Dust Traps and the Formation of Cavities in Transition Discs: A millimetre to sub-millimetre comparison survey
Brodie J. Norfolk, Sarah T. Maddison, Christophe Pinte, Nienke van der Marel, Richard A. Booth, Logan Francis, Jean-François Gonzalez, François Ménard, Chris M. Wright, Gerrit van der Plas, Himanshi Garg
MMNRAS , 1–14 (2015) Preprint 5 February 2021 Compiled using MNRAS L A TEX style file v3.0
Dust Traps and the Formation of Cavities in Transition Discs: Amillimetre to sub-millimetre comparison survey
Brodie J. Norfolk ★ , Sarah T. Maddison , Christophe Pinte , ,Nienke van der Marel , , Richard A. Booth , Logan Francis , , Jean-François Gonzalez ,François Ménard , Chris M. Wright , Gerrit van der Plas , Himanshi Garg Centre for Astrophysics and Supercomputing (CAS), Swinburne University of Technology, Hawthorn, Victoria 3122, Australia Monash Centre for Astrophysics (MoCA) and School of Physics and Astronomy, Monash University, Clayton Vic 3800, Australia Univ. Grenoble Alpes, CNRS, IPAG, F-38000 Grenoble, France Department of Physics & Astronomy, University of Victoria, Victoria, BC V8P 5C2 Herzberg Astronomy & Astrophysics Programs, National Research Council of Canada, 5071 West Saanich Road, Victoria BC V9E 2E7, Canada Institute of Astronomy, University of Cambridge, Madingley Road, Cambridge, CB3 0HA, UK Univ Lyon, Univ Claude Bernard Lyon 1, ENS de Lyon, CNRS, Centre de Recherche Astrophysique de Lyon UMR5574, F-69230, Saint-Genis-Laval, France School of Science, University of New South Wales, PO Box 7916, Canberra, BC 2610, Australia
Accepted XXX. Received YYY; in original form ZZZ
ABSTRACT
The origin of the inner dust cavities observed in transition discs remains unknown. The segre-gation of dust and size of the cavity is expected to vary depending on which clearing mechanismdominates grain evolution. We present the results from the Discs Down Under program, an8.8 mm continuum Australia Telescope Compact Array (ATCA) survey targeting 15 transitiondiscs with large ( (cid:38)
20 au) cavities, and compare the resulting dust emission to Atacama Largemillimetre/sub-millimetre Array (ALMA) observations. Our ATCA observations resolve theinner cavity for 8 of the 14 detected discs. We fit the visibilities and reconstruct 1D radialbrightness models for 10 sources with a S/N > 5 𝜎 . We find that, for sources with a resolvedcavity in both wavebands, the 8.8 mm and sub-mm brightness distributions peak at the sameradius from the star. We suggest that a similar cavity size for 8.8 mm and sub-mm dust grainsis due to a dust trap induced by the presence of a companion. Key words: planet-disc interactions – stars: pre-main sequence – techniques: interferometric
Protoplanetary discs around young stars are the birth places of plan-ets. Of particular interest are the so-called transition discs, which ex-hibit inner cavities in their dust distribution. They were historicallybelieved to be the phase of transition between gaseous protoplane-tary discs and debris discs, but their exact place in the evolutionarytimeline (if any) still remains unknown (Espaillat et al. 2014). Theywere originally identified via a dip in the mid-infrared region oftheir spectral energy distributions (Strom et al. 1989; Wolk & Wal-ter 1996), suggesting a deficit of micron-sized grains in the innerdisc. Sub-millimetre interferometry has confirmed this deficit andresolved the inner cavity in the dust of numerous transition discs forthe largest ( >
20 au) cavities (e.g. Brown et al. 2009; Andrews et al.2011; Francis & van der Marel 2020).The deficit of central emission may result from grain growthdue to the vertical settling, radial drift, and coagulation of grains,resulting in a radial and vertical size sorting, and grain growth to ★ E-mail: [email protected] (SUT) sizes outside current observation capabilities (Tanaka et al. 2005).Another possible explanation is a dust trap; the trapping of dustat local maxima in the gas density (Whipple 1972; Pinilla et al.2012) that can overcome the radial drift barrier (Weidenschilling1977). Dust grains interior to this region accrete onto the host star,depleting the inner disc and forming a cavity. A number of physicalmechanisms can produce dust traps in transition discs, includingphotoevaporation, dynamical clearing from a companion, and dead-zones (Williams & Cieza 2011; Pinilla et al. 2016).As the accretion rate decreases, high-energy photons can im-pinge the accretion flow and photoevaporate the disc progressivelyfurther away. This limits the resupply of inner disc material andincreases the local gas pressure at the photoevaporation front, trap-ping dust at the inner edge of the disc (Hollenbach et al. 1994;Clarke et al. 2001; Alexander et al. 2006). Numerous photoevapo-ration models (Owen & Clarke 2012; Gorti et al. 2015; Owen et al.2017; Ercolano & Pascucci 2017; Ercolano et al. 2018) suggestthat discs with an inner deficit of emission due to photoevaporativewinds will have low accretion rates ( ≤ − M (cid:12) yr − ) and exhibitcavities with small radii ( ≤
10 au, see Figure 5 in Owen & Clarke © a r X i v : . [ a s t r o - ph . E P ] F e b Norfolk et al. (2012), or ≤
20 au, see Figure 6 in Ercolano & Pascucci (2017)).This is inconsistent with (sub)millimetre observations of transitiondiscs, exhibiting cavities of tens of au in size and high accretionrates (Manara et al. 2014).Dust traps can also result from the dynamical clearing of acompanion. The formation of a planetary mass companion willlocally deplete gas and/or dust in the disc and produce a pressurebump at the cavity edge (e.g. Paardekooper & Mellema 2004; Riceet al. 2006; Fouchet et al. 2007, 2010; Pinilla et al. 2012), whilethe tidal truncation of the disc by a binary companion (e.g. Lin &Papaloizou 1979) results in a gas pressure bump at the leading edgeof the truncation, consequentially trapping dust and forming a cavityas the inner disc accretes onto the host-star at viscous timescales( (cid:54) years) (Brauer et al. 2007). Dust traps formed by the presenceof a companion are expected to result in a grain size distributionthat mirrors the profile of the gas pressure at the maxima with largergrains strongly concentrating at the peak of the pressure profile, andsmaller grain exhibiting a more extended radial structure. This isdue to the relationship between the grains Stokes number and itssize, as larger grains with St ∼ < Recent high-resolution ( (cid:54) . (cid:48)(cid:48) ) observations of transition discswith ALMA have provided a growing catalogue of discs observedat sub-mm wavelengths which show resolved dust cavities. OurDiscs Down Under survey was designed to specifically complement this catalogue at 8.8 mm for southern hemisphere sources that areobservable by the ATCA.Our sample (see Table 1) includes four Herbig Ae stars(HD34282, HD100453, HD135344B, HD169142), and eleven TTauri stars (SZ Cha, CS Cha, HP Cha, HD143006, RY Lup,J1604, J160830.7, Sz111, SR24S, SR21, DoAr44). Three of theseare located in Lupus (RY Lup, J16083070, & Sz111), three inChamaeleon (SZ Cha, CS Cha, & HP Cha) and three in Ophiuchus(SR 24S, SR 21 and DoAr44). One target (J1604) is a member of theUpper Scorpius Association. Three of the targets have previouslybeen observed with ATCA in the 7 mm band at low resolution andwere unresolved (Sz111 and RY Lup, Lommen et al. 2010; CS Cha,Lommen et al. 2009), and one with the VLA at 7 mm (HD169142,Dent et al. 2006).This survey also targeted the protoplanetary disc HD163296,which is not a transition disc and hence is excluded from the analysis.Observational results for this source are presented in Appendix A. We used the ATCA radio telescope at 34 GHz (8.8 mm) to conductour survey from 2016 to 2018 (project code C3119). The CompactArray Broadband Backend (CABB) (Wilson et al. 2011) providesobservations with two bands that contain 2048 × . (cid:48)(cid:48) .The synthesised beam for each observation is detailed in Table 3.The astrometric accuracy of ATCA observations is primarily de-termined by the atmospheric conditions and the properties of thephase calibrators. The weather during the observations varied foreach science target, and the seeing monitor RMS path length noisefor each observation is summarised in Table 2.The science targets were observed with a sequence of 10 minon-source integration and 2 min integration of the gain/phase cali-brator. The bandpass and flux calibrators were observed for ∼
15 minand pointing checks were made on the phase calibrator every ∼ . Briefly, this involved: correcting for thefrequency-dependent gain using the miriad task mfcal ; then usingthe flux density of the ATCA primary flux calibrator, 1937-638, tore-scale the visibilities measured by the correlator using the miriadtask mfcal with the option nopassol set; and correcting for the gainof the system’s time variable properties due to changing conditionsusing the miriad task gpcal . To reduce the noise in our data whilemaintaining as complete an observational track as possible, weflagged all data with a seeing monitor RMS path length noise above400 𝜇 m using uvflag , and calibrator amplitude readings that deviatedmore than 10% from the mean flux using blflag . Any unusual spikesseen in the channel vs. amplitude or the channel vs. phase plots werealso flagged using uvflag .We extracted the ATCA fluxes by first concatenating the33 GHz and 35 GHz observations using the uvcat miriad task, sub-sequently producing observations centred at 34 GHz. We then usedthe fits miriad task to output the resulting uv data-set as a UVFITS MNRAS000
15 minand pointing checks were made on the phase calibrator every ∼ . Briefly, this involved: correcting for thefrequency-dependent gain using the miriad task mfcal ; then usingthe flux density of the ATCA primary flux calibrator, 1937-638, tore-scale the visibilities measured by the correlator using the miriadtask mfcal with the option nopassol set; and correcting for the gainof the system’s time variable properties due to changing conditionsusing the miriad task gpcal . To reduce the noise in our data whilemaintaining as complete an observational track as possible, weflagged all data with a seeing monitor RMS path length noise above400 𝜇 m using uvflag , and calibrator amplitude readings that deviatedmore than 10% from the mean flux using blflag . Any unusual spikesseen in the channel vs. amplitude or the channel vs. phase plots werealso flagged using uvflag .We extracted the ATCA fluxes by first concatenating the33 GHz and 35 GHz observations using the uvcat miriad task, sub-sequently producing observations centred at 34 GHz. We then usedthe fits miriad task to output the resulting uv data-set as a UVFITS MNRAS000 , 1–14 (2015) ust Traps and the Formation of Cavities in TDs Table 1.
Discs Down Under survey source list. Distances for our sample are taken from Gaia Collaboration (2018), except for HP Cha which is taken fromWhittet et al. (1997). We extract the mass accretion rates of SZ Cha from Manara et al. (2019), HD143006 from Manara et al. (2020), J1608 and Sz111 fromAlcalá et al. (2017), and the remaining sources from Francis & van der Marel (2020).
Target RA (ICRS) Dec. (ICRS) Distance (pc) ATCA Observing Date ALMA ID (cid:164) 𝑀 (log ( M (cid:12) / yr ) )HD 34282 05 16 00.48 -09 48 35.4 325 10/06/2018 2015.1.00192.S, 2017.1.01578.S < − . − . − . − . < − . − . − . − . − . − . − . − . − . − . − . Table 2.
ATCA observing log.
Date Obs.Time(min) Calibrators Ave.seeingrms ( 𝜇 m)Band-pass Flux Gain/Phase22/04/16 225 1253-055 1934-638 1606-39 38023/04/16 225 1253-055 1934-638 1622-253 38824/04/16 430 1253-055 1934-638 1622-253 24825/04/16 450 1253-055 1934-638 1622-253 15229/04/16 430 0537-441 1934-638 1057-797 20203/05/16 340 0537-441 1934-638 1057-797 30804/05/16 450 0537-441 1934-638 1129-58 9506/05/16 450 0537-441 1934-638 1606-39 15507/05/16 450 1253-055 1934-638 1606-39 17013/05/16 450 1253-055 1934-638 1451-400 9915/05/16 450 1253-055 1934-638 1804-251 23517/05/16 340 1253-055 1934-638 1622-310 30004/05/17 430 1253-055 1934-638 1622-253 18505/05/17 450 1253-055 1934-638 1606-39 11406/05/17 430 1253-055 1934-638 1606-39 13912/05/17 340 1253-055 1934-638 1622-253 23613/05/17 430 1253-055 1934-638 1606-39 19622/05/17 320 0537-441 1934-638 1057-797 31310/06/18 450 1921-293 1934-638 0514-161 18412/06/18 430 1253-055 1934-638 1817-254 218 file and binned the visibilities using python. The total integratedfluxes were taken at zero spacing (Table 3).After calibrating the data, images at 34 GHz were producedusing robust weighting of 0.5 for uniform weighting or a sup valueof 0 for natural weighting with the invert task. The sup parameterrefers to the area around the source (in arcseconds) where invert will attempt to suppress side-lobes. The dirty images were cleanedto 5 𝜎 (5 times the RMS noise level) using the clean task and thebeam was restored using the restor task. The resulting images for 13of the 14 detected sources are presented in Figure 1. The relativelypoor seeing during the 3 May 2016 synthesis track of SZ Cha,coupled with the phase decorrelations in the visibilities, requireddata flagged that resulted in an unresolved image. We thereforeexclude the image of SZ Cha and only present the flux determinedduring the period of best seeing in Table 3. Additionally, the 10 June2018 synthesis track resulted in a non-detection of HD34282, likely a product of the low peak brightness due to its distance. As a resultwe only present the 3 𝜎 upper limit in Table 3. The ALMA data for each source was selected from archival obser-vations to both resolve the inner cavity of each transition disc andrecover the largest spatial scales in the outer disc. The final calibratedvisibilities for HD143006 were sourced from the DSHARP survey(Andrews et al. 2018). For all other discs, the highest resolutiondata was in ALMA bands 6 and 7 (1.3 and 0.88 mm, respectively),except for SR 24S, which has higher resolution observations in band3 (3 mm). All ALMA data were reduced using the CASA pipelinefor the appropriate ALMA cycle. Spectral lines in the ALMA datawere flagged before extracting the continuum, which was time av-eraged to 30.5 s and a single channel per spectral window. ALMAvisibilities were extracted using the exportuvfits in CASA and sub-sequently dealt with in python, with the total integrated fluxes takenat zero spacing listed in Table 3.
We quantify the relative position of the emission in the ATCA andALMA observations using the code Frankenstein (Jennings et al.2020) to reconstruct the 1D radial brightness profiles of the disc ateach wavelength. Frankenstein is an open source code that uses aGaussian process to reconstruct the 1D radial brightness profile of adisc non-parametrically and can in principle measure radial featuressmaller than the clean beam size. We use an azimuthally symmetricbrightness profile model despite known asymmetries in our sampleas we are solely interested in obtaining the typical cavity radius forthis exercise. This is especially true for ATCA observations withelongated beams, where the poor uv-coverage effectively results indifferent sensitivity and bias at baselines ≥ 𝜆 and a 1D fit ofthe north-south/east-west elongation. While we would ideally preferto construct 2D models, such modelling is not feasible due to thelow resolution of our ATCA observations.Frankenstein infers the brightness at a set of N radial pointsgiven a disc outer radius (R max ) and assuming azimuthal symmetry.It then fits the profile to observed visibilities by using the discrete MNRAS , 1–14 (2015)
Norfolk et al.
Table 3.
Discs Down Under survey results of ATCA and ALMA continuum observations.
Target ATCA34 GHz flux(mJy) ATCA1 𝜎 rms(mJy) ATCA 𝜃 beam (as × as) ALMAfreq.(GHz) ALMAflux(mJy) ALMA1 𝜎 rms(mJy) ALMA 𝜃 beam (as × as) Spec.Slope( 𝛼 mm )HD34282 <0.290 𝑎 Notes: 𝑎 Upper limits are 3 𝜎 . Hankel transform (Baddour & Chouinard 2015) to relate the ob-served visibilities to the radial brightness profile, applying a non-parametric Gaussian Process (GP) prior. The GP prior is learnedfrom the data given two hyperparameters, 𝛼 and w smooth , and actsto dampen power in the reconstructed brightness profile on scaleswhere the signal-to-noise in the visibilities is low. The most per-tinent parameter is 𝛼 , which controls the signal-to-noise thresholdbelow which Frankenstein does not attempt to fit the data. w smooth introduces a coupling between adjacent points and prevents regionsof artificially low power arising from narrow gaps in the visibil-ities. For more details see Jennings et al. (2020). The brightnessreconstruction is not overly sensitive to variations in w smooth . Forour sample we set N = max (cid:54) (cid:48)(cid:48) , w smooth to either 0.01 or0.0001, and vary 𝛼 between 1.01 and 1.5 and then select the modelwith the lowest 𝜒 value. However, if the data was sufficiently noisyat long baselines and the model was over-fitting, 𝛼 was increaseduntil oscillations in the fit at long baselines disappeared. We con-strained our fits to be non-negative. We include Figure B1 as anexample of the fitting methodology using a variety of hyperpara-maters with the resulting 𝜒 values. Discs are deprojected using theinclination and position angle from high-resolution observations(see Table 4). We restrict our modelling to sources with 𝜎 ≥
5, lim-iting the number of sources in our sample that we model to CS Cha,HPCha, HD100453, HD135344B, RY Lup, J1608, Sz111, SR24S,DoAr44, and HD169142.
We detected 14 out of the 15 discs from the survey with ATCA.HD34282 was the only non-detection and we present the 3 𝜎 fluxupper limit in Table 3. We spatially resolved the continuum emis-sion of disc-like structures for 13 sources and show the CLEANedmaps in Figure 1. A deficit of emission is visible in the inner re-gions of HD100453, HD135344B, RY Lup, J1608, Sz111, SR24S,DoAr44, and HD169142. This can be representative of a ring-likestructure in the 8.8 mm grains of these discs assuming the bulk of the emission is from thermal dust. J1604 and SR21 also appear toexhibit a deficit of central emission, but due to the low SNR for bothobservations this is uncertain. The structure seen in the emission ofHD135344B, RY Lup, J1604, J1608, and SR24S hints at asymmet-rical dust structure which is expected to become more pronouncedat longer wavelengths (Birnstiel et al. 2013; van der Marel et al.2020). However for RY Lup this morphology is due to the discbeing approximately edge-on (Ansdell et al. 2016; Langlois et al.2018). The possible asymmetries in HD135344B and J1604 can notbe confirmed due to the significant beam elongation, but it seemslikely for HD135344B given the dust asymmetry in ALMA (vander Marel et al. 2016b). The observations of HP Cha, HD135344B,HD143006, J1604, and SR21 all suffered from a variety of poorweather conditions that required significant data flagging, resultingin sparse uv-coverage and low SNR. This makes it difficult to di-rectly compare disc structures in the ATCA and ALMA continuummaps for these sources. In extended configurations of the 12-m array, ALMA has a reso-lutions range from 0 . (cid:48)(cid:48) at 230 GHz to 0 . (cid:48)(cid:48) at 110 GHz. Tofacilitate comparison between the ATCA and ALMA observations,we degrade the resolution of the ALMA observations to match thatof the corresponding ATCA map using the restoringbeam option inthe tclean task in CASA (c.f. column 4 of Table 3 and column 3 ofFigure 1).The degraded ALMA continuum maps generally present mor-phology similar to the ATCA continuum maps. Two symmetricintensity peaks can be seen in the degraded ALMA images thatalign, within reasonable radial positional accuracy, for HD100453,RY Lup, J1608, Sz111, SR21, DoAr44, and HD169142. The de-graded ALMA maps for HD143006 and J1604 show morphologyindicative of an inner cavity and ring-like structure. However, this isnot reflected in the corresponding ATCA observations due to eitherthe poor observing conditions or the lack of this structure in the8.8 mm emission. For CS Cha and HP Cha, the degraded contin-uum maps exhibit elliptical structures similar to the corresponding MNRAS000
We detected 14 out of the 15 discs from the survey with ATCA.HD34282 was the only non-detection and we present the 3 𝜎 fluxupper limit in Table 3. We spatially resolved the continuum emis-sion of disc-like structures for 13 sources and show the CLEANedmaps in Figure 1. A deficit of emission is visible in the inner re-gions of HD100453, HD135344B, RY Lup, J1608, Sz111, SR24S,DoAr44, and HD169142. This can be representative of a ring-likestructure in the 8.8 mm grains of these discs assuming the bulk of the emission is from thermal dust. J1604 and SR21 also appear toexhibit a deficit of central emission, but due to the low SNR for bothobservations this is uncertain. The structure seen in the emission ofHD135344B, RY Lup, J1604, J1608, and SR24S hints at asymmet-rical dust structure which is expected to become more pronouncedat longer wavelengths (Birnstiel et al. 2013; van der Marel et al.2020). However for RY Lup this morphology is due to the discbeing approximately edge-on (Ansdell et al. 2016; Langlois et al.2018). The possible asymmetries in HD135344B and J1604 can notbe confirmed due to the significant beam elongation, but it seemslikely for HD135344B given the dust asymmetry in ALMA (vander Marel et al. 2016b). The observations of HP Cha, HD135344B,HD143006, J1604, and SR21 all suffered from a variety of poorweather conditions that required significant data flagging, resultingin sparse uv-coverage and low SNR. This makes it difficult to di-rectly compare disc structures in the ATCA and ALMA continuummaps for these sources. In extended configurations of the 12-m array, ALMA has a reso-lutions range from 0 . (cid:48)(cid:48) at 230 GHz to 0 . (cid:48)(cid:48) at 110 GHz. Tofacilitate comparison between the ATCA and ALMA observations,we degrade the resolution of the ALMA observations to match thatof the corresponding ATCA map using the restoringbeam option inthe tclean task in CASA (c.f. column 4 of Table 3 and column 3 ofFigure 1).The degraded ALMA continuum maps generally present mor-phology similar to the ATCA continuum maps. Two symmetricintensity peaks can be seen in the degraded ALMA images thatalign, within reasonable radial positional accuracy, for HD100453,RY Lup, J1608, Sz111, SR21, DoAr44, and HD169142. The de-graded ALMA maps for HD143006 and J1604 show morphologyindicative of an inner cavity and ring-like structure. However, this isnot reflected in the corresponding ATCA observations due to eitherthe poor observing conditions or the lack of this structure in the8.8 mm emission. For CS Cha and HP Cha, the degraded contin-uum maps exhibit elliptical structures similar to the corresponding MNRAS000 , 1–14 (2015) ust Traps and the Formation of Cavities in TDs ATCA observations. This suggests that any structure at 8.8 mm is notresolved by the clean reconstructed images. HD135344B exhibitstwo asymmetric intensity peaks indicative of its true asymmetricstructure.
We estimate the phase centre of the ATCA observations wherepossible by centering the image to the apparent centre in the ring,which is taken as the central deficit of emission, for about half of thesample (HD100453, HD135344B, RY Lup, J1608, Sz111, SR24S,DoAr44, and HD169142). For CS Cha, the phase centre is taken asthe stellar position correct by Gaia proper motions. HD135344B isknown to be asymmetric (van der Marel et al. 2016b), the positionof central emission deficit closely corresponds to the orbital radiusof the vortex derived by Cazzoletti et al. (2018) (81 au). For HP Chathis comparison could not be made as it was too faint for Gaiato measure its proper motion. These phase centres are consistent( ∼ . (cid:48)(cid:48) ) with the proper motion corrected Gaia positions and theaccuracy is likely a function of the beam size, which can be worse inthe north-south/east-west direction for sources with extended beams(e.g. the poor north-south accuracy for DoAr44). Offsets are similarto the astrometric accuracy seen in previous ATCA observations(Wright et al. 2015).The typical offsets of the phase centre ( ∼ . (cid:48)(cid:48) ) are compara-ble to the cavity size of our transition discs. This directly impactsthe Frankenstein 1D radial brightness models, in particular forsources where the 8.8 mm continuum map does not resolve an innercavity with a ring of emission but exhibits asymmetrical structure.This means that the sources that appear centrally peaked rather thanring-like (CS Cha and HP Cha) might be asymmetric as well at8.8 mm. Figure C1 highlights the effects on the 1D brightness mod-els of RY Lup when moving the phase centre, and cautions the inter-pretation of these models for sources that are possibly asymmetric(HP Cha, HD135344B, RY Lup, J1608, SR24S, and DoAr44) withrelatively large beam sizes ( (cid:38) . (cid:48)(cid:48) ). We present the real component of the deprojected ATCA and ALMAvisibilities in Figure 2, adopting the inclination and position an-gle for each source from high-resolution observations (see Table4). The imaginary components of the visibilities are presented inFigure D1. We also present in Figure 2 the Frankenstein recon-structed brightness profiles for the majority (10/15) of our sampleand Table 4 includes the radial offsets for the peak (col.2 and 3),R peak of the Frankenstein fits. Errors in the Frankenstein fit forsources with resolved features (e.g. a Gaussian peak) are calculatedvia the bootstrap method. This involves fitting the disc geometryas defined by the inclination and position angle, and calculatingthe model numerous times using different sub-samples of the data.The bootstrap method effectively tests the ’goodness’ of a fits cho-sen geometry and hyperparameters. As a result, if the data has arelatively high SNR and the choice of hyper-parameters is correct,sub-samples of the data will always reproduce approximately thesame fit. This is apparent in our ALMA visibility fits where the 1 𝜎 error is ∼ .
4% of the R peak values. For further details on the boot-strap method in Frankenstein see Jennings et al. (2020). We takethe uncertainty in the R peak values as the combination of the 1 𝜎 standard deviation of our bootstrap models and the full width half maximum of the observational beam size. For sources with unre-solved features, calculating R peak and the fwhm is not possible, andthe results for these sources are omitted from Table 4. We includeradial offsets from previous visibility fits in the literature and notethat our values are comparable (see Table 4, col.4). The visibilitiesfor ATCA observations that were not modelled with Frankenstein(HD34282, SZ Cha, HD143006, J1604, and SR21) and the corre-sponding ALMA visibilities with Frankenstein fits are includedin Appendix E.Our multi-wavelength visibility comparison shown in Figure2 suggests that the 8.8 mm and sub-mm disc structures share asimilar cavity size for HD100453, HD135344B, RY Lup, J1608,Sz111, SR24S, DoAr44, HD169142. This is indicated by the similarlocation of the null seen in the visibility data and similar radialpositions of the Gaussian ring peak seen in the Frankensteinreconstructed 1D brightness profiles. For CS Cha and HP Cha the1D radial brightness profiles suggest that the 8.8 mm and sub-mmdo not share a similar cavity size. For these sources, it is clear thatthe noise is larger than the correlated flux of the negative part in theALMA data. It is unlikely that the Frankenstein fit for low SNRat long baselines could resolve a cavity if one is indeed present. TheATCA 1D radial brightness profiles of HD100453, HD135344B,RY Lup, J1608, Sz111, SR24S, and DoAr44 show an inner deficit ofemission consistent with Figure 1. For the ATCA data of HD169142the Frankenstein fit is able to reproduce the two rings seen in theALMA image that is not resolved in the ATCA image in Figure1, this is attributed to Frankenstein’s ability to resolve angularscales smaller than the uniform-weighted clean beam for sourceswith relatively low SNR at long baselines. All of the ALMA radialbrightness profiles indicate a ring-like structure, as also shown inFigure 1. In our multi-wavelength comparison we have assumed that the8.8 mm emission is primarily from thermal dust, which shouldbe tested. Figure 3 shows the spectral slope between 0.8 mm and1 cm for sources with sufficient flux measurements between 0.8–3.3 mm for a linear least-squares fit (in log-log space). We include acalibration uncertainty of 10% for ALMA data and 20% for ATCAdata. Figure 3 shows that for the majority of our sources the 8.8 mmemission falls within the fitted error, indicating that the emission islikely due to thermal dust emission from large grains. For SZ Chaand RY Lup, we see excess emission above that of thermal dust,which is likely due to free-free emission. The lack of publicly avail-able ALMA Band 3 data for J1604 and DoAr44 result in large errorbands in the spectral slope. In Table 3 we include the spectral slope( 𝛼 mm ) calculated from our linear least-squares fit. The spectral in-dices for our sample range from 2 . − .
63. This is consistent witha distribution of large grains (Natta & Testi 2004; Draine 2006).
Continuum emission at millimetre wavelengths from discs aroundyoung stars often show excess emission above that expected from asimple extrapolation of thermal dust emission observed at shortermm wavelengths (Rodmann et al. 2006; Lommen et al. 2009; Ubachet al. 2012, 2017). This excess has been attributed to thermal free-free emission from an ionised wind, non-thermal processes such as
MNRAS , 1–14 (2015)
Norfolk et al. -0.06 -0.04 -0.02 0.00 0.02 0.04 0.06 0.07 0.09 0.11 mJy/beam mJy/beam mJy/beam mJy/beam -0.04 -0.02 0.00 0.02 0.04 0.05 0.07 0.09 0.11 mJy/beam -0.09 0.00 0.09 0.17 0.26 0.34 0.43 0.52 0.60 0.69 mJy/beam mJy/beam mJy/beam -0.06 -0.04 -0.02 0.00 0.02 0.04 0.06 0.08 0.10 0.12 mJy/beam mJy/beam -4.80 0.00 4.80 9.60 14.4019.2024.0028.8033.6038.40 mJy/beam -4.80 0.00 4.80 9.60 14.4019.2024.0028.8033.6038.40 mJy/beam -0.04 -0.02 -0.01 -0.00 0.01 0.02 0.04 0.05 0.06 0.07 mJy/beam -0.00 5.94 11.8817.8223.7629.7035.6441.5847.5253.46 mJy/beam mJy/beam mJy/beam -0.06 -0.05 -0.03 -0.02 -0.01 0.00 0.01 0.02 0.03 0.05 mJy/beam mJy/beam mJy/beam mJy/beam -0.03-0.02-0.01 0.00 0.01 0.02 0.03 0.04 0.05 0.06 mJy/beam mJy/beam -0.00 5.00 10.0015.0020.0025.0030.0035.0040.0045.00 mJy/beam -0.00 5.00 10.0015.0020.0025.0030.0035.0040.0045.00 mJy/beam -0.04 -0.03 -0.02 -0.01 -0.00 0.01 0.02 0.03 0.04 0.05 mJy/beam -0.00 0.49 0.99 1.48 1.98 2.47 2.97 3.46 3.96 4.46 mJy/beam mJy/beam mJy/beam
Figure 1.
The 13 spatially resolved discs from the Discs Down Under survey. The 2 (cid:48)(cid:48) 𝑥 (cid:48)(cid:48) maps show the 8.8 mm continuum ATCA maps, ALMA continuummaps, degraded ALMA continuum maps, and the ALMA continuum maps overlayed with ATCA continuum contours. The contours as printed at the bottomof each continuum map are a factor of the 1 𝜎 rms. MNRAS000
The 13 spatially resolved discs from the Discs Down Under survey. The 2 (cid:48)(cid:48) 𝑥 (cid:48)(cid:48) maps show the 8.8 mm continuum ATCA maps, ALMA continuummaps, degraded ALMA continuum maps, and the ALMA continuum maps overlayed with ATCA continuum contours. The contours as printed at the bottomof each continuum map are a factor of the 1 𝜎 rms. MNRAS000 , 1–14 (2015) ust Traps and the Formation of Cavities in TDs -0.03-0.02-0.01-0.00 0.01 0.02 0.03 0.05 0.06 0.07 mJy/beam mJy/beam -0.00 4.30 8.60 12.9017.2021.5025.8030.1034.4038.70 mJy/beam -0.00 4.30 8.60 12.9017.2021.5025.8030.1034.4038.70 mJy/beam -0.03-0.02-0.01 0.00 0.01 0.02 0.03 0.04 0.06 0.07 mJy/beam mJy/beam mJy/beam mJy/beam -0.07-0.06-0.04-0.02 0.00 0.02 0.04 0.06 0.07 0.09 mJy/beam mJy/beam mJy/beam mJy/beam -0.05-0.04-0.03-0.02-0.01 0.00 0.01 0.02 0.03 0.04 mJy/beam mJy/beam mJy/beam mJy/beam -0.05-0.04-0.03-0.01-0.00 0.01 0.03 0.04 0.05 0.07 mJy/beam mJy/beam mJy/beam mJy/beam -0.06-0.03 0.00 0.03 0.06 0.09 0.12 0.15 0.18 0.22 mJy/beam -0.00 0.15 0.29 0.44 0.58 0.73 0.87 1.02 1.16 1.31 mJy/beam mJy/beam mJy/beam Figure 1.
Continued. chromospheric emission from the young stellar object, or a combi-nation of both (Dullemond et al. 2007; Millan-Gabet et al. 2007).For thermal free-free emission, the total integrated flux can vary bya factor of 20 −
40% over a long period, while for non-thermal emis-sion the total integrated flux can vary by up to a factor (cid:38)
MNRAS , 1–14 (2015)
Norfolk et al.
Deprojected baseline (k ) R e a l n o r m CSCha
Radius ["] B r i g h t n e ss n o r m Deprojected baseline (k ) R e a l n o r m HPCha
Radius ["] B r i g h t n e ss n o r m Deprojected baseline (k ) R e a l n o r m HD100453
Radius ["] B r i g h t n e ss n o r m Deprojected baseline (k ) R e a l n o r m HD135344B
Radius ["] B r i g h t n e ss n o r m Deprojected baseline (k ) R e a l n o r m RYLup
Radius ["] B r i g h t n e ss n o r m Deprojected baseline (k ) R e a l n o r m J1608
Radius ["] B r i g h t n e ss n o r m Deprojected baseline (k ) R e a l n o r m Sz111
Radius ["] B r i g h t n e ss n o r m Deprojected baseline (k ) R e a l n o r m SR24S
Radius ["] B r i g h t n e ss n o r m Deprojected baseline (k ) R e a l n o r m DoAr44
Radius ["] B r i g h t n e ss n o r m Deprojected baseline (k ) R e a l n o r m HD169142
Radius ["] B r i g h t n e ss n o r m ATCA observationsALMA observations Frankenstein ATCA fitFrankenstein ALMA fit 1 bootstrap ATCA1 bootstrap ALMA
Figure 2.
Normalised real component of the ATCA (blue) and ALMA (red) visibilities as a function of the deprojected baseline, and normalised Frankensteinmodel radial brightness profiles. The visibilities are overlaid by Frankenstein models (curves), and the model radial brightness profiles are shown withrespective 1 𝜎 bootstrap error bar (see Sect. 4.3.2 for further details on the error calculation). The 1D radial brightness profile models derived by Frankensteinhighlight that the majority of our sample (HD100453, HD135344B,RY Lup, J1608, Sz111, SR24S, DoAr44, and HD169142) share asimilar R peak in both the ATCA and ALMA data. The models ex-hibit a Gaussian ring fit indicative of a resolved inner cavity inFourier space that is likely due to the concentration of grains indust traps and hence, we take 𝑅 peak as representative cavity size.Neither CS Cha or HP Cha share similar R peak values betweenthe ATCA and ALMA data. For these sources, the ATCA 1D ra-dial brightness profiles suggest a disc without an inner deficit ofemission, and the ALMA fits suggest a ring-like morphology. Asa result, if we plot R peak (ALMA) versus R peak (ATCA) (see Figure4), our sample of transition discs can be classified into two groups:sources with similar R peak values and those which only show a cav- ity in the ALMA data. Sources in the first group from our sampleinclude HD100453, HD135344B, RY Lup, J1608, Sz111, SR24S,DoAr44, and HD169142. Other transitions discs in the literaturewhich exhibit an approximately one-to-one correlation between the8.8 mm and sub-mm R peak values include HD100546 (Wright et al.2015; Pinilla et al. 2018), GM Aur (Macías et al. 2018), LkCa15(Isella et al. 2014; Pinilla et al. 2018), and HD142527 (Casassuset al. 2015). The second group includes CS Cha and HP Cha fromour sample, and HD97048 from van der Plas et al. (2017a). For thepurposes of this paper, group 1 sources will be labelled resolved cav-ity transition discs (RC-TDs) and group 2 sources will be labellednon-detected cavity transition discs (NC-TDs).The non-detection of cavities at 8.8 mm for the NC-TDs canreflect the lack of a cavity in the large grain population, but couldalso be due to large asymmetric structure resulting in either a poormodel fit or inaccurate phase centering, unresolved features due MNRAS000
Normalised real component of the ATCA (blue) and ALMA (red) visibilities as a function of the deprojected baseline, and normalised Frankensteinmodel radial brightness profiles. The visibilities are overlaid by Frankenstein models (curves), and the model radial brightness profiles are shown withrespective 1 𝜎 bootstrap error bar (see Sect. 4.3.2 for further details on the error calculation). The 1D radial brightness profile models derived by Frankensteinhighlight that the majority of our sample (HD100453, HD135344B,RY Lup, J1608, Sz111, SR24S, DoAr44, and HD169142) share asimilar R peak in both the ATCA and ALMA data. The models ex-hibit a Gaussian ring fit indicative of a resolved inner cavity inFourier space that is likely due to the concentration of grains indust traps and hence, we take 𝑅 peak as representative cavity size.Neither CS Cha or HP Cha share similar R peak values betweenthe ATCA and ALMA data. For these sources, the ATCA 1D ra-dial brightness profiles suggest a disc without an inner deficit ofemission, and the ALMA fits suggest a ring-like morphology. Asa result, if we plot R peak (ALMA) versus R peak (ATCA) (see Figure4), our sample of transition discs can be classified into two groups:sources with similar R peak values and those which only show a cav- ity in the ALMA data. Sources in the first group from our sampleinclude HD100453, HD135344B, RY Lup, J1608, Sz111, SR24S,DoAr44, and HD169142. Other transitions discs in the literaturewhich exhibit an approximately one-to-one correlation between the8.8 mm and sub-mm R peak values include HD100546 (Wright et al.2015; Pinilla et al. 2018), GM Aur (Macías et al. 2018), LkCa15(Isella et al. 2014; Pinilla et al. 2018), and HD142527 (Casassuset al. 2015). The second group includes CS Cha and HP Cha fromour sample, and HD97048 from van der Plas et al. (2017a). For thepurposes of this paper, group 1 sources will be labelled resolved cav-ity transition discs (RC-TDs) and group 2 sources will be labellednon-detected cavity transition discs (NC-TDs).The non-detection of cavities at 8.8 mm for the NC-TDs canreflect the lack of a cavity in the large grain population, but couldalso be due to large asymmetric structure resulting in either a poormodel fit or inaccurate phase centering, unresolved features due MNRAS000 , 1–14 (2015) ust Traps and the Formation of Cavities in TDs Table 4.
Radial brightness profile properties of the ATCA and ALMA images for sources modelled with Frankenstein. This includes the radial offset ofthe peak brightness R peak (col. 2 and 3). We compare our Frankenstein fits with previous visibility models in col.4. Also included are the 1D profile peakpositions from other models in the literature from infrared (IR) and 13CO observations, and the observed inclination and position angle of the target disc.Superscripts in the column headers refer to the position in the reference list of the last column.
Source 𝑅 peak ATCA obs(") 𝑅 peak ALMA obs(") 𝑅 peak ALMA a (") 𝑅 peak IR b (") 𝑅 peak c (") Incl . d ( ° ) PA d ( ° ) Ref.(a,b,c,d)CS Cha <0.390 𝑢 . + . − . - 0.102 - 8 161 -, 8, -, 5HP Cha <0.340 𝑢 . + . − . - - - 37 164 -, -, -, 6HD100453 0 . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . - 0.286 68 109 1, -, 14, 7J1608 0 . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . - 0.225 -53 40 1, -, 14, 7SR24S 0 . + . − . . + . − . . + . − . - - 46 25 4, -, -, 1DoAr44 0 . + . − . . + . − . . + . − . . + . − . , 0 . + . − . , 0 . + . − . , 0.118 0.121 13 5 2, 13, 15, 20 . + . − . * 0 . + . − . * 0 . + . − . * - - Notes: 𝑢 Upper limits are taken as the maximum cavity size (if a cavity exists) from the error for each source in Figure 4.*Second peak.
References (1) Pinilla et al. (2018), (2) Fedele et al. (2017), (3) van der Plas et al. (2019), (4) Pinilla et al. (2017), (5) Francis & van der Marel (2020) (6)Cazzoletti et al. (2018), (7) Ansdell et al. (2018), (8) Ginski et al. (2018), (9) Benisty et al. (2017), (10) Stolker et al. (2017), (11) Villenave et al. (2019), (12)Casassus et al. (2018), (13) Birchall et al. (2019), (14) van der Marel et al. (2018), (15) Carney et al. (2018), (16) van der Marel et al. (2016a). to spatial resolution limitations, or the presence of other sourcesof emission obscuring disc features as suggested by van der Plaset al. (2017a) for HD97048. The emission for CS Cha and HP Chaappears to be dominated by thermal dust (see Figure 3). Due to thelack of proper motions and the astrometric uncertainty of ATCA,it is possible that the emission of HP Cha is asymmetric and thechosen phase centre is incorrect (e.g. see Figure C1), as asymmetriesare expected to become more pronounced or even only revealed atlonger wavelengths (Birnstiel et al. 2013; van der Marel et al. 2020).However, it is equally likely, given the low spatial resolution of theATCA observations, that a cavity exists for CS Cha and HP Cha butremains unresolved. Using Frankenstein to fit low SNR visibilitieswill result in poor fits not representative of the true disc features.As a result, the presence of a cavity at 8.8 mm cannot be confirmedfor CS Cha and HP Cha, and more observations are required.The exact mechanism driving large cavity formation in tran-sition discs remains unknown (Williams & Cieza 2011). Howeverhistorically the observed deficit of inner disc emission has beenattributed to either grain growth, photoevaporation, or the trappingof dust due to either dead zones or the dynamical interaction with acompanion. Grain growth models (Laibe et al. 2008; Birnstiel et al.2012b; Gonzalez et al. 2015) predict, due to the settling, growth, andinward drift of grains, that the disc will contain a size-sorted radialdistribution of dust, with larger grains preferentially concentratedtowards the mid-plane and host star. Previous reviews of photoe- vaporative models (Ercolano & Pascucci 2017) conclude that thedeficit of inner emission seen in transition discs with large cavi-ties ( ≥ ≥ − M (cid:12) yr − cannot beattributed entirely to photoevaporation. Therefore, we suggest thatneither grain growth nor photoevaporation are likely to be domi-nating the grain evolution in our RC-TD sources. Our sample doesnot show evidence of radial size sorting (Figure 2), exhibit largecavities (Figure 4), and the majority have modest levels of accretion(Table 1).Magnetohydrodynamical (MHD) models of transition discsfrom Pinilla et al. (2016) predict a factor of 2 difference in the radialpeak positions for sub-mm to mm grains at dust traps located at theedge of dead zones (see Figure 4 at 1Myr in Pinilla et al. 2016)which is excluded by our one-to-one correlation seen in Figure4. Additionally, MHD models require a high turbulence viscositycoefficient ( 𝛼 ≈ − ) in active regions of the disc (Pinilla et al.2016; Ueda et al. 2019), which is inconsistent with observationsof classical protoplanetary discs (e.g. 𝛼 ≈ − − − from Pinteet al. 2016; Flaherty et al. 2020) and previous studies of transitiondiscs that model how inner companions influence cavity formation(Dong et al. 2017; Toci et al. 2020). Dong et al. (2017) showed withhydrodynamical simulations and radiative transfer models that theturbulence viscosity coefficient for HD169142 had to be lower than10 − for a single planet to form two distinct ring that agreed with theobservations. However Toci et al. (2020), using a similar procedure MNRAS , 1–14 (2015) Norfolk et al. =2.74 HD34282 =3.64 SZCha =2.29 CSCha =3.08 HPCha =2.65 HD100453 =3.06 HD135344B =2.78 HD143006 =3.59 RYLup =3.04 J1604.3-2130 =2.79 J16083070 Wavelength [mm] F l u x [ m J y ] =2.52 Sz111 =2.6 SR24S =3.42 SR21A =2.1 DoAr44 =2.86 HD169142
Figure 3.
Spectral slopes for 14 sources in the Discs Down Under survey with observed emission between 0.8 and 3.3 mm. The error bars shown in both panelsincludes a calibration uncertainty of 10% (ALMA) and 20% (ATCA). Over-plotted with grey shading are linear least-squares fits error limits to parts of thespectral slope between 0.8 and 3.3 mm. but with two planets, were able to reproduce the two observed ringswith 𝛼 = × − . Turbulence in the active regions of transitiondiscs may not be well enough constrained to reach a firm conclusionregarding its role in cavity formation.The dynamical interaction with a companion is proposed to bethe driving force behind most features seen in transition discs in-cluding the formations of dust traps. For our RC-TDs, van der Marelet al. (2015a, 2016a, 2018) showed gas cavities inside dust cavitiesfor SR21, HD135344B, DoAr44, J1604, Sz111, J16083070. vander Plas et al. (2019) utilised SPH and radiative transfer codes toshow some disc features seen in HD100453 can be attributed to anundetected, low mass close companion within the disc’s cavity. Thestudy of HD100453 is further refined by Gonzalez et al. (2020) andNealon et al. (2020), who show that while the outer disc morphologycan be caused by a companion star on an inclined orbit exterior tothe disc, the inner cavity can be explained by a (cid:46) J planet, lessmassive than previously suggested (van der Plas et al. 2019), at 15-20 au. Using similar 1D radial brightness profiles for the majorityof our sources (HD135344B, RY Lup, J1608, Sz111, SR24S, andDoAr44), Pinilla et al. (2017) and Pinilla et al. (2018) suggest thatthe clearing of the inner cavity could be attributed to the dynamicalinteraction with an embedded planet. For HD169142, Fedele et al.(2017) uses a thermo-chemical code to model the gas and dust dis-tribution. Their results suggest that dynamical interaction betweenthe disc and two giant embedded planets results in the depletionof the inner disc, and creates two pressure bumps that aid in thetrapping of dust at each observable ring position. It still remains unclear for the majority of transition discs if a companion is indeedresponsible for the formation of the cavity, whether this companionis an embedded planet or a binary star. What is notable about oursample is that for the RC-TDs the 8.8 mm and sub-mm grains sharea similar cavity size (R peak value). This is also seen in SR24S andHD142527 (included in Figure 4). Pinilla et al. (2019) presents aplanet-disc model for SR24S that predicts both mm and sub-mmgrains will share a similar radial peak in the dust density. Price et al.(2018) extensively modelled HD142527 with 3D hydrodynamicalsimulations considering several possible orbits for the M-dwarf bi-nary companion presented in Lacour et al. (2016). They concludethat all the observable disc features can be attributed to the tidaltruncation of the disc from the binary companion. Given that ourRC-TDs and HD142527 follow the one-to-one correlation shownin Figure 4, we suggest similar physical mechanisms affecting thegrain evolution for these discs may have occurred.We include in Table 4 the IR and CO R peak values wherepossible for our sample. The IR and CO R peak values are consis-tently smaller than the corresponding ATCA/ALMA R peak values,and for sources with both IR and CO observations, the respectiveR peak values are similar. The discrepancy between the ATCA andALMA versus IR and CO R peak values was originally classified asthe "missing cavities" problem by Dong et al. (2012). With ATCAand ALMA observations tracing large dust grains (sub-mm to mmsizes) and IR observations tracing small dust grains ( ∼ 𝜇 m sizes)our comparison is in agreement with dust evolution models whichshow large dust grains accumulate at pressure maxima (dust traps), MNRAS000
Spectral slopes for 14 sources in the Discs Down Under survey with observed emission between 0.8 and 3.3 mm. The error bars shown in both panelsincludes a calibration uncertainty of 10% (ALMA) and 20% (ATCA). Over-plotted with grey shading are linear least-squares fits error limits to parts of thespectral slope between 0.8 and 3.3 mm. but with two planets, were able to reproduce the two observed ringswith 𝛼 = × − . Turbulence in the active regions of transitiondiscs may not be well enough constrained to reach a firm conclusionregarding its role in cavity formation.The dynamical interaction with a companion is proposed to bethe driving force behind most features seen in transition discs in-cluding the formations of dust traps. For our RC-TDs, van der Marelet al. (2015a, 2016a, 2018) showed gas cavities inside dust cavitiesfor SR21, HD135344B, DoAr44, J1604, Sz111, J16083070. vander Plas et al. (2019) utilised SPH and radiative transfer codes toshow some disc features seen in HD100453 can be attributed to anundetected, low mass close companion within the disc’s cavity. Thestudy of HD100453 is further refined by Gonzalez et al. (2020) andNealon et al. (2020), who show that while the outer disc morphologycan be caused by a companion star on an inclined orbit exterior tothe disc, the inner cavity can be explained by a (cid:46) J planet, lessmassive than previously suggested (van der Plas et al. 2019), at 15-20 au. Using similar 1D radial brightness profiles for the majorityof our sources (HD135344B, RY Lup, J1608, Sz111, SR24S, andDoAr44), Pinilla et al. (2017) and Pinilla et al. (2018) suggest thatthe clearing of the inner cavity could be attributed to the dynamicalinteraction with an embedded planet. For HD169142, Fedele et al.(2017) uses a thermo-chemical code to model the gas and dust dis-tribution. Their results suggest that dynamical interaction betweenthe disc and two giant embedded planets results in the depletionof the inner disc, and creates two pressure bumps that aid in thetrapping of dust at each observable ring position. It still remains unclear for the majority of transition discs if a companion is indeedresponsible for the formation of the cavity, whether this companionis an embedded planet or a binary star. What is notable about oursample is that for the RC-TDs the 8.8 mm and sub-mm grains sharea similar cavity size (R peak value). This is also seen in SR24S andHD142527 (included in Figure 4). Pinilla et al. (2019) presents aplanet-disc model for SR24S that predicts both mm and sub-mmgrains will share a similar radial peak in the dust density. Price et al.(2018) extensively modelled HD142527 with 3D hydrodynamicalsimulations considering several possible orbits for the M-dwarf bi-nary companion presented in Lacour et al. (2016). They concludethat all the observable disc features can be attributed to the tidaltruncation of the disc from the binary companion. Given that ourRC-TDs and HD142527 follow the one-to-one correlation shownin Figure 4, we suggest similar physical mechanisms affecting thegrain evolution for these discs may have occurred.We include in Table 4 the IR and CO R peak values wherepossible for our sample. The IR and CO R peak values are consis-tently smaller than the corresponding ATCA/ALMA R peak values,and for sources with both IR and CO observations, the respectiveR peak values are similar. The discrepancy between the ATCA andALMA versus IR and CO R peak values was originally classified asthe "missing cavities" problem by Dong et al. (2012). With ATCAand ALMA observations tracing large dust grains (sub-mm to mmsizes) and IR observations tracing small dust grains ( ∼ 𝜇 m sizes)our comparison is in agreement with dust evolution models whichshow large dust grains accumulate at pressure maxima (dust traps), MNRAS000 , 1–14 (2015) ust Traps and the Formation of Cavities in TDs R peak mm (au) R p e a k s u b - mm ( a u ) CSChaHPCha HD100453HD135344B RYLup J1608Sz111SR24S DoAr44HD169142 GM AurHD 97048 HD 142527LkCa 15HD100546
Detected excess emissionNo excess emissionOur sampleOther TDs
Figure 4.
The 0 . − . peak radial offsets as a function of the 7 − peak radial offsets. The dashed line represents an equal radial offset at bothwavelength ranges. The pentagon and triangle markers indicate whetheremission other than thermal dust emission has been detected in the source asno excess or detected excess respectively. The pink markers represent R peak values from our Frankenstein modelling, and the blue markers representR peak values from the literature including: HD100546 0 .
86 mm data fromPinilla et al. (2018) and 7 mm data from Wright et al. (2015); GM Aur0 . .
43 mm datafrom Pinilla et al. (2018) and 7 mm data from Isella et al. (2014); HD1425271 . .
85 mm and 8 . while smaller dust grains are coupled to the gas and follow the ac-cretion flow into the inner disc (Birnstiel et al. 2012a; Pinilla et al.2012; Zhu et al. 2012). In this work we present the largest 8.8 mm continuum survey to dateof transition discs with large cavities, and compare the resulting dustemission to ALMA observations.(i) For the 15 discs observed with ATCA, 14 were detected, 13were spatially resolved, and 8 exhibited morphology indicative ofring-like structure.(ii) The spectral slopes indicate that the emission is dominatedby thermal dust for most sources, suggesting the ATCA emission istracing the large grain population. Only SZ Cha and RY Lup showedexcess emission above that from thermal dust emission.(iii) We use the Frankenstein code to model 1D radial brightnessprofiles, which reveal that both the 8.8 mm ATCA and sub-mmALMA emission in most of the discs share a similar radial positionof the peak brightness, R peak , suggesting a similar cavity size.(iv) The R peak values for the millimetre and sub-mm data is con-sistently larger than the corresponding IR and CO R peak values.This is in agreement with models indicating large grains prefer- entially concentrating at maxima in the gas pressure, whilst small ∼ 𝜇 m-sized grains are coupled to the gas.(v) We suggest that the large cavities in these transition discsresult from a dust trap, likely induced by a companion. ACKNOWLEDGEMENTS
The authors thank the referee for their constructive commentsand suggestions. We thank Mathew Agnew and Elodie Thilliezfor useful discussions with our ATCA observing proposal. B.J.Nis supported by an Australian Government Research TrainingProgram (RTP) Scholarship. CP acknowledges funding fromthe Australian Research Council via grants FT170100040 andDP180104235. N.M. acknowledges support from the Banting Post-doctoral Fellowships program, administered by the Government ofCanada. RAB acknowledges support from the STFC consolidatedgrant ST/S000623/1. This work has also been supported bythe European Union’s Horizon 2020 research and innovationprogramme under the Marie Sklodowska-Curie grant agreementNo 823823 (DUSTBUSTERS). J.-F.G. acknowledges funding fromANR (Agence Nationale de la Recherche) of France under contractnumber ANR-16-CE31-0013 (Planet-Forming-) and thank theLABEX Lyon Institute of Origins (ANR-10-LABX-0066) of theUniversité de Lyon for its financial support within the programme‘Investissements d’Avenir’ (ANR-11-IDEX-0007) of the Frenchgovernment operated by the ANR. C. M. Wright acknowledgesfinancial support from the Australian Research Council via FutureFellowship FT100100495. The Australia Telescope Compact Arrayis part of the Australia Telescope which is funded by the Common-wealth of Australia for operation as a National Facility managed byCSIRO. This research has made use of NASA’s Astrophysics DataSystem. The National Radio Astronomy Observatory is a facilityof the National Science Foundation operated under agreementby the Associated Universities, Inc. ALMA is a partnership ofESO (representing its member states), NSF (USA) and NINS(Japan), together with NRC (Canada) and NSC and ASIAA(Taiwan) and KASI (Republic of Korea), in cooperation with theRepublic of Chile. The Joint ALMA Observatory is operatedby ESO, AUI/ NRAO and NAOJ. This paper makes use of thefollowing ALMA data: ADS/JAO.ALMA
MNRAS , 1–14 (2015) Norfolk et al.
DATA AVAILABILITY
The ATCA observational data used in this paper is available fromthe Australian National Telescope Facility Archive at https://atoa.atnf.csiro.au/ under project code C3119. TheALMA observational data is available from the ALMA sciencearchive at https://almascience.nrao.edu/aq/ underthe project IDs listed in Table 1. 1D radial brightness mod-els used the Frankenstein code which is available from https://github.com/discsim/frank . Reduced observation dataand models will be shared on reasonable request to the correspond-ing author.
REFERENCES
Alcalá J. M., et al., 2017, A&A, 600, A20Alexander R. D., Clarke C. J., Pringle J. E., 2006, MNRAS, 369, 229Andrews S. M., Wilner D. J., Espaillat C., Hughes A. M., Dullemond C. P.,McClure M. K., Qi C., Brown J. M., 2011, ApJ, 732, 42Andrews S. M., et al., 2018, ApJ, 869, L41Ansdell M., et al., 2016, ApJ, 828, 46Ansdell M., et al., 2018, ApJ, 859, 21Astropy Collaboration et al., 2013, A&A, 558, A33Baddour N., Chouinard U., 2015, Journal of the Optical Society of AmericaA, 32, 611Benisty M., et al., 2017, A&A, 597, A42Birchall E. K., et al., 2019, MNRAS, 486, 3721Birnstiel T., Klahr H., Ercolano B., 2012a, A&A, 539, A148Birnstiel T., Andrews S. M., Ercolano B., 2012b, A&A, 544, A79Birnstiel T., Dullemond C. P., Pinilla P., 2013, A&A, 550, L8Brauer F., Dullemond C. P., Johansen A., Henning T., Klahr H., Natta A.,2007, A&A, 469, 1169Brown J. M., Blake G. A., Qi C., Dullemond C. P., Wilner D. J., WilliamsJ. P., 2009, ApJ, 704, 496Carney M. T., et al., 2018, A&A, 614, A106Casassus S., et al., 2015, ApJ, 812, 126Casassus S., et al., 2018, MNRAS, 477, 5104Cazzoletti P., et al., 2018, A&A, 619, A161Clarke C. J., Gendrin A., Sotomayor M., 2001, MNRAS, 328, 485Dent W. R. F., Torrelles J. M., Osorio M., Calvet N., Anglada G., 2006,Monthly Notices of the Royal Astronomical Society, 365, 1283Dong R., et al., 2012, ApJ, 750, 161Dong R., Li S., Chiang E., Li H., 2017, ApJ, 843, 127Draine B. T., 2006, ApJ, 636, 1114Dullemond C. P., Hollenbach D., Kamp I., D’Alessio P., 2007, inReipurth B., Jewitt D., Keil K., eds, Protostars and Planets V. p. 555( arXiv:astro-ph/0602619 )Ercolano B., Pascucci I., 2017, Royal Society Open Science, 4, 170114Ercolano B., Weber M. L., Owen J. E., 2018, MNRAS, 473, L64Espaillat C., et al., 2014, in Beuther H., Klessen R. S., Dullemond C. P., Hen-ning T., eds, Protostars and Planets VI. p. 497 ( arXiv:1402.7103 ),doi:10.2458/azu_uapress_9780816531240-ch022Fedele D., et al., 2017, A&A, 600, A72Flaherty K., et al., 2020, ApJ, 895, 109Flock M., Ruge J. P., Dzyurkevich N., Henning T., Klahr H., Wolf S., 2015,A&A, 574, A68Fouchet L., Maddison S. T., Gonzalez J. F., Murray J. R., 2007, A&A, 474,1037Fouchet L., Gonzalez J. F., Maddison S. T., 2010, A&A, 518, A16Francis L., van der Marel N., 2020, ApJ, 892, 111Gaia Collaboration 2018, VizieR Online Data Catalog, p. I/345Garg H., et al., 2020, arXiv e-prints, p. arXiv:2010.15310Ginski C., et al., 2018, A&A, 616, A79Gonzalez J. F., Laibe G., Maddison S. T., Pinte C., Ménard F., 2015, Planet.Space Sci., 116, 48Gonzalez J.-F., et al., 2020, MNRAS, submitted Gorti U., Hollenbach D., Dullemond C. P., 2015, ApJ, 804, 29Hollenbach D., Johnstone D., Lizano S., Shu F., 1994, ApJ, 428, 654Hunter J. D., 2007, Computing in Science & Engineering, 9, 90Isella A., Chandler C. J., Carpenter J. M., Pérez L. M., Ricci L., 2014, ApJ,788, 129Isella A., et al., 2018, The Astrophysical Journal, 869, L49Jennings J., Booth R. A., Tazzari M., Rosotti G. P., Clarke C. J., 2020, arXive-prints, p. arXiv:2005.07709Lacour S., et al., 2016, A&A, 590, A90Laibe G., Gonzalez J. F., Fouchet L., Maddison S. T., 2008, A&A, 487, 265Langlois M., et al., 2018, A&A, 614, A88Lin D. N. C., Papaloizou J., 1979, MNRAS, 188, 191Lommen D., Maddison S. T., Wright C. M., van Dishoeck E. F., WilnerD. J., Bourke T. L., 2009, Astronomy and Astrophysics, 495, 869Lommen D. J. P., et al., 2010, Astronomy and Astrophysics, 515, A77Macías E., et al., 2018, ApJ, 865, 37Manara C. F., Testi L., Natta A., Rosotti G., Benisty M., Ercolano B., RicciL., 2014, A&A, 568, A18Manara C. F., Mordasini C., Testi L., Williams J. P., Miotello A., Lodato G.,Emsenhuber A., 2019, A&A, 631, L2Manara C. F., et al., 2020, A&A, 639, A58McMullin J. P., Waters B., Schiebel D., Young W., Golap K., 2007, CASAArchitecture and Applications. p. 127Millan-Gabet R., Malbet F., Akeson R., Leinert C., Monnier J., Waters R.,2007, in Reipurth B., Jewitt D., Keil K., eds, Protostars and Planets V.p. 539 ( arXiv:astro-ph/0603554 )Natta A., Testi L., 2004, in Johnstone D., Adams F. C., Lin D. N. C.,Neufeeld D. A., Ostriker E. C., eds, Astronomical Society of the PacificConference Series Vol. 323, Star Formation in the Interstellar Medium:In Honor of David Hollenbach. p. 279Nealon R., Cuello N., Gonzalez J.-F., van der Plas G., Pinte C., AlexanderR., Ménard F., Price D. J., 2020, MNRAS, acceptedOwen J. E., Clarke C. J., 2012, MNRAS, 426, L96Owen J. E., Ercolano B., Clarke C. J., 2017, MNRAS, 472, 2955Paardekooper S. J., Mellema G., 2004, A&A, 425, L9Pinilla P., Benisty M., Birnstiel T., 2012, A&A, 545, A81Pinilla P., Flock M., Ovelar M. d. J., Birnstiel T., 2016, A&A, 596, A81Pinilla P., et al., 2017, ApJ, 839, 99Pinilla P., et al., 2018, ApJ, 859, 32Pinilla P., Benisty M., Cazzoletti P., Harsono D., Pérez L. M., Tazzari M.,2019, ApJ, 878, 16Pinte C., Dent W. R. F., Ménard F., Hales A., Hill T., Cortes P., de Gregorio-Monsalvo I., 2016, ApJ, 816, 25Price D. J., et al., 2018, MNRAS, 477, 1270Rice W. K. M., Armitage P. J., Wood K., Lodato G., 2006, MNRAS, 373,1619Rodmann J., Henning T., Chandler C. J., Mundy L. G., Wilner D. J., 2006,A&A, 446, 211Ruge J. P., Flock M., Wolf S., Dzyurkevich N., Fromang S., Henning T.,Klahr H., Meheut H., 2016, A&A, 590, A17Sault R. J., Teuben P. J., Wright M. C. H., 1995, in Shaw R. A., Payne H. E.,Hayes J. J. E., eds, Astronomical Society of the Pacific ConferenceSeries Vol. 77, Astronomical Data Analysis Software and Systems IV.p. 433 ( arXiv:astro-ph/0612759 )Stolker T., et al., 2017, ApJ, 849, 143Strom K. M., Strom S. E., Edwards S., Cabrit S., Skrutskie M. F., 1989, AJ,97, 1451Tanaka H., Himeno Y., Ida S., 2005, ApJ, 625, 414Toci C., Lodato G., Fedele D., Testi L., Pinte C., 2020, ApJ, 888, L4Ubach C., Maddison S. T., Wright C. M., Wilner D. J., Lommen D. J. P.,Koribalski B., 2012, MNRAS, 425, 3137Ubach C., Maddison S. T., Wright C. M., Wilner D. J., Lommen D. J. P.,Koribalski B., 2017, MNRAS, 466, 4083Ueda T., Flock M., Okuzumi S., 2019, ApJ, 871, 10Villenave M., et al., 2019, A&A, 624, A7Weidenschilling S. J., 1977, MNRAS, 180, 57Whipple F. L., 1972, in Elvius A., ed., From Plasma to Planet. p. 211MNRAS000
Alcalá J. M., et al., 2017, A&A, 600, A20Alexander R. D., Clarke C. J., Pringle J. E., 2006, MNRAS, 369, 229Andrews S. M., Wilner D. J., Espaillat C., Hughes A. M., Dullemond C. P.,McClure M. K., Qi C., Brown J. M., 2011, ApJ, 732, 42Andrews S. M., et al., 2018, ApJ, 869, L41Ansdell M., et al., 2016, ApJ, 828, 46Ansdell M., et al., 2018, ApJ, 859, 21Astropy Collaboration et al., 2013, A&A, 558, A33Baddour N., Chouinard U., 2015, Journal of the Optical Society of AmericaA, 32, 611Benisty M., et al., 2017, A&A, 597, A42Birchall E. K., et al., 2019, MNRAS, 486, 3721Birnstiel T., Klahr H., Ercolano B., 2012a, A&A, 539, A148Birnstiel T., Andrews S. M., Ercolano B., 2012b, A&A, 544, A79Birnstiel T., Dullemond C. P., Pinilla P., 2013, A&A, 550, L8Brauer F., Dullemond C. P., Johansen A., Henning T., Klahr H., Natta A.,2007, A&A, 469, 1169Brown J. M., Blake G. A., Qi C., Dullemond C. P., Wilner D. J., WilliamsJ. P., 2009, ApJ, 704, 496Carney M. T., et al., 2018, A&A, 614, A106Casassus S., et al., 2015, ApJ, 812, 126Casassus S., et al., 2018, MNRAS, 477, 5104Cazzoletti P., et al., 2018, A&A, 619, A161Clarke C. J., Gendrin A., Sotomayor M., 2001, MNRAS, 328, 485Dent W. R. F., Torrelles J. M., Osorio M., Calvet N., Anglada G., 2006,Monthly Notices of the Royal Astronomical Society, 365, 1283Dong R., et al., 2012, ApJ, 750, 161Dong R., Li S., Chiang E., Li H., 2017, ApJ, 843, 127Draine B. T., 2006, ApJ, 636, 1114Dullemond C. P., Hollenbach D., Kamp I., D’Alessio P., 2007, inReipurth B., Jewitt D., Keil K., eds, Protostars and Planets V. p. 555( arXiv:astro-ph/0602619 )Ercolano B., Pascucci I., 2017, Royal Society Open Science, 4, 170114Ercolano B., Weber M. L., Owen J. E., 2018, MNRAS, 473, L64Espaillat C., et al., 2014, in Beuther H., Klessen R. S., Dullemond C. P., Hen-ning T., eds, Protostars and Planets VI. p. 497 ( arXiv:1402.7103 ),doi:10.2458/azu_uapress_9780816531240-ch022Fedele D., et al., 2017, A&A, 600, A72Flaherty K., et al., 2020, ApJ, 895, 109Flock M., Ruge J. P., Dzyurkevich N., Henning T., Klahr H., Wolf S., 2015,A&A, 574, A68Fouchet L., Maddison S. T., Gonzalez J. F., Murray J. R., 2007, A&A, 474,1037Fouchet L., Gonzalez J. F., Maddison S. T., 2010, A&A, 518, A16Francis L., van der Marel N., 2020, ApJ, 892, 111Gaia Collaboration 2018, VizieR Online Data Catalog, p. I/345Garg H., et al., 2020, arXiv e-prints, p. arXiv:2010.15310Ginski C., et al., 2018, A&A, 616, A79Gonzalez J. F., Laibe G., Maddison S. T., Pinte C., Ménard F., 2015, Planet.Space Sci., 116, 48Gonzalez J.-F., et al., 2020, MNRAS, submitted Gorti U., Hollenbach D., Dullemond C. P., 2015, ApJ, 804, 29Hollenbach D., Johnstone D., Lizano S., Shu F., 1994, ApJ, 428, 654Hunter J. D., 2007, Computing in Science & Engineering, 9, 90Isella A., Chandler C. J., Carpenter J. M., Pérez L. M., Ricci L., 2014, ApJ,788, 129Isella A., et al., 2018, The Astrophysical Journal, 869, L49Jennings J., Booth R. A., Tazzari M., Rosotti G. P., Clarke C. J., 2020, arXive-prints, p. arXiv:2005.07709Lacour S., et al., 2016, A&A, 590, A90Laibe G., Gonzalez J. F., Fouchet L., Maddison S. T., 2008, A&A, 487, 265Langlois M., et al., 2018, A&A, 614, A88Lin D. N. C., Papaloizou J., 1979, MNRAS, 188, 191Lommen D., Maddison S. T., Wright C. M., van Dishoeck E. F., WilnerD. J., Bourke T. L., 2009, Astronomy and Astrophysics, 495, 869Lommen D. J. P., et al., 2010, Astronomy and Astrophysics, 515, A77Macías E., et al., 2018, ApJ, 865, 37Manara C. F., Testi L., Natta A., Rosotti G., Benisty M., Ercolano B., RicciL., 2014, A&A, 568, A18Manara C. F., Mordasini C., Testi L., Williams J. P., Miotello A., Lodato G.,Emsenhuber A., 2019, A&A, 631, L2Manara C. F., et al., 2020, A&A, 639, A58McMullin J. P., Waters B., Schiebel D., Young W., Golap K., 2007, CASAArchitecture and Applications. p. 127Millan-Gabet R., Malbet F., Akeson R., Leinert C., Monnier J., Waters R.,2007, in Reipurth B., Jewitt D., Keil K., eds, Protostars and Planets V.p. 539 ( arXiv:astro-ph/0603554 )Natta A., Testi L., 2004, in Johnstone D., Adams F. C., Lin D. N. C.,Neufeeld D. A., Ostriker E. C., eds, Astronomical Society of the PacificConference Series Vol. 323, Star Formation in the Interstellar Medium:In Honor of David Hollenbach. p. 279Nealon R., Cuello N., Gonzalez J.-F., van der Plas G., Pinte C., AlexanderR., Ménard F., Price D. J., 2020, MNRAS, acceptedOwen J. E., Clarke C. J., 2012, MNRAS, 426, L96Owen J. E., Ercolano B., Clarke C. J., 2017, MNRAS, 472, 2955Paardekooper S. J., Mellema G., 2004, A&A, 425, L9Pinilla P., Benisty M., Birnstiel T., 2012, A&A, 545, A81Pinilla P., Flock M., Ovelar M. d. J., Birnstiel T., 2016, A&A, 596, A81Pinilla P., et al., 2017, ApJ, 839, 99Pinilla P., et al., 2018, ApJ, 859, 32Pinilla P., Benisty M., Cazzoletti P., Harsono D., Pérez L. M., Tazzari M.,2019, ApJ, 878, 16Pinte C., Dent W. R. F., Ménard F., Hales A., Hill T., Cortes P., de Gregorio-Monsalvo I., 2016, ApJ, 816, 25Price D. J., et al., 2018, MNRAS, 477, 1270Rice W. K. M., Armitage P. J., Wood K., Lodato G., 2006, MNRAS, 373,1619Rodmann J., Henning T., Chandler C. J., Mundy L. G., Wilner D. J., 2006,A&A, 446, 211Ruge J. P., Flock M., Wolf S., Dzyurkevich N., Fromang S., Henning T.,Klahr H., Meheut H., 2016, A&A, 590, A17Sault R. J., Teuben P. J., Wright M. C. H., 1995, in Shaw R. A., Payne H. E.,Hayes J. J. E., eds, Astronomical Society of the Pacific ConferenceSeries Vol. 77, Astronomical Data Analysis Software and Systems IV.p. 433 ( arXiv:astro-ph/0612759 )Stolker T., et al., 2017, ApJ, 849, 143Strom K. M., Strom S. E., Edwards S., Cabrit S., Skrutskie M. F., 1989, AJ,97, 1451Tanaka H., Himeno Y., Ida S., 2005, ApJ, 625, 414Toci C., Lodato G., Fedele D., Testi L., Pinte C., 2020, ApJ, 888, L4Ubach C., Maddison S. T., Wright C. M., Wilner D. J., Lommen D. J. P.,Koribalski B., 2012, MNRAS, 425, 3137Ubach C., Maddison S. T., Wright C. M., Wilner D. J., Lommen D. J. P.,Koribalski B., 2017, MNRAS, 466, 4083Ueda T., Flock M., Okuzumi S., 2019, ApJ, 871, 10Villenave M., et al., 2019, A&A, 624, A7Weidenschilling S. J., 1977, MNRAS, 180, 57Whipple F. L., 1972, in Elvius A., ed., From Plasma to Planet. p. 211MNRAS000 , 1–14 (2015) ust Traps and the Formation of Cavities in TDs Table A1.
Integrated 7 mm flux, residual rms, synthesised beam size 𝜃 𝑏 ,and position angle for 34 GHz observations of HD163296. Target ATCA34 GHz flux(mJy) ATCA1 𝜎 rms(mJy) ATCA 𝜃 beam (" x ") ATCAPA beam ( ° )HD 163296 2.45 0.0026 1.31 x 0.400 4.59 Whittet D. C. B., Prusti T., Franco G. A. P., Gerakines P. A., Kilkenny D.,Larson K. A., Wesselius P. R., 1997, A&A, 327, 1194Williams J. P., Cieza L. A., 2011, ARA&A, 49, 67Wilson W. E., et al., 2011, MNRAS, 416, 832Wolk S. J., Walter F. M., 1996, AJ, 111, 2066Wright C. M., et al., 2015, MNRAS, 453, 414Zhu Z., Nelson R. P., Dong R., Espaillat C., Hartmann L., 2012, ApJ, 755, 6van der Marel N., van Dishoeck E. F., Bruderer S., Pérez L., Isella A., 2015a,A&A, 579, A106van der Marel N., Pinilla P., Tobin J., van Kempen T., Andrews S., Ricci L.,Birnstiel T., 2015b, ApJ, 810, L7van der Marel N., van Dishoeck E. F., Bruderer S., Andrews S. M., Pontop-pidan K. M., Herczeg G. J., van Kempen T., Miotello A., 2016a, A&A,585, A58van der Marel N., Cazzoletti P., Pinilla P., Garufi A., 2016b, ApJ, 832, 178van der Marel N., et al., 2018, ApJ, 854, 177van der Marel N., et al., 2020, arXiv e-prints, p. arXiv:2010.10568van der Plas G., et al., 2017a, A&A, 597, A32van der Plas G., Ménard F., Canovas H., Avenhaus H., Casassus S., PinteC., Caceres C., Cieza L., 2017b, A&A, 607, A55van der Plas G., et al., 2019, A&A, 624, A33
APPENDIX A: HD163296
Here we present the 34 GHz observations of HD163296. This sourcewas part of the Discs Down Under survey but is not classified as atransition disc. We include the continuum 34 GHz results in TableA1 and show the continuum map in Figure A1. The continuum mapshows significant central emission likely from the combination ofboth the inner disc and ring structure seen in the recent DSHARPsurvey (Isella et al. 2018), our observations fail to resolve these fea-tures due to the lack of sufficiently long baselines and uv-coverage.No excess of emission is seen in Figure A2.
APPENDIX B: FRANKENSTEIN HYPER-PARAMETERSWEEP
Here we include the Frankenstein hyper-parameter sweep forHD100453. Figure B1 highlights the Frankenstein’s fit depen-dence on the selection of 𝛼 and w smooth . The fit is not overly sen-sitive to w smooth which determines the length over which adjacentpoints are coupled. For our sample, 𝛼 is varied between 1.01 and1.5, the choice of 𝛼 for each disc is determined by the character-istics of the visibility data. Values of 𝛼 close to 1.01 are typicallychosen for noisy data sets as Frankenstein attempts to fit longerbaseline visibilities with lower SNR. As expected, Figure B1 showsthat low choices of 𝛼 result in lower 𝜒 values and R peak values thatvary minimally. However, if 𝛼 is sufficiently increased the data is fitpoorly (as the SNR threshold is increased) and our Gaussian ringbrightness profile model may become a Gaussian centred on 𝑟 = 𝜒 values for these models, the differencein R peak values between 𝑟 = 𝑟 = 𝑅 ring is not a concern. Figure A1. Wavelength [mm] F l u x [ m J y ] =2.73 HD163296
Figure A2.
The spectral slope for HD163296. Purple markers represent0 . − . APPENDIX C: VARYING THE PHASE CENTRE
We present the visibility data and Frankenstein’s fit dependenceon the phase centre. For face-on sources with an obvious deficit ofcentral emission and a clear ring-like morphology, phase centring issimple. However, for low SNR inclined/edge-on sources finding theexact phase centre can be troublesome. Figure C1 clearly indicateshow changing the phase centre can either produce a Gaussian-like
MNRAS , 1–14 (2015) Norfolk et al.
Deprojected baseline (k ) R e a l n o r m ATCA 34GHz
Radius ["] B r i g h t n e ss n o r m =1.01 ws =0.01 =72.661 =1.01 ws =0.0001 =72.661 =1.05 ws =0.01 =72.663 =1.05 ws =0.0001 =72.664 =1.1 ws =0.01 =72.664 =1.1 ws =0.0001 =72.664 =1.3 ws =0.01 =72.67 =1.3 ws =0.0001 =72.664 =1.5 ws =0.01 =72.67 =1.5 ws =0.0001 =72.664 Deprojected baseline (k ) R e a l n o r m ALMA 281GHz
Radius ["] B r i g h t n e ss n o r m =1.01 ws =0.01 =1053.178 =1.01 ws =0.0001 =1053.207 =1.05 ws =0.01 =1053.201 =1.05 ws =0.0001 =1053.217 =1.1 ws =0.01 =1053.221 =1.1 ws =0.0001 =1053.283 =1.3 ws =0.01 =1053.609 =1.3 ws =0.0001 =1053.65 =1.5 ws =0.01 =1053.675 =1.5 ws =0.0001 =1053.676 Figure B1.
Frankenstein fit sensitivity to variations in both 𝛼 and w smooth , and the corresponding 𝜒 values. or Gaussian ring-like 1D radial brightness profile. It’s inconsequen-tial whether we to fit this non-parametrically with Frankenstein orwith a parametric model, this is an intrinsic difficulty in fitting a ra-dial profile to a non-axisymmetric disc with uncertain phase centre.Therefore, we suggest fitting these sources should be approachedwith caution. APPENDIX D: IMAGINARY COMPONENT OF THEVISIBILITIES
Here we include the normalised imaginary component of the visibil-ities as a function of the deprojected baseline for sources modelledwith Frankenstein seen in Fig. 2. For sources with relatively highSNR in the ATCA data (HD100453, SR24S, and HD169142) theimaginary component similarly has a high SNR and a low phaseinstability. For the rest of our sample, the imaginary component ofthe ATCA indicates large phase instabilities.
APPENDIX E: SOURCES WITHOUT FRANKENSTEINATCA FITS
We include the normalised imaginary component of the visibili-ties as a function of the deprojected baseline for ATCA data notmodelled with Frankenstein and the corresponding ALMA datawith Frankenstein models. We chose not to model the ATCA dataof these sources due to the low SNR ( ≤ 𝜎 ) of the observations,this is obvious as shown by the large phase instability in FigureE1. In Figure E2 the Frankenstein fits exhibit some deviation inthe reconstructed 1D radial brightness profiles. This is likely due to asymmetrical structure seen in each of the corresponding contin-uum maps. In our analysis we use the same ALMA observations asPinilla et al. (2018) for SZ Cha (Band 7), J1604 (Band 6), and SR21(Band 7). For HD34282 and HD143006, we make use of Band 6data and Pinilla et al. (2018) analyses Band 7 data. For HD34282our R peak value is comparable to the radial location of the peakemission presented in Pinilla et al. (2018) (see Table E1). Howeverour models for SZ Cha, HD143006, J1604, and SR21 deviate be-yond the 1 𝜎 error. The final calibrated visibilities for HD143006was sourced from the DSHARP survey (Andrews et al. 2018) andas a result, our analysis reveals more disc substructure in compar-ison to Pinilla et al. (2018). For SZ Cha, J1604, and SR21, ourcontinuum maps show a peak emission at r ≈ . , . , and0.355 arcseconds respectively, which is in close agreement with ourFrankenstein fit R peak values. For SZ Cha, we use robust = = − .
10 and dDec = − .
05. Thediscrepancy in the model peak brightness between this work andPinilla et al. (2018) for these three sources is likely due to limita-tions introduced by adopting a singular functional form of the discstructure to fit the visibilities parametrically. These limitations aremost evident in the choice of the parametric model profile whichwill typically poorly fit long baseline data ( ≥ 𝜆 ). The accuracyof the fit to long baseline data (on any scale) strongly influences therecovered profile’s features including the radial position of the peakbrightness. This paper has been typeset from a TEX/L A TEX file prepared by the author.MNRAS , 1–14 (2015) ust Traps and the Formation of Cavities in TDs Deprojected baseline (k ) R e a l n o r m Frankenstein ATCA fitATCA observations I m a g n o r m Radius ["] B r i g h t n e ss n o r m Deprojected baseline (k ) R e a l n o r m Frankenstein ATCA fitATCA observations I m a g n o r m Radius ["] B r i g h t n e ss n o r m mJy/beam mJy/beam Figure C1.
The effect of varying the phase centre of RY Lup on the normalised real and imaginary components of the visibilities as a function of thedeprojected baseline, normalised Frankenstein model radial brightness profiles, and the 8.8 mm ATCA continuum map at two different phase centres, withoffsets dRA = . (cid:48)(cid:48) and dDec = . (cid:48)(cid:48) . CSCha
HPCha
HD100453
HD135344B
RYLup
Deprojected baseline (k ) I m a g n o r m J1608
Sz111
SR24S
DoAr44
HD169142
ATCA observations ALMA observations
Figure D1.
Normalised imaginary component of the visibilities as a function of the deprojected baseline for sources modelled with Frankenstein seen inFig. 2.MNRAS , 1–14 (2015) Norfolk et al.
Deprojected baseline (k ) R e a l n o r m HD34282
Deprojected baseline (k ) I m a g n o r m Deprojected baseline (k ) R e a l n o r m SZCha
Deprojected baseline (k ) I m a g n o r m Deprojected baseline (k ) R e a l n o r m HD143006
Deprojected baseline (k ) I m a g n o r m Deprojected baseline (k ) R e a l n o r m J1604
Deprojected baseline (k ) I m a g n o r m Deprojected baseline (k ) R e a l n o r m SR21
Deprojected baseline (k ) I m a g n o r m Figure E1.
Normalised real and imaginary components of the visibilities as a function of the deprojected baseline for ATCA data with an SNR less than 5 𝜎 .MNRAS000
Normalised real and imaginary components of the visibilities as a function of the deprojected baseline for ATCA data with an SNR less than 5 𝜎 .MNRAS000 , 1–14 (2015) ust Traps and the Formation of Cavities in TDs Deprojected baseline (k ) R e a l n o r m HD34282 I m a g n o r m Radius ["] B r i g h t n e ss n o r m Deprojected baseline (k ) R e a l n o r m SZCha I m a g n o r m Radius ["] B r i g h t n e ss n o r m Deprojected baseline (k ) R e a l n o r m HD143006 I m a g n o r m Radius ["] B r i g h t n e ss n o r m Deprojected baseline (k ) R e a l n o r m J1604 I m a g n o r m Radius ["] B r i g h t n e ss n o r m Deprojected baseline (k ) R e a l n o r m SR21 I m a g n o r m Radius ["] B r i g h t n e ss n o r m -0.00 0.14 0.28 0.43 0.57 0.71 0.85 1.00 1.14 1.28 mJy/beam mJy/beam mJy/beam -0.00 0.49 0.99 1.48 1.98 2.47 2.97 3.46 3.96 4.46 mJy/beam mJy/beam ALMA observations Frankenstein ALMA fit 1 bootstrap ALMA
Figure E2.
The normalised real and imaginary components of the visibilities as a function of the deprojected baseline, normalised Frankenstein model radialbrightness profiles, and the sub-mm ALMA continuum map for sources not included in Figure 2.MNRAS , 1–14 (2015) Norfolk et al.
Table E1.
Radial brightness profile properties of the ALMA for sources notincluded in Figure 2. This includes the radial position of the peak brightnessR peak . Also included are the 1D profile peak positions from other models inthe literature and the observed inclination and position angle of the targetdisc. Superscripts in the column headers refer to the position in the referencelist of the last column.
Source 𝑅 peak obs(") 𝑅 peak Lit a (") Incl . b ( ° ) PA d ( ° ) Ref.(a,b)HD34282 0 . + . − . . + . − .
59 117 1,2SZ Cha 0 . + . − . . + . − .
47 154 1,1HD143006 0 . + . − . . + . − . . + . − . *0 . + . − . **J1604 0 . + . − . . + . − . . + . − . . + . − .
16 14 1,2
Notes: *Second Peak**Third Peak
References (1) Pinilla et al. (2018), (2) Francis & van der Marel (2020) MNRAS000