exoALMA Program: Disk Kinematics & Planets

• exoALMA Program is a comprehensive ALMA Large Program that maps the gas and dust structure in protoplanetary disks to detect embedded planets via kinematic and thermodynamic signatures.
• It employs high-resolution Band 7 observations to resolve substructures like rings, gaps, and spirals, optimizing sensitivity and velocity resolution across 15 selected disks.
• Advanced imaging techniques and forward modeling validate non-Keplerian features and turbulence, setting a new standard for studying planet-disk interactions.

The exoALMA Program is an ALMA Large Program centered on the gas and dust structure of planet-forming disks, with the explicit aim of detecting and characterizing embedded planets through the kinematic and thermodynamic perturbations they imprint on molecular-line emission. It targets 15 nearby, bright, and spatially extended protoplanetary disks, using Band 7 observations of $^{12}$CO $J=3\!-\!2$, $^{13}$CO $J=3\!-\!2$, CS $J=7\!-\!6$, and $0.9$ mm continuum to recover disk geometry, emitting surfaces, temperatures, turbulence, pressure structure, and localized departures from Keplerian rotation at high spatial and spectral fidelity (Teague et al., 25 Apr 2025).

1. Survey design and scientific scope

exoALMA was designed around a coherent science case: to map the distributions and motions of gas and dust in a sample of large disks and relate substructures such as rings, gaps, spirals, azimuthal asymmetries, vortices, and kinematic anomalies to embedded planets, $(\mathrm{magneto\!-\!})$hydrodynamic instabilities, and disk winds. The survey emphasized Band 7 because the native correlator setup provides fine velocity sampling, and because Band 7 offers favorable brightness-temperature sensitivity for deep line cubes across a multi-source sample. The program images line cubes at channel spacings of $100\!-\!200~{\rm m\,s^{-1}}$, while the native spectral resolution reaches $\approx 27~{\rm m\,s^{-1}}$ at line center; continuum imaging reaches $\sim 90$ mas with $J=3\!-\!2$0 sensitivity, and fiducial molecular-line imaging reaches $J=3\!-\!2$1 with $J=3\!-\!2$2 K per $J=3\!-\!2$3 channel (Teague et al., 25 Apr 2025).

The sample was deliberately biased toward bright, radially extended disks with gas radii of roughly $J=3\!-\!2$4, inclinations between $J=3\!-\!2$5 and $J=3\!-\!2$6, and minimal contamination from envelopes or cloud emission. The targets are DM Tau, AA Tau, LkCa 15, HD 34282, MWC 758, CQ Tau, SY Cha, PDS 66/MP Mus, HD 135344B/SAO 206462, HD 143006, RXJ1604.3-2130 A/J1604, RXJ1615.3-3255/J1615, V4046 Sgr, RXJ1842.9-3532/J1842, and RXJ1852.3-3700/J1852. This is not a population-representative sample; it is a deliberately biased one optimized for high-fidelity kinematic work, and that bias is fundamental to interpreting program-level trends (Teague et al., 25 Apr 2025).

A central result of the survey design is that, at the achieved sensitivity, all disks appear dynamically structured when observed in sufficient detail. Extensive dust-continuum substructure is found across the sample, and the line data recover gas kinematics with uncertainties of order $J=3\!-\!2$7, revealing both localized perturbations and global departures from purely Keplerian rotation (Teague et al., 25 Apr 2025).

2. Calibration, imaging, and released data products

exoALMA required a calibration and imaging strategy tailored to Band 7, multi-configuration datasets, and spatially extended CO emission. The program combined short- and long-baseline 12-m observations, and for seven of the 15 sources it also incorporated ACA data to recover the largest angular scales. Because the survey explicitly targeted subtle non-Keplerian features, calibration emphasized uv-plane alignment, correction of phase decoherence through iterative self-calibration, and separation of spatial alignment from flux rescaling so that spurious asymmetries would not be imprinted on the line cubes (Loomis et al., 28 Apr 2025).

The collaboration delivered three principal imaging families: a fiducial set with circularized $J=3\!-\!2$8 beams, a high-resolution set down to $J=3\!-\!2$9 mas, and a low-resolution $^{13}$0 set optimized for surface-brightness sensitivity. Continuum images, spectral cubes, moment products, residual velocity maps, deprojected rotation curves, emission surfaces, empirical temperature structures, and radial profiles are publicly available through the ALMA archive and the exoALMA portal, together with calibrated visibilities and metadata on beams, channel spacing, and geometry (Teague et al., 25 Apr 2025).

A notable methodological feature of the program is the use of independent imaging paradigms to validate marginal structures. In seven disks with apparent non-Keplerian features in $^{13}$1CO $^{13}$2, regularized maximum-likelihood imaging with MPoL independently reproduced the same channel-map anomalies seen in CLEAN. That cross-check was framed not as a cosmetic comparison, but as a test of whether localized kinks, arcs, and other low-contrast structures survive fundamentally different image reconstruction procedures (Zawadzki et al., 27 Apr 2025).

3. Geometric reconstruction, tomography, and forward modeling

The program’s analysis backbone is built around Discminer, disksurf, bettermoments, and related forward-modeling tools. Discminer fits channel maps with parameterized geometry, Keplerian rotation, line profiles, and emitting surfaces, including two-component fitting for inclined disks to separate front- and back-side emission. disksurf provides non-parametric surface extraction from channel-map asymmetries, and bettermoments is used for peak-intensity and centroid products. In this framework, a standard projected velocity field is written as

$^{13}$3

while brightness temperatures follow

$^{13}$4

These relations are used operationally to derive emitting layers, temperatures, and residual velocity fields from the cubes (Izquierdo et al., 28 Apr 2025).

The same infrastructure supports explicit line-profile tomography. For optically thick lines such as $^{13}$5CO, exoALMA III models front and back surfaces with a “mask” combination, whereas optically thinner tracers such as CS can be combined additively to represent overlapping emission from both sides of the disk. This was developed to mitigate backside contamination in inclined disks and to extract more reliable centroids, linewidths, and geometric parameters than single-component moment methods (Izquierdo et al., 28 Apr 2025).

Forward modeling occupies a distinct place in the program. exoALMA VII benchmarked the hydrodynamics and radiative-transfer backbone by comparing FARGO3D, Idefix, Athena++, PLUTO, and Phantom for planet–disk interactions, followed by MCFOST and RADMC-3D for synthetic $^{13}$6CO cubes. Across the grid-based hydrodynamics codes, density and velocity perturbations were strongly consistent, with velocity perturbation scatter $^{13}$7. MCFOST and RADMC-3D agreed in disk temperature to $^{13}$8 across the domain, yielding brightness-temperature differences within $^{13}$9 K in synthetic $J=3\!-\!2$0CO channel maps. Planet localization with DISCMINER then showed only a few percent scatter in recovered radius and azimuth across all hydro–RT combinations, establishing a reliability floor for the program’s embedded-planet forward-modeling strategy (Bae et al., 25 Apr 2025).

4. Vertical structure, thermodynamics, and disk masses

One of exoALMA’s defining contributions is the empirical reconstruction of gaseous emission layers and two-dimensional temperature structure. Across the full sample, $J=3\!-\!2$1CO $J=3\!-\!2$2 traces the upper disk atmosphere with mean $J=3\!-\!2$3, while $J=3\!-\!2$4CO and CS probe lower layers with mean $J=3\!-\!2$5 and $J=3\!-\!2$6, respectively. The program derives 2D $J=3\!-\!2$7 temperature distributions and reports that localized substructure in emission surfaces and peak-intensity profiles is present in nearly every disk, often co-incident across tracers and with kinematic perturbations. Four disks display evidence of potential photo-desorption in the outer disk, and the mean $J=3\!-\!2$8CO emission height correlates positively with disk mass (Galloway-Sprietsma et al., 28 Apr 2025).

Disk masses are then attacked with multiple, partially independent methods. Rotation-curve modeling of $J=3\!-\!2$9CO and $J=7\!-\!6$0CO in ten disks yields dynamical disk masses, density scale radii, and effective $J=7\!-\!6$1 values. These fits include stellar gravity, disk self-gravity, pressure support, and vertical thermal stratification, and they find all modeled disks gravitationally stable. The inferred $J=7\!-\!6$2 spans $J=7\!-\!6$3, and the mean gas-to-dust ratio is approximately $J=7\!-\!6$4, not statistically consistent with the canonical value of $J=7\!-\!6$5 under optically thin dust assumptions (Longarini et al., 25 Apr 2025).

A complementary inversion uses CO emitting heights and temperature structure to infer $J=7\!-\!6$6, $J=7\!-\!6$7, and total disk masses from optically thick lines. Applied to 14 disks, this method finds that masses derived from emitting heights are systematically lower than dynamical masses if one assumes an ISM CO abundance, implying a median CO depletion factor of $J=7\!-\!6$8 (Rosotti et al., 28 Apr 2025). A further chemical benchmark compares CO+$J=7\!-\!6$9-based gas masses with kinematic masses in 19 disks and finds agreement typically within a factor of $0.9$0, with CO+$0.9$1 masses on average lower by a factor $0.9$2 (Trapman et al., 27 Apr 2025).

exoALMA XIV adds a distinct density diagnostic by detecting pressure-broadened $0.9$3CO $0.9$4 wings in RX J1604.3-2130 A. In the $0.9$5 au region that contains the mm dust ring, the pressure-broadening analysis yields a peak gas surface density of $0.9$6, a gas mass of $0.9$7, and a turbulence estimate of $0.9$8. The midplane pressure peak matches the dust ring location, directly supporting radial dust trapping at a gas-pressure maximum (Yoshida et al., 28 Apr 2025).

5. Non-Keplerian structure, turbulence, and embedded planets

Kinematic complexity is ubiquitous in exoALMA. The full $0.9$9CO $(\mathrm{magneto\!-\!})$0 channel maps reveal substructures in 13 of the 15 disks, including large-scale arcs or spiral arms, localized velocity kinks, and multiple faint filaments on the disk surface. In six systems—AA Tau, SY Cha, J1842, J1615, LkCa 15, and HD 143006—the channel morphology is consistent with planet wakes, with inferred orbital radii between $(\mathrm{magneto\!-\!})$1 and $(\mathrm{magneto\!-\!})$2 au and planet masses between $(\mathrm{magneto\!-\!})$3 and $(\mathrm{magneto\!-\!})$4 from comparison to hydrodynamic and radiative-transfer models (Pinte et al., 25 Apr 2025).

Program-wide residual analysis reinforces that picture while also broadening it. The 2D atlas of deviations from smooth Keplerian disks shows that all 15 targets exhibit large-scale residual structure in centroid velocity, line width, and peak intensity. Five disks—CQ Tau, MWC 758, HD 135344B, HD 34282, and SY Cha—show non-axisymmetric spiral-like residuals, preferentially in Herbig Ae/Be systems, whereas some sources such as J1852, PDS 66, and V4046 Sgr are dynamically quieter despite showing global residuals (Fukagawa et al., 13 Mar 2026). Tomographic analysis of full line profiles extends the search beyond centroids alone: localized line broadening, Doppler flips, centroid-uncertainty enhancements, and positive skewness in optically thick CO are used to separate planet-driven signatures from instability-driven ones. Applied to HD 135344B, this framework identifies candidate planets at $(\mathrm{magneto\!-\!})$5, $(\mathrm{magneto\!-\!})$6, and $(\mathrm{magneto\!-\!})$7 au; applied to MWC 758, it instead favors vertical-velocity spirals associated with eccentricity or warps rather than a localized outer planet (Izquierdo et al., 13 Mar 2026).

exoALMA also approaches turbulence in two complementary ways. In DM Tau, a full 3D MCFOST Bayesian fit to high-resolution $(\mathrm{magneto\!-\!})$8CO $(\mathrm{magneto\!-\!})$9 and CS $100\!-\!200~{\rm m\,s^{-1}}$0 data explored more than $100\!-\!200~{\rm m\,s^{-1}}$1 models and recovered a non-thermal broadening parameter $100\!-\!200~{\rm m\,s^{-1}}$2, corresponding to $100\!-\!200~{\rm m\,s^{-1}}$3, with CS independently supporting the CO-based inference (Hardiman et al., 2 Feb 2026). At the same time, a program-wide vertical-flow analysis finds that vertical motions are detected in most disks, typically at amplitudes of a few tens of $100\!-\!200~{\rm m\,s^{-1}}$4, but reaching $100\!-\!200~{\rm m\,s^{-1}}$5 ($100\!-\!200~{\rm m\,s^{-1}}$6) in spiral features of MWC 758 and CQ Tau and up to $100\!-\!200~{\rm m\,s^{-1}}$7 ($100\!-\!200~{\rm m\,s^{-1}}$8) in the outer disk of MWC 758 (Benisty et al., 13 Mar 2026).

Not all observed dust asymmetries map cleanly to vortices or planets. In four disks with strong crescent-like continuum emission—HD 135344B, HD 143006, HD 34282, and MWC 758—exoALMA XVII compared analytical and hydrodynamical vortex models to the $100\!-\!200~{\rm m\,s^{-1}}$9CO and $\approx 27~{\rm m\,s^{-1}}$0CO data and found no distinctive, co-spatial vortex kinematic signature at the dust crescents (Wölfer et al., 28 Apr 2025). Conversely, exoALMA XVIII argues that moderate warps of order $\approx 27~{\rm m\,s^{-1}}$1 can reproduce many widespread large-scale $\approx 27~{\rm m\,s^{-1}}$2 residuals, including the spiral-like kinematics of MWC 758, and can also generate $\approx 27~{\rm m\,s^{-1}}$3 K variations in CO brightness temperature (Winter et al., 15 Jul 2025). Synthetic observations of VSI, MRI, and GI similarly show that exoALMA-quality data should resolve large-scale instability-driven residuals; comparison to the survey suggests a scarcity of ring- and arc-like VSI signatures, while MRI- or GI-like spirals remain plausible in several disks (Barraza-Alfaro et al., 28 Apr 2025).

6. Programmatic significance, caveats, and trajectory

Taken together, the exoALMA papers define a program in which disk geometry, vertical structure, chemistry, turbulence, and embedded-planet signatures are treated as coupled observables rather than isolated diagnostics. The combination of uniformly calibrated Band 7 data, public visibilities and cubes, explicit imaging cross-checks, and hydro–RT benchmarking means that continuum substructures, emission surfaces, rotation curves, linewidths, and channel-map perturbations can be compared within a common framework rather than across heterogeneous archives (Teague et al., 25 Apr 2025).

Several limitations are intrinsic and repeatedly emphasized. The survey sample is biased toward bright, extended disks and should not be read as a demographic census of all protoplanetary disks. Optical depth, emission-surface geometry, backside contamination, beam smearing, channelization, continuum subtraction, and correlated image noise all remain important systematics. Many analyses adopt simplified chemistry, vertically averaged or locally isothermal disks, constant turbulence parameters, or single-planet interpretations. Warps, winds, instability-driven flows, and planet wakes can project into superficially similar residual morphologies, and the papers consistently treat those degeneracies as active modeling problems rather than solved classifications (Bae et al., 25 Apr 2025).

The program’s trajectory is correspondingly multi-pronged. Published outlook sections point to expanded multi-line chemistry using $\approx 27~{\rm m\,s^{-1}}$4CO, $\approx 27~{\rm m\,s^{-1}}$5CO, and C$\approx 27~{\rm m\,s^{-1}}$6O; broader use of heavier species such as CS for non-thermal broadening; systematic grids in planet mass, orbital radius, inclination, and flaring; non-LTE excitation; dust evolution; and increasingly explicit 3D forward modeling of vertical flows, warps, and line-profile asymmetries (Bae et al., 25 Apr 2025). A plausible implication is that exoALMA’s lasting importance lies not in any single detection claim, but in having established a program-scale standard for combining deep molecular-line imaging, continuum morphology, geometry-aware tomography, and benchmarked forward models into a single, internally cross-validated description of the protoplanetary disks in which giant planets are expected to form.