The eccentric dust ring of Fomalhaut: planetary perturbations or gas-dust interactions?

Transcription

1 of : planetary perturbations or? 1,2, Alexis Brandeker 1,2, Göran Olofsson 1,2 1 Department of Astronomy, Stockholm University 2 Stockholm University Astrobiology Centre eccentric / 16

2 Outline eccentric 2 / 16

3 What are? circumstellar, gas-poor disks consisting of dust, comets and leftover planetesimals Kuiper belt / asteroid belt analogues size: from a few AU to >1000 AU dust continuously replenished by collisions several hundred examples known today eccentric Image credit: NASA 3 / 16

4 images. system A&A 540, A125 (2012) Image Pixel a ( ) Integration time(s) SB b SNR b total map per pixel c (mjy/ 2 ) PACS 70 µm PACS 160 µm SPIRE 250 µm SPIRE 350 µm SPIRE 500 µm spectral type: A-star distance: 8 pc age: 400 Myr e of the projected image. (b) Surface brightness and signal-to-noise ratio in the brightest pixel of the image. (c) In the central part the source is located. with semi-major axis of 150 AU 3, 86, 140, 190 and 280 AU at the distance to ata were processed in the Herschel interactive onment (HIPE, Ott 2010) version6.0,applypipeline steps. flux conversion was done f the response calibration. Signal glitches due pacts were masked using the PACS photmmtin HIPE on the detector timeline. n a first projected. A high-pass filter was applied to requency drifts. To avoid the introduction of artee coarse map to mask out the disk from the derior to the application of the filter. scale of er was 15 frames for the blue maps (70 µm), and e red maps (160 µm). detector time-series summed up into a map using the PACS photpixel scale for the 70 µm mapwassetto1, r the 160 µmmapwas2.sourcewascovpendent observations, scanning the sky in two tions. We combined the two detector time series se together into the final maps. RE data were reduced using HIPE. high atio of these data allowed us to created overth pixel scales of 3,4.2 and 7 in the 250 µm, Image credit: Acke et al µm bands respectively,equating to approx- Fig. 1. Herschel PACS 70 µm imageof.positionsofthe star and of the planet candidate are indicated. hatched eccentric 4 / 16

5 tness and signal-to-noise ratio in the brightest pixel of the image. (c) In the central part istance to nteractive.0,applywas done tches due hotmmten a first lied to ren of artem the dee scale of µm), and me-series CS photsetto1, wascovky in two me series eccentric high ted over- Herschel PACS 70 µm Image credit: Acke et al / 16

6 Paul Kalas 1, James R. Graham 1 & Mark Clampin 2 Is a planet Sun and >15 pershaping cent of nearby stars are surrounded the by dusty? disks that must be collisionally replenished by asteroids and comets, as the dust would otherwise be depleted on timescales <10 7 years (ref. 1). oretical studies show that the structure of a dusty disk can be modified by the gravitational influence of planets 2 4, but the observational evidence is incomplete, at least in part because maps of the thermal infrared emission from the disks have low linear resolution (35 AU in the best case 5 ). Optical images provide higher resolution, but the closest examples (AU Mic and b Pic) are edge-on 6,7, preventing the direct measurement of the azimuthal and radial disk structure that is required for fitting theoretical models of planetary perturbations. Here we report the detection of optical light reflected from the dust grains orbiting (HD ). system is inclined 248 away from edge-on, enabling the measurement of disk structure around its entire circumference, at a linear resolution of 0.5 AU. dust is Indications for a planet shaping the (Kalas of planetary-mass et al. objects2005): orbiting. dust emission centre at submillimetre of dust wavelengths with offset by 15 AU from star sharp inner edge distributed in a belt 25 AU wide, with a very sharp inner edge at a radial distance of 133 AU, andwemeasureanoffsetof15au between the belt s geometric centre and. Taken together, the sharp inner edge and offset demonstrate the presence s circumstellar dust was discovered at thermal infrared wavelengths with the Infrared Astronomical Satellite 8,9, and maps of angular resolution suggested a of material residing between 100 and 140 AU radius 10,11. is only 7.7 parsecs (1 pc ¼ 3.3 light years) from the Sun. Using the Advanced Camera for Surveys (ACS) on board the Hubble Space Telescope (HST), we detect its dust complex at optical wavelengths, with angular resolution 100 times sharper thanprevioussubmillimetremaps (Fig. 1; Methods). is surrounded by a narrow dust belt inclined to our line of sight and with significant asymmetry in its azimuthal brightness distribution. We perform a nonlinear least-squares fit to the ellipse search for planetary candidates traced by the brightest points along the belt. semimajor and semiminor axes are ^ 1.8 AU and 57.5 ^ 0.7 AU, respectively. position angle of the ellipse is PA ¼ ^ 0.38, with inclination to the line of sight i ¼ ^ 0.48 (assuming that the structure is intrinsically circular). se values are consistent with the belt properties previously inferred via modelling of submillimetre data 11. geometric centre of the elliptical fit to s belt is offset from the star by 13.4 ^ 1.0 AU at PA ¼ Assuming that the star and belt are coplanar, the projected offset translates to 15.3 AU in the plane of the belt. A keplerian orbit with non-zero eccentricity traces an ellipse where the star is at one focus and the centre of the ellipse is offset by a value, f, equivalent to the semimajor axis, a, times offset of the star relative to the dust belt and an asymmetric scatte phase function produce the azimuthal brightness asymmetries. In the rotated frame shown in Fig. 1, we refer to the upper left quadrant of the belt as quadrant one (Q1), with Q2, Q3 and Q4 Figure 1 Optical detection of dust in s Kuiper belt. a, Falsecolour coronagraphic HST image. (Methods). Image North credit: is 668Kalas clockwiseet fromal. vertical Optical b, Residual image after subtracting a model of a dust belt (Methods). Green lines trace the projected boundaries of the detected belt ( AU). white diamond and asterisk mark the centres of the belt and the star, eccentric 6 / 16

7 Detection of detection of with HST (Kalas et al. 2008, 2011, 2013) eccentric Image credit: NASA and ESA 7 / 16

8 nature of puzzling spectral properties: detected in optical (e.g. Kalas et al. 2013)... but non-detection in mid-ir (e.g. Janson et al. 2012) inconsistent with models of planetary spectra detection in reflected starlight need additional reflecting surface eccentric Image credit: Paul Kalas (UC Berkeley) 8 / 16

9 nature of proposed models: circumplanetary dust disk (Kalas et al. 2008) transient dust cloud (Kalas et al. 2008) swarm of irregular satellites around 2 10 M planet (Kennedy & Wyatt 2011) eccentric Image credit: ESA, NASA, and L. Calcada (ESO for STScI) 9 / 16

10 and the revisited barycenter is located 15.8 kau from A. orbit of highly eccentric ns of the periapse (left) unlikely to be responsible for system may eccentricity be most / morphology of the disk! c? an increasingly comtherefore helpful to elements before as-. Kalas et al Fig. 26. Notional sketch of the system viewed face-on. radial locations of features are ap- eccentric 10 / 16

11 Sharp (eccentric) s through CH alternative without invoking planets: dust affects gas temperature - gas affects dust dynamics operates for stars with spectral type as late as K works for a wide range of dust and gas masses LETTER a 0.3 b c d x/h x/h x/h Lyra & Kuchner t/t orb z/h max( d d0 )/ 0 d / 0 at t = 100 T orb < d >/ 0 y/h t/t orb eccentric 11 / 16

12 photoelectric instability 532 BESLA & WU start with generic dust enhancement (e.g. from collision) gas is primarily heated by photoelectrons from the dust photoelectric instability: dust heats gas by photoelectric effect gas temperature rises gas pressure rises dust concentrates where gas pressure is maximum... Fig. 1. Midplane gas and dust densities in our transitional phase disk. While the gas density (solidbesla line) falls & offwu as a smooth 2007 power-law, the dust density (dotted line) is a Gaussian centered at 70 AU, with a FWHM of 20 AU. gas and dust masses are Mgas ¼ Mdust ¼ 0:1 M and both components have the same vertical scale-height of H ¼ 0:05r. We stress that although we use these densities for our fiducial parameters, the proposed mechanism works over a large range of gas and dust masses (x 5). the electrostatic potential of the grain. value of depends (weakly) on stellar spectrum, local electron density, and gas temperature. It is important not to limit the photon energy range to the FUV range (more on this below). Gas orbiting around a star is subject to the forces of gravity and gas pressure. We again assume that gas moves only on cir- eccentric Fig. 2. Gas heating the transitional disk arou be at local thermal equilib of 100 to distinguish the hancement (70 AU), phot and cooling processes, re nismwithin the vicinityo plotted). Cosmic-rays dom are calculated assuming lo hydrogen; in the absence hydrogen. asymmetri by H2 self-shielding of th Igno collisional within each 12 ds / 16 bin is

13 photoelectric instability HR 4796A (A0) ; ; ; ; ; ; Pic (A5) ; ; ; ; ; ; Star/Spectral Type Distance (pc) N[H 2] N[H] O i (44.1 m) O i (63.2 m) O i (145.5 m) C ii (157.7 m) start with generic dust enhancement (e.g. from collision) gas is primarily heated by photoelectrons from the dust Column Densities a (cm 2 ) Line Fluxes at Earth b (ergs s cm 2 s 1 ) TABLE 2 LTE Predictions for Various Debris Disks column density of H2 will be N½H2Š ¼1:2 ; cm 2 in our model. This is much greater than the observed limit. Does this rule out the presence of a significant amount of gas in the HR 4796A disk? gas thermal timescale at 70 AU (¼ nkbt/ 20 yr for T ¼ 200 K, where is the heating rate) is short compared to the dynamical time (400 yr at 70 AU), so gas rising to high altitudes will fall beyond the influence of the dust and cool rapidly. vertical isothermal assumption is therefore invalid and we do not expect much vertical expansion of the gas disk. Instead, Hgas Hdust 3:5 AU. This reduces the line-of-sight column density to N½H2Š 7 ; cm 2. Furthermore, H2 molecules located well above the disk midplane are destroyed rapidly and the ratio of molecular to atomic species is reduced compared to that at the midplane. We therefore conclude that the observed upper limit on N½H2Š cannot yet exclude the presence of a significant amount of gas in the disk. upper limits on other gases (Chen & Kamp 2004 ) are less constraining than that of H2 and are compatible with our model. Apart from HR 4796A, there is also a stent upper limit on N½H2Š in the Pic disk (Lecavelier des Etangs et al. 2001): N½H2Š < cm 2.Thisdiskisedge-on,sotherearenotthe same complications in interpreting the column densities as for HR 4796A. Most likely there is not a significant amount of gas in the Pic disk (also see Fernandez et al. 2006). In corroboration with this conclusion, the disk does not exhibit a narrow. O i 63.2 m flux. In the transitional disks we study the gas is heated by photoelectrons from the dust grains and cooled mostly by radiation in the O i 63.2 mline.estimatedlinefluxesat the Earth are listed in Table 2. factor of 10 below that reported by Kamp et al. (2003). More investigation is warranted to resolve this discrepancy. All objects should be easily detectable by the PACS instrument in the upcoming Herschel mission.for instance, PACS should be able to achieve a 5 detection of the HR 4796A system in a mere 6 minutes. With a diffraction limit of 8 00, Herschel may even spatially resolve some of the close-by systems. Metallic gas (C, O, Fe, Na,...) has been detected in the Pic disk (Brandeker et al. 2004; Roberge et al. 2006). This gas may be produced du collisions of dust grains and contains little or no hydrogen (Fernandez et al. 2006). We calculate the O i/c ii fluxes expected in such a hydrogen-poor disk and find them to be comparable to those listed in Table 2. This is because the electron density is not significantly reduced when hydrogen is absent from the gas (x 4.1). photoelectric heating rate remains largely unchanged from that of a hydrogen-rich disk. So even though the total gas mass of the disk is very low, there is good reason to expect, and therefore to search for, emission in fine-structure cooling lines. C ii m flux. Detection of the 157.7m Cii line has been reported for the Pic disk (using ISO; Kampetal.2003) with a flux of 1:8 ; ergs cm 2 s 1 (4 detection). This is 30 times greater than our computed value of 6:0 ;10 15 ergs cm 2 s 1 (Table 2) and is surprising. Ionized carbon has been observed in absorption with a column density of N½C iiš 2:5 ; cm 2 (Robergeetal.2006).Without knowing the actual gas/dust masses and distributions, We can obtain an upper limit to the C ii flux by assuming that every C ii observed in absorption resides in the excited state and spontaneously radiates. Assuming a vertical scale height of 0.05 and no selfabsorption, we obtain a flux of 2:4 ; ergs cm 2 s 1 still a photoelectric instability: dust heats gas by photoelectric effect gas temperature rises gas pressure rises dust concentrates where gas pressure is maximum... Fig. 5. Graphical demonstration of the evolution of the, seen in scattered light. surface brightness profiles are taken from Figs. 3 and 4, with the color white indicating brighter regions. initial dust enhancement (left) has a FWHM of 20 AU. gas within the enhancement heats up and concentrates grains into a distribution with a smaller FWHM (10 AU; center) and sharp inner and outer edges. Subsequent evolution of this new distribution produces a bright with a FWHM of 5 AU. We have assumed that small grains are regenerated from larger grains after being initially lost from the system. [See the electronic edition of the Journal for a color version of this figure.] Besla & Wu BESLA & WU Vol. 655 eccentric 12 / 16

14 Search for gas in the search for gas emission within the : C II 158 µm and O I 63 µm: Herschel PACS (non-detections) CO 867 µm: ALMA (non-detection) known structure allows forward modelling approach general strategy: 1 derive upper limits on O I / C II / CO gas luminosity 2 convert into total gas mass using the gas ionisation / thermal balance code ontario eccentric 13 / 16

15 PACS observations eccentric 14 / 16

16 preliminary result for C II rough first analysis indicates that PACS data can give stent upper limits 5σ upper limits: solar abundances: M of gas ( ɛ 10) C overabundant 300 : M of gas ( ɛ 400) eccentric 15 / 16

17 Summary 1 system appears as an increasingly complex planetary system 2 exact nature of is unclear 3 Current constraints on the orbital elements of suggest a different origin of the morphology: c ( with other stars) 4 We use Herschel PACS and ALMA to search for gas in the to test whether could be at work eccentric 16 / 16