The data presented here are archived at, and the full project data (2011.0.00635.SSB) will be publicly available at the ALMA Science Data Archive,
View Abstract

Cited By...


Navigate This Article