betaFeedback

WorksPhysics and Astronomy

ALMA detection of continuum emission from GQ Lup B

Abstract

GQ Lup B is a massive (10-40 MJup_{\text{Jup}}) companion orbiting the young T-Tauri star GQ Lup A at a projected distance of ∼\sim110 au (0.7''). It is outside the circumprimary disk in the submm continuum, but its ∼ 63∘{\sim}\,63^\circ inclined orbit intersects the primary's gas disk twice per period. We report a 5.8σσ detection of continuum emission at the location of GQ Lup B, obtained with the Atacama Large Millimeter/submillimeter Array (ALMA) at 0.89 mm (band 7). We associate this emission with a dusty disk around the companion, already detected by VLT/NACO and JWST/MIRI in the mid-Infrared. To date, this is the third disk observed at mm wavelengths around a companion spanning the brown dwarf and planetary mass regimes, following PDS 70 c, a gas giant embedded in the parent circumstellar disk, and SR 12 c, a planetary-mass companion at 1200 au from the primary. We retrieve a point source with peak intensity 54.5±9.4 μJy beam−154.5\pm9.4\,μ\rm Jy\, beam^{-1}, co-spatial with GRAVITY astrometry. The emission is spatially unresolved, placing an upper limit on the source radius of 2.3 au (from a synthesized beam of ∼ 29×25\sim\,29\times25 mas). For comparison, this is smaller than 1/3 of the Hill's radius (3-5 au for GQ Lup B, depending on the assumed mass), the theoretical upper limit for circum-planetary disks. Assuming optically thin thermal emission from mm-size grains in a disk-like structure, we infer a mass of 0.0032 M⊕_\oplus, while the optically thick limit yields an emission radius of 0.2 au. Assuming a gas to dust ratio of 100, our results indicate a very low disk mass to accretion rate ratio, of the order of ∼104\sim10^4 years, analogously to the highest accretors in the stellar mass regime. In the case of GQ Lup B, this can be explained by its periodic crossings of the circumprimary disk, which could deliver a substantial amount of gas to the companion's disk over a long timescale.

Introduction

Typical tracers of an active planet formation process are observed at optical and infrared wavelengths. UV excess, Hα\alpha emission and near-IR hydrogen recombination lines trace ongoing accretion and shocks, [81, 4, 26] while continuum excess, silicate features in the mid-IR and dust-induced polarization are linked to µm-sized dust grains in the most irradiated surroundings of the protoplanet [69, 22, 18].

Observing these tracers requires high angular resolution and contrast, hence the most accessible targets are massive, bright and well-separated planetary mass companions (PMCs, [42, 11]). These are gravitationally bound objects with masses broadly in the range 1–30 MJupM_{\mathrm{Jup}}, at separations ≳100\gtrsim100 au from the star and outside the primary disk. Despite forming via pathways more characteristic of stars, their near-planetary masses suggest that they share properties with giant planets, including temperatures, surface gravities, and chemical compositions. PMCs also represent an excellent laboratory to investigate the physics of disks around substellar mass objects. Disks around planetary-mass companions could provide a valuable and more accessible comparison to Circum-Planetary Disks (CPDs) [15, 73, 41, 64, 47] around planets embedded in the primary disks, helping us to understand their physical properties and evolution, and how they affect planetary formation mechanisms.

For many PMCs rich observational datasets already exist, spanning direct imaging, accretion rate measurements and infrared spectral energy distributions (SEDs), that could reveal the presence of a disk. [51, 22, 21, 25, 32] (Santamaría-Miranda et al. 2018; Bowler et al. 2020). However, none of these observations are suitable for investigating the bulk of a disk dust component, as the emission is likely optically thick and only probes the hottest layers of the disk, those near the surface and close to the central object. The main component of the dust reservoir instead is likely located in the colder region of the disk, in the midplane. It can be probed only by (sub)mm observations, which however may include contribution from free-free emission [27]. Detecting disks around PMOs and CPDs at millimeter wavelengths is extremely challenging and rare, as their typical dust mass seems to be lower than what can be probed by typical sensitivities of current surveys. Despite several ALMA campaigns targeting PMCs and embedded protoplanets [12, 40, 30] (Wu et al. 2017a; Wu et al. 2020), only two sources have been detected: PDS 70c [37, 8], inside the primary’s disk, and SR 12 c (Wu et al. 2022), one of the widest-orbit PMCs known to date. With this paper we present the detection of 0.89 mm continuum emission at the location of the massive companion GQ Lup B.

GQ Lup is located in the Lupus star forming region, at a distance of 154.1 pc [33]. GQ Lup A is a T-Tauri star with spectral type K7V, an age of 2-5 Myr and a mass of 1.03 M⊙_\odot [28, 46]. It hosts a circumstellar disk, whose dust component extends up to 59 ± 12 au, with an inclination with respect to the line of sight of 60.5 ± 0.5° [46]. Previous ALMA observations show signs of a dust gap at ∼10 au from the star [44] and an asymmetric CO disk ∼ 200 au wide [46]. Scattered light images with VLT-SPHERE show a single spiral arm in the southern region of the disk [69], likely caused by the interaction with the massive companion GQ Lup B. This 10-40 MJup_\mathrm{Jup} object identified in 2005 [54], is located at a projected separation of 0.7″ from the primary. Recent studies, based on almost twenty years of astrometry and precise radial velocity measurements, found that its orbit is inclined by 63−6+663^{+6}_{-6}° with respect to the circumstellar disk, with a semimajor axis of 97 au [70]. This orbital configuration suggests that the companion punctures through the gas component of the primary disk twice per orbit, perturbing its kinematics and originating asymmetries in the disk morphology [63]. Other than with the circumprimary disk, the companion’s orbit is misaligned also with the stellar spin axis (relative inclination ∼ 52−24+1952^{+19}_{-24}° [13, 70]) possibly implying that GQ Lup B formed outside of the primary disk, from the same parent cloud of the primary star. A similar 12^{12}CO/13^{13}CO ratio in the atmosphere of GQ Lup A (51−8+1051^{+10}_{-8}) and B (53−6+753^{+7}_{-6}) [55] indeed suggests they originated from a shared material reservoir. To complete the picture, there is a third object gravitationally bound to the system [1], GQ Lup C. It is a low mass star (0.15 M⊙_\odot) at a projected distance of 2400 au, which possibly participates in the intricate dynamical architecture of the entire system and may have played a role in determining the highly inclined orbit of GQ Lup B.

The presence of a disk around GQ Lup B has long been discussed. Shortly after its discovery in 2005, hydrogen emission lines showed that the companion was still accreting material [49]. The hypothesis was further strengthened by the detection of UV excess [81], tracing ongoing accretion, followed by mid-IR continuum excess which was explicitly linked with a disk-like structure [63]. Recent JWST/MIRI observations confirmed that GQ Lup B’s IR SED cannot be explained without accounting for the hot dust in a disk around the companion, with simple semi-analytical modeling suggesting a compact disk and moderately grown dust grains [22]. Previous ALMA observations set a 3σ upper limit of 150 μJy for ALMA band 7 continuum [46] (Wu et al. 2017b), leaving important disk parameters such as dust mass and structure unconstrained.

Observations and data reduction

We present new ALMA band 7 observations, combining two programs carried out between July 2024 and June 2025, targeting the GQ Lup system with 5 different baseline setups (ID 2023.1.01340.S and 2024.1.00780.S, PIs Ya-Lin Wu and Gabriele Cugno, respectively). In total, the observations consist of 16 Execution Blocks distributed over 13 nights. A summary of the MOUS (Member Observation Unit Sets) used in this work is presented in Table 1, including the baseline ranges, precipitable water vapor (PWV) at the time of the observation, frequency support and total integration time. Throughout this paper, different names will be used to refer to different MOUS: ACA refers to the data from the Atacama Compact Array; SB, IB, and LB refer to the Short, Intermediate, and Long baseline setup in program 2024.1.00780.S; wIB refers to the only MOUS of program 2023.1.01340.S.

IDNameDateBaselines (L5 – L80) (m)Frequency (GHz)Number of EBsPWV (mm)On source time (h)
2023.1.01340.SwIB2024/07/1399 – 1003332.81 – 348.7520.251.546
2024.1.00780.SSB2024/10/0333 – 222330.56 – 345.8210.890.672
2024.1.00780.SACA2025/04/07-109 – 32330.56 – 345.8230.911.739
2024.1.00780.SIB2025/04/28-2996 – 1056330.56 – 345.8230.912.167
2024.1.00780.SLB2025/06/10-16208 – 1915330.46 – 346.2870.814.872

Table 1. Summary of the observations used in this work.

Program 2023.1.01340.S was set up to optimize continuum detection. Its MOUS (wIB) consists of two Execution Blocks (EB) with four spectral windows each. Those spectral windows are centered in 333.787, 335.737, 347.736 and 345.782 GHz, with bandwidths of 1875, 1875, 1875 and 937.5 MHz respectively; the maximum spectral resolution is 0.565 MHz (∼ 500 m/s). Program 2024.1.00780.S was tuned to optimize continuum detection and also study gas kinematics, as part of a joint ALMA-JWST proposal. Because of this, the ACA, SB and IB MOUS have two narrow spectral windows at high spectral resolution (a detailed discussion on gas kinematics will be the subject of a following work). Their EBs have four spectral windows, centered in 330.567, 332.981, 343.627 and 345.774 GHz with nominal spectral resolution of 0.031, 0.977, 0.977, 0.031 MHz (∼ 0.03, 0.9, 0.9, 0.03 km/s) and bandwidth of 58.59, 1875, 1875 and 58.59 MHz. LB MOUS were instead set to maximize continuum sensitivity: each one has four spectral windows centered in 331.400, 333.300, 343.450 and 345.350 GHz, with 1875 MHz bandwidth and 0.977 MHz spectral resolution.

Self calibration

After calibrating the raw data with CASA 6.6.1.17 pipeline (version 2024.1.0.8, CASA Team et al. 2022), we performed self-calibration to improve the Signal to Noise Ratio (S/N) and increase chances of detecting the companion, adapting the procedure used by the exoALMA collaboration [65, 45] to our specific combination of datasets.

Pipeline-calibrated measurement sets were first quality checked in amplitude-vs-channel and amplitude-vs-baseline space. No major outliers were found at this stage and consequently no data were rejected. Subsequently, a pseudo-continuum measurement set was created for each execution block individually by flagging channels containing spectral line emission (12^{12}CO, 13^{13}CO, CS, SO) within ±15 km/s of the systemic velocity and averaging the remaining data into 250 MHz channels. We checked for phase decoherence in individual EBs by looking for sudden decreases in amplitude against time [45], which were observed in most of our datasets. Therefore we performed one round of phase-only self-calibration on each EB, combining scans and spectral windows over the full observation duration. This reduced non-thermal noise and improved peak S/N by up to 50%, providing cleaner input for the subsequent alignment step. We then proceeded to combine different execution blocks into a single measurement set. Initially we combined all of the data from program 2024.1.00780.S, which share a common pointing center and frequency support. We then added the two EBs from program 2023.1.01340.S to the combined measurement set.

More specifically, starting with the most recent program, 2024.1.00780.S, EBs were aligned in the uv-plane by minimizing the weighted difference between visibilities gridded onto a common grid, considering only cells with overlapping coverage. This was performed by leveraging the script by Loomis et al. [45]. We chose the best long baselines EB as reference (the one carried out on 2025/06/14, because of its almost circular beam and high S/N). Alignment then proceeded hierarchically: LB first, then IB referenced to the aligned LB, then SB to the aligned IB, then ACA to the aligned SB.

After spatial alignment, the self-calibration procedure was carried out starting from the shortest baselines EBs. We started by combining ACA and SB and performed self-calibration on the joint measurement set. Multiple phase-only rounds at decreasing solution intervals (360,180,120,60,30,18 s) were performed. To create the clean models used to calculate the gain tables we set a 6σ6\sigma threshold, masking an ellipse centered in the star’s position with semimajor axis 0.7 arcsec, minor axis 0.35 arcsec and PA 346∘346^\circ. After a full cycle of phase self-calibration, before adding longer baselines, flux offsets between EBs (from absolute calibration errors or baseline-dependent decoherence, choosing once again the highest S/N EB as reference) were corrected by rescaling the observed amplitudes. The whole phase self-calibration step was then started over with the rescaled EBs, checking that the flux offsets had been indeed corrected. The same process was then applied iteratively by progressively adding longer-baseline data (ACA+SB → ACA+SB+IB → ACA+SB+IB+LB), assessing phase decoherence and amplitude offsets at every step. After completing these rounds of phase-only self calibration, we applied two amplitude and phase self calibration rounds on EB-long and scan-long intervals.

The EBs from program 2023.1.01340.S were then added to the final measurement set with the same logic: first they were aligned in the uv plane to the visibilities obtained in the previous step, then phase only self-calibration was performed before rescaling fluxes; the process was started over on the rescaled EB and finally amplitude self-calibration was applied to the final measurement set. A 30 s binning was finally applied to produce the definitive measurement set used in all the imaging procedures described in the next paragraph.

Imaging

The final measurement set was imaged using the CASA v.6.5.4 tCLEAN algorithm. Given the very high dynamic range of the reconstructed images (≳2900\gtrsim2900), even minor bandwidth smearing can dominate the residuals away from the phase-center. We therefore used the Multi-Term Multi-Frequency Synthesis (mtmfs) deconvolver, with a first order polynomial (nterms = 2) and multi-frequency synthesis (mfs) spectral mode. Since we combined 7m with 12m MOUS, we used a mosaic gridder. To assign the clean components in the deconvolution algorithm, we masked the same ellipse used during self-calibration and a circle centered in the companion’s position (radius 0.1″, center RA offset 0.698″, center DEC offset 0.114″).

To assess the robustness of the structures we retrieve, we performed a grid of CLEANed images, using Briggs weighting with varying robust parameters. Depending on the robust parameter, our images have an rms ranging between 9.4 and 53.7 μ\muJy beam−1^{-1} with the corresponding beam size spanning from 0.13 to 0.02 arcsec. The rms was calculated on an emission-free region, inside an annulus extending between 2 and 6 arcsec from the pointing center.

Results

Millimeter continuum emission from the GQ Lup system

In Figure 1 we report the 890 μ\mum continuum image we retrieve after combining all of our data. The primary disk is relatively bright, with a spatially integrated flux density of 70.11 ±\pm 0.05 mJy, calculated in the image plane with robust = 2 (over an elliptical area with PA = 347°, semi-axes 0.4″ and 0.5″). This value is lower than the continuum flux reported in MacGregor et al. [46] (77.8 ±\pm 0.2 mJy), obtained by fitting a two-dimensional gaussian to the continuum emission.

Figure 1. 890 μm continuum image of GQ Lup (robust=2, beam ∼0.13×0.11′′\sim0.13\times0.11''). The contours represent the 3,4, and 5σ5\sigma contours. Given the very high dynamic range of the image (∼2900\sim2900), we used an Asinh normalization of the color scale.

Using the galario package, we fitted the visibilities modeling the disk with a central Gaussian with two external Gaussian rings, obtaining an inclination of 59.25 ±\pm 0.01° and a position angle of 348.46 ±\pm 0.02°, roughly consistent with the estimate obtained for the gas disk from the CO emission in MacGregor et al. [46] (ii = 60.5 ±\pm 0.5°, PA = 346 ±\pm 1°). The disk appears fairly compact, extending up to ∼\sim 50 au from the star. In Figure 1 we report 3, 4, and 5σ\sigma contours, which highlight the presence of extended, non axisymmetric emission in the northern and southern sides of the disk at the 5σ\sigma level. This emission, which remains visible even when varying the imaging parameters (see Fig. A.1), is not reported in previous ALMA observations, which however had lower S/N and larger beams. The possible physical origin of these structures will be discussed in a forthcoming work, which will include different fitting procedures in the visibility plane.

Most notably, a point source is visible at the companion’s location. The most significant detection (5.8σ\sigma) is obtained with robust = 2, with a peak intensity of 54.5 μ\muJy beam−1^{-1} and a 9.4 μ\muJy beam−1^{-1} rms calculated in the same emission-free annulus described in the previous section. We also calculate the rms in an annulus between 0.5″ and 0.8″ from the star, excluding a circle with 0.1″ radius centered in the companion’s position. Over this area, the rms is 14.3 μ\muJy, lowering the significance of the detection to 3.8 σ\sigma. We note, however, that this last annulus does not appear emission-free, as there is a faint filament visible at a ∼3σ\sim3\sigma level in the east side of the disk (Figure 2). This structure does not seem to be produced by imaging artifacts, as is apparent from the residuals images in Fig. B.1. We also note some structure on the east side of the companion’s position, also at a ∼3σ\sim3\sigma level.

Figure 2. Left: scattered light image of GQ Lup taken with VLT/SPHERE (van Holstein et al. [69]). A white contour (at 40% of the peak of the image) serves as a reference for the other panels. The single spiral arm visible in the southern part of the image is attributed to the companion, which appears in the right side of the image. Although useful for spatial comparison, the apparent polarized flux of GQ Lup B is caused by the subtraction of the stellar component, as explained in the original paper (van Holstein et al. [69]). Center: saturated version of the continuum image from Figure 1, where the color scale has been limited to a range of 0 to 50 μ\muJy beam−1^{-1}. The dashed gray square, with sides of 0.5″, is centered on GQ Lup B. Right: zoomed-in view of the same region as in the central panel: the red X marks the companion’s position from GRAVITY on 2023/3/8 (Venkatesan et al. [70]), while the white contour represents the scattered light contour.

In Figure 2 we show a comparison with the position retrieved by GRAVITY astrometry (Venkatesan et al. [70]) and with a scattered light image obtained with SPHERE (van Holstein et al. [69]; Stolker et al. [63]). The peak of the intensity in our ALMA image is ∼\sim 0.02″ from the position from GRAVITY reported in Figure 2 (2023/3/8) and ∼\sim 0.01″ from the peak of the image from SPHERE. The latter is useful for spatial comparison, but as noted in the original article (van Holstein et al. [69]), GQ Lup B’s emission is not polarized up to ∼\sim 0.2%. The apparent signal is instead spuriously created by subtracting the stellar polarization. Since all observations were taken within a year from July 2024 to June 2025, and given the companion’s orbital period of ∼\sim 1000 years (Stolker et al. [63]; Venkatesan et al. [70]), we can safely assume that its position did not change significantly between different measurement sessions. Specifically, GRAVITY data collected between August 2021 and March 2023 (Venkatesan et al. [70]) reveal a proper motion of GQ Lup B relative to its host star of ∼\sim 4 mas in the same time span, five times smaller than the smallest synthesized beam we can obtain. We therefore ignore any proper motion in our analysis. The coincidence of the point source seen with ALMA to both the IR signals suggests that their origin is co-spatial. We interpret this emission as thermal emission from millimeter-sized grains in an unresolved disk surrounding GQ Lup B.

To verify that the emission is not due to imaging artifacts we carried out two sets of tests. Following Benisty et al. [8], we produced multiple images varying the robust parameters (and consequently the beam sizes, see Fig. A.1 in the appendix). If the emission was due to the contamination from GQ Lup A’s disk, the peak intensity would increase with more extended beams, since the resolution element would include a higher contribution from the primary. This is not the case for GQ Lup B: regardless of the beam size the observed peak intensity remains stable within a 1σ1\sigma interval from the reported value of 54.5 μ\muJy beam−1^{-1} (see Figure 3). The constant peak intensity also indicates that the emission is unresolved independently of the imaging parameters. Furthermore, Fig. C.1 highlights how the radial profile of the emission behaves as a point source with peak intensity just above the cleaning threshold. The implications for the disk’s size are discussed in the next section.

Peak intensity at the companion’s location obtained with different Briggs robust parameters

Figure 3. Peak intensity at the companion’s location obtained with different Briggs robust parameters, which translate into different beam sizes.

The images we obtain have a very high dynamic range, with the peak intensity of the primary disk reaching up to 2900 times the rms noise level. When searching for faint structures, this must be treated with caution, as sidelobes from the bright primary could potentially be mistaken for real emission. To address this, we produced an additional image set masking only the primary disk (see Figure 4 and B.1 in the appendix), preventing the cleaning algorithm from placing any CLEAN components at the planet’s location. Since the algorithm iteratively subtracts model components convolved with the dirty beam, which is responsible for the sidelobes, any spurious emission arising from those sidelobes should be progressively reduced as the cleaning proceeds deeper. If the 5.8σ5.8\sigma point source was indeed an artifact, we would expect its flux density in the residual image to decrease with the clean stopping threshold. Instead, the flux density measured at the companion’s location in the residual image remains constant, further indicating that the detected emission is not related to the primary’s sidelobes.

Peak intensity at the companion’s location in the residual map

Figure 4. Peak intensity at the companion’s location in the residual map, masking only the primary disk and stopping the imaging procedure at different thresholds.

Discussion

Our detection represents the third observation of continuum emission from a bound planetary-mass object in the millimeter regime, the second around a wide-orbit PMC (Wu et al. 2022). The mid-IR excess observed in the GQ Lup B’s SED has previously been associated with a dusty disk ([22]). Therefore, it is intuitive to associate the emission we retrieve at 890 μ\mum with thermal emission from dust grains in a disk-like structure. However, recent studies on PDS 70 c ([27, 16]) highlighted that other mechanisms (e.g. free-free emission from accretion shocks) could also play a role at submm wavelengths. Only a multi-band characterization and thorough radiative transfer modeling can distinguish between different scenarios. In the next sections, we discuss the implications of our single measurement within the assumption that the point source we detect is caused only by thermal emission from a dusty disk.

We note once again that GQ Lup B’s IR emission observed with SPHERE is not polarized. If we consider only emission from a disk, this could imply a low inclination ([69]) or alternatively relatively grown dust grains (a≳5 μa \gtrsim5\,\mum, [22]).

Flux comparison with stars and PMOs

In Figure 5 we compare the submm flux density of GQ Lup B to other disks around stars and planetary mass objects, rescaled to a common frequency (240 GHz) and distance (150 pc). We adopt a spectral index of 2.2 for protoplanetary disks ([2]) and 2 for CPDs and disks around PMOs, as measured for PDS 70 c ([31, 27]).

Figure 5. Measured fluxes and upper limits for stars and planetary-mass objects, rescaled to a common frequency and distance, as a function of the host mass. Stars (Manara et al. [48]) are reported as circles, while squares refer to planetary-mass objects. Other than PDS 70c and SR 12c, we report the free floating OTS 44 (Bayo et al. [7]) and the low-mass brown dwarf 2MASS J12073346-3932539 A (2M1207A) (Ricci et al. [57]). The faded dot represents the tentative detection of PDS 70 b (Benisty et al. [8]; Fasano et al. [31]). Triangles represent 5σ5\sigma upper limits for GSC 6214-210 b (Bowler et al. [12]), DH Tau b, CT Cha b and FU Tau b (Wu et al. 2020), and for the embedded planets WISPIT 2c and WISPIT 2b (Benisty et al. [9]; Facchini et al. [30]). Red represents our measurement of GQ Lup B. The color scale accounts for the host’s accretion rate (see discussion in Sect. 4.3). Stars without an accretion rate measurement are colored in gray. Disks around PMOs seem to broadly follow the relation found for stars, represented by the dashed gray line. For GQ Lup B we assume a mass of 20 MJupM_{\mathrm{Jup}} as a mean value between different estimates, which vary between 10 and 30 MJupM_{\mathrm{Jup}} (Marois et al. [49]; Seifahrt et al. [61]; Stolker et al. [63]; Cugno et al. [22])

In general, disks around stars seem to follow a common trend, with slope β=1.43±0.10\beta= 1.43 \pm0.10 and intercept α=1.37±0.06\alpha= 1.37 \pm0.06 ([3, 48, 30]), with a considerable scatter around the best fit relation. This large scatter, together with the scarce number of detections, makes it premature to assess whether also planetary-mass objects follow the same relation. GQ Lup B exhibits lower peak intensity in ALMA band 7 than PDS 70c and SR 12c (54.5 ±\pm 9.4, 86 ±\pm 16 and 127 ±\pm 14 μ\muJy beam−1^{-1}, respectively), but when rescaled to a common distance the measurements seem to align (28.6 ±\pm 4.9, 23.2 ±\pm 4.4 and 34.3 ±\pm 3.8 μ\muJy beam−1^{-1}, at 240 GHz if the sources were at 150 pc). These already available detections indicate the sensitivity threshold necessary to detect disks around substellar objects at mm wavelengths, increasing the sample needed to understand similarities and differences between circumstellar and circumplanetary physics.

Mass and radius

Constraints on the radial extent Since the emission around GQ Lup B is unresolved, only limited constraints on its size and mass can be derived. An agnostic estimate on the disk size can be given directly by inspecting the submm continuum images. As shown in Figure 3, the disk remains unresolved even at the highest angular resolution achievable with our dataset (∼0.03′′\sim0.03^{\prime\prime}): if it was resolved, we would expect the peak intensity to drop as the angular resolution increases. Instead, the peak intensity remains constant across all the images independently of the synthesized beam size. Varying the robust parameter, we can improve the angular resolution at the cost of increasing the noise. The smallest synthesized beam for which the signal-to-noise ratio exceeds 3 is ∼29×25\sim29 \times25 mas, obtained with robust -0.5. At the system’s distance of 154.1 pc, this corresponds to approximately 4.6 au. Therefore we adopt ∼2.3\sim2.3 au as a conservative upper limit on the outer radius of the dust component of the disk. This is slightly smaller than one third of the companion’s Hill radius (3-5 au, when adopting the most recent orbital parameters from Venkatesan et al. [70]), which is usually cited as the theoretical upper limit for embedded CPDs (Martin & Lubow [50]; Shabram & Boley [62]). The upper limit we find for GQ Lup B is also comparable with the upper limit of 5 au for SR 12 c and 1.2 au for PDS 70c, confirming that independently of the separation from the primary, these disks are likely compact in their (sub)mm continuum emission.

Simulations in Rab et al. [56] showed that PMC disks can be optically thick at millimeter wavelengths. Under this assumption, we can obtain an estimate for the disk outer radius, since in this regime the observed flux density scales with the emitting surface. Following Isella et al. [38, 37]; Wu et al. 2022 we start by assuming a temperature profile for the disk’s dust:

T(r)4=T⋆,irr4+Tirr,p4+Tacc4.(1)T(r)^4 = T_{\star,\mathrm{irr}}^4 + T_{\mathrm{irr,p}}^4 + T_{\mathrm{acc}}^4. \tag*{(1)}

The first component is an approximation of the stellar contribution, which we compute as

Tirr,⋆4(R)=ϕL⋆8πR2σSB.(2)T_{\mathrm{irr,\star}}^4(R) = \frac{\phi L_{\star}}{8\pi R^2 \sigma_{\mathrm{SB}}}. \tag*{(2)}

This is the simplified expression for a passively heated, flared disk in radiative equilibrium (Chiang & Goldreich [17]; D’Alessio et al. [24]; Dullemond et al. [29]), where L⋆L_{\star} is the stellar bolometric luminosity and σSB\sigma_{\mathrm{SB}} is the Stefan-Boltzmann constant. In the original papers ϕ\phi is the flaring angle of the disk surface: in this case it represents the fraction of stellar radiation that hits the disk surface due to its inclination, and we take it as the sin of the grazing angle between the irradiation and the disk plane; dd is the distance between the companion and the star. Also the companion contributes to the total temperature profile Isella et al. [37]:

Tirr,p4=μLp4πσSBr2(3)T_{\mathrm{irr,p}}^{4} = \frac{\mu L_{\mathrm{p}}}{4\pi\sigma_{\mathrm{SB}}r^{2}} \tag*{(3)}

but in this case μ\mu is the flaring angle of the surface of the disk (which is unknown and we set it to 0.1), LpL_{\mathrm{p}} the bolometric luminosity of the planet and rr the radial distance from the companion. All the values that we adopt are reported in the appendix, in Table D.1. Finally,

Tacc4(r)=3GMpM˙acc8πσSBr3(1−rpr)(4)T_{\mathrm{acc}}^{4}(r) = \frac{3GM_{\mathrm{p}}\dot{M}_{\mathrm{acc}}}{8\pi\sigma_{\mathrm{SB}}r^{3}}\left(1-\sqrt{\frac{r_{\mathrm{p}}}{r}}\right) \tag*{(4)}

represents the contribution to the temperature profile given by accretion heating (D’Alessio et al. [24]). MpM_{\mathrm{p}} is the companion’s mass, M˙acc\dot{M}_{\mathrm{acc}} its accretion rate and rpr_{\mathrm{p}} its radius.

In the case of GQ Lup B, we find that for radii smaller than 2.3 au, the irradiation from the companion itself (L=10−2.16 L⊙L = 10^{-2.16}\,L_{\odot}) is the dominant term (see Figure 6). The contribution of the accretion term to the global temperature profile is negligible, while the irradiation from the star, which effectively acts as a background, plays only a marginal role. We note that in our models this background term strongly depends on the grazing angle, which is unknown. While for GQ Lup B and SR 12 c this is almost irrelevant (we set it to 45∘45^{\circ}), the stellar irradiation term completely dominates the temperature profile of CPDs closer to the star and more accurate modeling should also take into account shadowing from the primary disk.

Figure 6. Different contributions to the temperature profile of GQ Lup B. The red line, which represents the total temperature profile, interrupts at the upper limit we retrieve from the beam size.

Having a temperature profile, we are able to convert this to an emergent intensity assuming the full Planck law (Bν(T(r))B_{\nu}(T(r))). The total flux FνF_{\nu} can be obtained by integrating from a chosen inner radius rinr_{\mathrm{in}} to the outer radius routr_{\mathrm{out}}:

Fν=2π⟨cos⁡i⟩d2∫rinroutBν(T(r))r dr,(5)F_{\nu} = \frac{2\pi\langle\cos i\rangle}{d^{2}}\int_{r_{\mathrm{in}}}^{r_{\mathrm{out}}} B_{\nu}(T(r))r\,\mathrm{d}r, \tag*{(5)}

where dd is the distance of the system and ii is the disk inclination with respect to the line of sight. We find that the choice of the inner radius is not relevant for the final result, as the dust temperature decreases relatively slowly over radius. Finally, we use Eq. (5) to compute the outer radius which best reproduces the observed mm flux. All the parameters we used are listed in Appendix A in Table D.1. We obtain an outer radius of 0.20 au (418 RJupR_{\mathrm{Jup}}), smaller than SR 12 c (0.34 au) and PDS 70 c (0.38 au, setting ϕ=1∘\phi= 1^{\circ}). As pointed out in Wu et al. (2022) however, this value likely represents a lower limit: if dust emission was optically thin, it would scale with the dust mass and a more extended disk could have a similar brightness.

Mass estimate from optically thin assumption While in the optically thick regime the flux density scales as the emitting surface, in the optically thin assumption it scales with the total mass. In the most simple scenario, at frequency ν\nu the mass of optically thin dust is related to its flux density FνF_{\nu} by Hildebrand [36]:

Mdust=Fνd2κνBν(Tmean),(6)M_{\mathrm{dust}} = \frac{F_{\nu}d^{2}}{\kappa_{\nu}B_{\nu}(T_{\mathrm{mean}})}, \tag*{(6)}

where κν\kappa_{\nu} is the dust opacity and Bν(Tmean)B_{\nu}(T_{\mathrm{mean}}) is the Planck function evaluated at a mean representative temperature TmeanT_{\mathrm{mean}}. For the dust opacity we assume mm-sized grains (κν=3.63 cm2 g−1\kappa_{\nu}=3.63\ \mathrm{cm}^{2}\ \mathrm{g}^{-1}, Birnstiel [10]). We calculate a mean temperature of the disk of ∼56\sim56 K by averaging the temperature profile (Eq. (1)) from a minimum radius of 4 RJupR_{\mathrm{Jup}} (a mean estimate of the companion’s radius, Demars et al. [26]) to the upper limit of the disk’s radial extent, 2.3 au. We find a mass of 0.0032 M⊕M_{\oplus}, less than half of and a quarter of the analog estimates we obtain for PDS 70 c and SR 12 c (0.006 and 0.012 M⊕M_{\oplus}), respectively. Our estimates for PDS 70 c and SR 12 c agree with what is obtained in Benisty et al. [8] and Wu et al. (2022), but we note that these values, especially for PDS 70 c, strongly depend on the temperature profile, which is set by the choices on the parameters in Table D.1.

Accounting for the optical depth A more comprehensive approach to infer disk properties from a single submillimeter flux measurement is to assume a dust surface density profile Σ(r)\Sigma(r) to account for the dust optical depth. Once again following Isella et al. [38, 37] and Wu et al. (2022), Eq. (5) becomes

Fν=2πcos⁡id2∫rinrout[1−exp⁡(−Σ(r)κνcos⁡i)]Bν[T(r)]r dr,(7)F_{\nu} = \frac{2\pi\cos i}{d^{2}}\int_{r_{\mathrm{in}}}^{r_{\mathrm{out}}}\left[1-\exp\left(-\frac{\Sigma(r)\kappa_{\nu}}{\cos i}\right)\right]B_{\nu}[T(r)]r\,\mathrm{d}r, \tag*{(7)}

where we assume Σ=Σ0r−1\Sigma= \Sigma_{0}r^{-1}. We then calculate the flux density for a grid of disk dust masses MdM_{\mathrm{d}} and outer radii routr_{\mathrm{out}}. Figure 7 shows the combinations of these parameters for which the calculated FνF_{\nu} equals the observed peak intensity of GQ Lup B. We assume once again mm dust grains with opacity κν=3.63 cm2 g−1\kappa_{\nu}=3.63\ \mathrm{cm}^{2}\ \mathrm{g}^{-1}. In the same figure we show the same curve for PDS 70c and SR 12 c, together with dashed lines representing upper limits on other candidates for disks around PMOs. All of the parameters we used are listed in Table D.1. In particular, at a first approximation we assume PDS 70 b and c and WISPIT 2 b and c to be aligned with the primary disk: we set the grazing angle to 1∘1^{\circ}, and for ii we use the primary disk’s inclination. All the other objects (DH Tau b, GQ Lup B, FU Tau b, SR 12 c, CT Cha b) are wide orbit Planetary Mass Companions, and a priori there are no constraints on their disk’s orientation. Hence, for them we set both angles to 45∘45^{\circ}. As highlighted by Figure 6, for PMCs the temperature profile is dominated by the companion’s irradiation, and the grazing angle of the primary’s irradiation plays only a negligible role in the observed emission. Also the inclination with respect to the L.O.S. plays a marginal role, except for highly inclined disks (i≳80∘i \gtrsim80^\circ).

Figure 7. Continuous lines represent different combinations of dust masses and radial extents that can reproduce the observed millimeter flux density for GQ Lup B, SR 12 c and PDS 70 c. The lines stop with a vertical dash, representing the maximum radius estimated from the angular resolution of the observations. Dashed lines represent upper limits on other candidates (Bowler et al. 2015; Wu et al. 2020; Facchini et al. 2026; Benisty et al. 2026). All the parameters are listed in Table D.1.

Although coming from a relatively simple model, these curves highlight the different possible structure of the disks. A small source means high densities and optically thick dust: its emission is related only to the emitting surface, giving no constraints on the total mass. If the dust disk is instead larger, it quickly becomes optically thin, and the observed emission is related to its total mass. The curves behave differently at large radii because of the assumed temperature profile: emission is related to the temperature, which depends on the distance from the companion in different ways according to which term of Eq. (1) dominates. If the major component is the stellar irradiation, TT no longer depends on rr, the curve becomes flat and we have a direct measure of the disk mass (as in PDS 70 b and WISPIT 2 b and c). However, if the stellar irradiation is low (as for SR 12 c), at large rr the temperature drops and larger masses are needed to reproduce the observed flux density. In this case inferring the disk mass is not trivial.

While little can be said regarding the exact mass and radius of the dust disk around GQ Lup B, in Figure 7 its curve lies below those of SR 12 c and PDS 70 c, naively suggesting a smaller or less massive disk. This can be explained considering the relatively high irradiation from the companion (see the third column in Table D.1), which heats up the dust increasing the mm flux density.

Accretion rate

Manara et al. [48] compares accretion rates for planetary-mass objects with stars in nearby star-forming regions. Following Manara et al. [48], on the x-axis we plot an estimate of the total disk mass, using Eq. (6) with T=20T = 20K to infer the dust mass and multiplying it by 100 to account for the gas reservoir, in the assumption that the gas-to-dust mass ratio is uniform across the sample. Dashed gray lines indicate the survival timescale of a gas disk with the reported mass and accretion rate. For stars, evolution models predict a correlation between the accretion rate and the global disk mass ([35, 39, 58]). Large samples from ALMA and spectroscopic surveys revealed that the relation is indeed typically distributed around Mdisk/Macc=1M_{\rm disk}/M_{\rm acc} = 1 Myr with significant scatter (Figure 8), and does not appear to depend on stellar mass (Manara et al. [48]). Initial results suggest that brown dwarfs are systematically located in the lower part of the distribution (Sanchis et al. [59]), though the sample remains limited.

Accretion rates for disks around stars and planetary-mass objects

Figure 8. Accretion rates for disks around stars (circles) and planetary-mass objects (squares, the same discussed in Figure 5), against dust mass estimates (Eq. (6), T=20T=20 K) multiplied by 100 to account for the gas mass. Triangles represent upper limits. The color scale here accounts for the mass of the disk host, while dashed gray lines represent depletion timescales.

The few planetary-mass objects for which mm detections are available (the same listed for Figure 5, reported as squares in both figures) appear broadly consistent with the stellar trend, but the limited sample and large scatter once again prevent any firm conclusion. We omit error bars on mass accretion rates (M˙acc\dot{M}_{\rm acc}) because of their strong correlation with the models and parameters used: different estimates from the same measurement can vary up to an order of magnitude. Nonetheless, GQ Lup B appears on the high end of the scatter, showing a relatively high accretion rate (∼10−9.5 M⊙ yr−1\sim10^{-9.5}\,M_{\odot}\,{\rm yr}^{-1}; Stolker et al. [63]) for its mm-flux density. Notably, the mean temperature derived in the previous section (∼56\sim56 K) is significantly higher than the 20 K assumed here: higher temperatures imply that the observed mm-flux density is compatible with even lower disk masses, which would shift the red dot in Figure 8 further to the left.

Assuming the accretion rate to be constant in time, the gas component of the disk would be depleted in less than ∼104\sim10^4 yr, two orders of magnitude shorter than the system’s age (3–5 Myr, MacGregor et al. [46]). Two main observational caveats apply to this depletion timescale. First, the inferred MM depends on the correction adopted for the extinction (AV∼2.3A_V \sim2.3 mag); previous work adopting lower estimates (Wu et al. 2015), inferred an accretion rate lower by an order of magnitude. Second, line emission studies show that accretion in GQ Lup B is likely variable (Stolker et al. [63]; Demars et al. [26]), implying that the time-averaged accretion rate could be lower than the single value shown here. Even accounting for these caveats, the depletion timescale assumes GQ Lup B to be isolated, which is not the case. Recent orbital studies (Veneti et al. 2025) retrieved an orbital period of ∼1000\sim1000 yr and a highly inclined orbit (∼63∘\sim63^\circ with respect to the primary disk) with a 97 au semimajor axis. Since the CO disk seems to extend up to 200 au (MacGregor et al. [46]), the companion would puncture through the primary’s gas disk twice per orbit, with a timescale of ∼500\sim500 yr. This periodic interaction could continuously replenish the gas reservoir of the disk of GQ Lup B, naturally explaining the observed high accretion rate due to an increased inflow of material. However, the companion disk mass is inferred from the dust emission, whereas the PMO does not appear to cross the circumprimary mm-continuum disk. If the gas-to-dust ratio of the outer regions of the circumprimary disk is significantly above the factor of 100 assumed to produce Figure 8, the gas mass of the disk orbiting GQ Lup B may be significantly higher than anticipated.

GQ Lup B is not the only planetary-mass companion with an accretion rate high enough to suggest gas depletion on timescales shorter than the system’s age. CT Cha b ([21] (Wu et al. 2015)) is also located in the upper region of Figure 8: given the upper limit on its mm flux density, its disk mass could even be lower than that of GQ Lup B, making the implied accretion timescale even shorter and the tension with the companion’s age even more pronounced. CT Cha b lies at a projected separation of ∼511\sim511 au from its host star, well outside the primary disk, meaning that its gas reservoir cannot be continuously replenished as in GQ Lup. This makes the persistence of an actively accreting disk even harder to explain, and suggests that additional mechanisms as late infall may be at play. A thorough discussion of accretion models is beyond the scope of this paper. We note, however, that some theoretical frameworks predict episodic accretion ([6, 71, 19]), in which extended periods of near-quiescence are interrupted by short, intense bursts of accretion triggered by gravitational and magnetorotational instabilities. If such variability is common among PMCs, as seems plausible by recent findings ([23, 19, 80, 32]), instantaneous accretion rate measurements may not be representative of the time-averaged value, potentially alleviating the tension with the disk mass estimates. We have also neglected the role of satellite formation, which can alter the disk structure and thus the observational signatures accessible to current facilities. To carry out our analysis, we had to make use of several assumptions (dust opacity, gas mass, inclination, radial extent, temperature), which often have no direct observational measurement. Multi-wavelength characterization of Planetary-Mass Companions is therefore crucial to robustly constrain properties of disks around PMOs.

Conclusions

In this work we analyze ALMA continuum observations at 0.89 mm of the GQ Lup system, which hosts one of the best-known planetary-mass companions, GQ Lup B. Our main results are the following:

  • We report the continuum detection of GQ Lup B. We retrieve a 5.8σ5.8\sigma point source and, following the analysis carried out for the mid-IR excess, we associate it with the unresolved dust in a disk around the companion. This is the third millimeter detection of a gravitationally bound planetary-mass object, after PDS 70 c and SR 12 c.

  • The emission is co-spatial with both GRAVITY astrometry and SPHERE scattered light emission, and shows a peak intensity of 54.5±9.4 μ54.5 \pm9.4\,\muJy.

  • The source is unresolved at the best angular resolution we can achieve, implying a 2.3 au upper limit on the emission radius.

  • In the optically thick approximation, we retrieve a dust disk radius of 0.2 au, whereas in the optically thin regime we estimate a dust mass of 0.0032 M⊙0.0032\,M_{\odot}.

  • GQ Lup B exhibits a remarkably high accretion rate for its millimeter flux density: this may be partly sustained by its eccentric orbit intersecting the primary’s gas disk every ∼500\sim500 yr, replenishing the circumplanetary gas reservoir.

Our millimeter detection completes an already rich dataset, including NIR and MIR spectral energy distributions, and upcoming JWST/NIRSpec 3 – 5 μ\mum spectra and JWST/MIRI mid-IR photometry. A natural next step would be to make use of this wealth of information in a thorough radiative transfer model, taking into account all the already available wavelengths. We also note that multi-frequency mm observations are needed to distinguish between different emission mechanisms (thermal emission from cold dust grains or free-free emission from accretion shocks). All of this is left for future works, while here we only focus on the 890 μ\mum flux density we observe with ALMA, considering only dust thermal emission.

The growing set of millimeter detections of planetary-mass objects is starting to provide a statistical overview, spanning a wide range of orbital configurations. In this frame, PMCs are emerging as a crucial laboratory to investigate the physics of disks around substellar objects, bridging circumstellar and circumplanetary environments. Our detection demonstrates the sensitivity required to detect these faint structures at submillimeter wavelengths: enlarging the current sample will be crucial to assess whether disks around substellar objects are similar to scaled-down versions of circumstellar disks, or additional physical processes govern their formation and evolution.

Acknowledgements. This paper makes use of the following ALMA data: ADS/JAO.ALMA 2023.1.01340.S, 2024.1.00780.S. ALMA is a partnership of ESO (representing its member states), NSF (USA) and NINS (Japan), together with NRC (Canada), NSTC and ASIAA (Taiwan), and KASI (Republic of Korea), in cooperation with the Republic of Chile. The Joint ALMA Observatory is operated by ESO, AUI/NRAO and NAOJ. S.F. acknowledges financial contribution from the European Union (ERC, UNVEIL, 101076613). Views and opinions expressed are however those of the authors only and do not necessarily reflect those of the European Union or the European Research Council. Neither the European Union nor the granting authority can be held responsible for them. P.C. acknowledges support by the ANID BASAL project FB210003. J.B. is supported in part by JWST Cycle 3 Program JWST-GO-06739 provided by NASA through a grant from the Space Telescope Science Institute, which is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS5-03127. M.B. has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (PROTOPLANETS, grant agreement No. 101002188).

Imaging varying robust

Figure A.1. Images obtained with different Briggs robust parameters, which translate into different beam sizes. The bottom row represents the zoomed-in area around the companion, indicated with a gray square in the top row. The color scale is the same as in Figure 2.

Residuals of imaging varying threshold, masking only the primary disk

Figure B.1. Residual map of the images obtained masking only the disk of the primary star and stopping the imaging procedure at different thresholds. The bottom row represents the zoomed-in area around the companion, indicated with a gray square in the top row. The color scale is the same as in Figure 2, and the maps are the same used for Figure 4.

Radial profile comparison

Figure C.1. Fig. C.1. Azimuthal average of the clean beam and the PSF of the image obtained with robust = 2 (Figure 1), compared with the azimuthally averaged profile of the emission, centered in the position of the companion (taken as the position of the peak of the residual map). The radial profile broadly follows the PSF profile, coherently with an unresolved point source slightly above the cleaning threshold.

Parameters used in modeling

PMCMpM_{\mathrm{p}} (MJupM_{\mathrm{Jup}})log⁡10(Lp/L⊙)\log_{10}(L_{\mathrm{p}}/L_{\odot})M˙acc\dot{M}_{\mathrm{acc}} (M⊙ yr−1M_{\odot}\,\mathrm{yr}^{-1})L⋆L_{\star} (L⊙L_{\odot})DD (pc)dd (au)F345 GHzF_{345\,\mathrm{GHz}} (mJy)
GQ Lup B20 [1]−2.16-2.16 [2]10−9.510^{-9.5} [1]0.8 [3]154 [4]97 [5]0.054
PDS 70 b4 [6]−3.80-3.80 [7]1.30×10−111.30 \times10^{-11} [8]0.36 [7]112 [4]22 [9]0.080 [10]
PDS 70 c7 [6]−3.80-3.80 [7]10−1110^{-11} [8]0.36 [7]112 [4]34 [9]0.086 [10]
SR 12 c16 [11]−2.95-2.95 [12]10−1110^{-11} [12]0.12 [12]112.5 [4]1200 [11]0.127 [12]
WISPIT 2 b4.9 [13]−3.5-3.5 [14]2.25×10−122.25 \times10^{-12} [14]0.7 [14]133 [4]57 [14]<0.040< 0.040 [15]
WISPIT 2 c9 [16]−3.6-3.6 [16]<10−13< 10^{-13} [16]0.7 [14]133 [4]15 [16]<0.040< 0.040 [15]
FU Tau b19 [17]−2.51-2.51 [17]6×10−116 \times10^{-11} [17]0.04 [18]126 [4]726 [17]<0.200< 0.200 [19]
DH Tau b14 [19]−2.72-2.72 [20]5.01×10−125.01 \times10^{-12} [20]0.27 [19]135 [4]330 [19]<0.200< 0.200 [19]
CT Cha b17 [19]−2.62-2.62 [19]6.00×10−106.00 \times10^{-10} [21]0.57 [19]190 [4]440 [19]<0.090< 0.090 [19]
GSC 6214-210 b14 [22]−2.8-2.8 [22]7.00×10−127.00 \times10^{-12} [20]0.38 [22]145 [4]330 [22]<0.367< 0.367 [22]

Table D.1. Table D.1: Parameters used to calculate the disk temperature profiles.

Notes. The original measurements with the relative uncertainties can be found in the references listed below.

References. [1] Stolker et al. [63]; [2] Marois et al. [49]; [3] Demars et al. [26]; [4] Gaia Collaboration et al. [33]; [5] Venkatesan et al. [70]; [6] Wang et al. [72], Trevascus et al. [67]; [7] Müller et al. [53]; [8] Haffert et al. [34]; [9] Mesa et al. [52]; [10] Benisty et al. [8]; [11] Finley et al. [32]; [12] Wu et al. (2022); [13] van Capelleveen et al. [68]; [14] Close et al. [20]; [15] Facchini et al. [30]; [16] Lawlor et al. [43]; [17] Wu et al. (2023); [18] Bailer-Jones et al. [5]; [19] Wu et al. (2020) (Table 2); [20] Zhou et al. [81]; [21] Wu et al. (2015); [22] Bowler et al. [12].

References

  1. [1]Alcalá, J., Majidi, F., Desidera, S., et al. 2020, A&A, 635, L1
  2. [2]Andrews, S. M. 2020, ARA&A, 58, 483
  3. [3]Andrews, S. M., Elder, W., Zhang, S., et al. 2021, ApJ, 916, 51
  4. [4]Aoyama, Y., Marleau, G.-D., Ikoma, M., & Mordasini, C. 2021, ApJ, 917, L30
  5. [5]Bailer-Jones, C., Rybizki, J., Fouesneau, M., Deleitner, M., & Andrae, R. 2021, AJ, 161, 147
  6. [6]Baraffe, I., Chabrier, G., & Gallardo, J. 2009, ApJ, 702, L27
  7. [7]Bayo, A., Joergens, V., Liu, Y., et al. 2019, in Proceedings of Science, Vol. 331, Sissa Medialab
  8. [8]Benisty, M., Bae, J., Facchini, S., et al. 2021, ApJ, 916, L2
  9. [9]Benisty, M., Facchini, S., Teague, R., et al. 2026, ApJ, 1009, L32
  10. [10]Birnstiel, T. 2024, ARA&A, 62, 157
  11. [11]Bowler, B. P. 2016, PASP, 128, 102001
  12. [12]Bowler, B. P., Andrews, S. M., Kraus, A. L., et al. 2015, ApJ, 805, L17
  13. [13]Bowler, B. P., Tran, Q. H., Zhang, Z., et al. 2023, AJ, 165, 164
  14. [15]Canup, R. M. & Ward, W. R. 2002, AJ, 124, 3404
  15. [16]Casassus, S., Cárcamo, M., Domínguez-Jamett, O., et al. 2026, A&A, 710, L10
  16. [17]Chiang, E. & Goldreich, P. 1997, ApJ, 490, 368
  17. [18]Choksi, N. & Chiang, E. 2025, MNRAS, 537, 2945
  18. [19]Close, L. M., Males, J. R., Li, J., et al. 2025a, AJ, 169, 35
  19. [20]Close, L. M., van Capelleveen, R. F., Weible, G., et al. 2025b, ApJ, 990, L9
  20. [21]Cugno, G. & Grant, S. L. 2025, ApJ, 991, L46
  21. [22]Cugno, G., Patapis, P., Banzatti, A., et al. 2024, ApJ, 966, L21
  22. [23]Currie, T., Lawson, K., Schneider, G., et al. 2022, Nat. Astron., 6, 751
  23. [24]D’Alessio, P., Cantó, J., Calvet, N., & Lizano, S. 1998, ApJ, 500, 411
  24. [25]Darcis, M., Haffert, S. Y., Stolker, T., et al. 2026, A&A, 710, A328
  25. [26]Demars, D., Bonnefoy, M., Dougados, C., et al. 2023, A&A, 676, A123
  26. [27]Domínguez-Jamett, O., Casassus, S., Baobab Liu, H., et al. 2025, A&A, 702, A18
  27. [28]Donati, J.-F., Gregory, S., Alencar, S., et al. 2012, MNRAS, 425, 2948
  28. [29]Dullemond, C., Dominik, C., & Natta, A. 2001, ApJ, 560, 957
  29. [30]Facchini, S., Curone, P., Benisty, M., et al. 2026, ApJ, 998, L16
  30. [31]Fasano, D., Benisty, M., Curone, P., et al. 2025, A&A, 699, A373
  31. [32]Finley, C. O., Bowler, B. P., Wu, Y.-L., et al. 2026, AJ, 172, 134
  32. [33]Gaia Collaboration et al. 2022, VizieR Online Data Catalog, 1355, I
  33. [34]Haffert, S., Bohn, A., De Boer, J., et al. 2019, Nat. Astron., 3, 749
  34. [35]Hartmann, L., Calvet, N., Gullbring, E., & D’Alessio, P. 1998, ApJ, 495, 385
  35. [36]Hildebrand, R. H. 1983, QJRAS, 24, 267
  36. [37]Isella, A., Benisty, M., Teague, R., et al. 2019, ApJ, 879, L25
  37. [38]Isella, A., Chandler, C. J., Carpenter, J. M., Pérez, L. M., & Ricci, L. 2014, ApJ, 788, 129
  38. [39]Jones, M. G., Pringle, J., & Alexander, R. 2012, MNRAS, 419, 925
  39. [40]Jorquera, S., Pérez, L. M., Chauvin, G., et al. 2021, AJ, 161, 146
  40. [41]Keith, S. L. & Wardle, M. 2014, MNRAS, 440, 89
  41. [42]Lachapelle, F.-R., Lafrenière, D., Gagné, J., et al. 2015, ApJ, 802, 61
  42. [43]Lawlor, C., van Capelleveen, R. F., Bourdarot, G., et al. 2026, ApJ, 1000, L38
  43. [44]Long, D. E., Zhang, K., Teague, R., & Bergin, E. A. 2020, ApJ, 895, L46
  44. [45]Loomis, R. A., Facchini, S., Benisty, M., et al. 2025, ApJ, 984, L7
  45. [46]MacGregor, M. A., Wilner, D. J., Czekala, I., et al. 2017, ApJ, 835, 17
  46. [47]Maeda, N., Otsuki, K., Sutsugui, R., et al. 2024, ApJ, 968, 62
  47. [48]Manara, C., Ansdell, M., Rosotti, G., et al. 2023, Protostars and Planets VII, 534, 539
  48. [49]Marois, C., Macintosh, B., & Barman, T. 2007, ApJ, 654, L151
  49. [50]Martin, R. G. & Lubow, S. H. 2011, MNRAS, 413, 1447
  50. [51]Martinez, R. A. & Kraus, A. L. 2022, AJ, 163, 36
  51. [52]Mesa, D., Keppler, M., Cantalloube, F., et al. 2019, A&A, 632, A25
  52. [53]Müller, A., Keppler, M., Henning, T., et al. 2018, A&A, 617, L2
  53. [54]Neuhäuser, R., Guenther, E., Wuchterl, G., et al. 2005, A&A, 435, L13
  54. [55]Picos, D. G., Snellen, I. A., de Regt, S., et al. 2025, A&A, 693, A298
  55. [56]Rab, C., Kamp, I., Ginski, C., et al. 2019, A&A, 624, A16
  56. [57]Ricci, L., Cazzoletti, P., Czekala, I., et al. 2017, AJ, 154, 24
  57. [58]Rosotti, G. P., Clarke, C. J., Manara, C. F., & Facchini, S. 2017, MNRAS, 468, 1631
  58. [59]Sanchis, E., Testi, L., Natta, A., et al. 2021, A&A, 649, A19
  59. [61]Seifahrt, A., Neuhäuser, R., & Hauschildt, P. H. 2007, A&A, 463, 309
  60. [62]Shabram, M. & Boley, A. C. 2013, ApJ, 767, 63
  61. [63]Stolker, T., Haffert, S. Y., Kesseli, A. Y., et al. 2021, AJ, 162, 286
  62. [64]Szulágyi, J., Masset, F., Lega, E., et al. 2016, MNRAS, 460, 2853
  63. [65]Teague, R., Benisty, M., Facchini, S., et al. 2025, ApJ, 984, L6
  64. [67]Trevascus, D., Blunt, S., Christiaens, V., et al. 2025, A&A, 698, A19
  65. [68]van Capelleveen, R. F., Ginski, C., Kenworthy, M. A., et al. 2025, ApJ, 990, L8
  66. [69]van Holstein, R. G., Stolker, T., Jensen-Clem, R., et al. 2021, A&A, 647, A21
  67. [70]Venkatesan, V., Blunt, S., Wang, J., et al. 2025, ApJ, 993, 69
  68. [71]Wagner, K., Follette, K. B., Close, L. M., et al. 2018, ApJ, 863, L8
  69. [72]Wang, J., Vigan, A., Lacour, S., et al. 2021, AJ, 161, 148
  70. [73]Ward, W. R. & Canup, R. M. 2010, AJ, 140, 1168
  71. [80]Zhou, Y., Bowler, B. P., Sanghi, A., et al. 2025, ApJ, 980, L39
  72. [81]Zhou, Y., Herczeg, G. J., Kraus, A. L., Metchev, S., & Cruz, K. L. 2014, ApJ, 783, L17

Paper details

vlnc:W_jxkmk73z
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents