To determine the origin of the spiral structure observed in the dust continuum emission of Elias 2–27 we analyze multiwavelength continuum ALMA data with a resolution of ∼0.″2 (∼23 au) at 0.89, 1.3, and 3.3 mm. We also study the kinematics of the disk with 13CO and C18O ALMA observations in the J = 3–2 transition. The spiral arm morphology is recovered at all wavelengths in the dust continuum observations, where we measure contrast and spectral index variations along the spiral arms and detect subtle dust-trapping signatures. We determine that the emission from the midplane is cold and interpret the optical depth results as signatures of a disk mass higher than previous constraints. From the gas data, we search for deviations from Keplerian motion and trace the morphology of the emitting surfaces and the velocity profiles. We find an azimuthally varying emission layer height in the system, large-scale emission surrounding the disk, and strong perturbations in the channel maps, colocated with the spirals. Additionally, we develop multigrain dust and gas hydrodynamical simulations of a gravitationally unstable disk and compare them to the observations. Given the large-scale emission and highly perturbed gas structure, together with the comparison of continuum observations to theoretical predictions, we propose infall-triggered gravitational instabilities as the origin for the observed spiral structure.

Recent multi-wavelength Atacama Large Millimeter/submillimeter Array (ALMA) observations of the protoplanetary disk orbiting around Elias 2–27 revealed a two-armed spiral structure. The observed morphology, together with the young age of the star and the disk-to-star mass ratio estimated from dust-continuum emission, make this system a perfect laboratory to investigate the role of self-gravity in the early phases of star formation. This is particularly interesting if we consider that gravitational instabilities could be a fundamental first step for the formation of planetesimals and planets. In this Letter, we model the rotation curve obtained by CO data of Elias 2–27 with a theoretical rotation curve, including both the disk self-gravity and the star contribution to the gravitational potential. We compare this model with a purely Keplerian one and with a simple power-law function. We find that (especially for the 13CO isotopologue) the rotation curve is better described by considering not only the star, but also the disk self-gravity. We are thus able to obtain for the first time a dynamical estimate of the disk mass of 0.08 ± 0.04 M ⊙ and the star mass of 0.46 ± 0.03 M ⊙ (in the more general case), the latter being comparable with previous estimates. From these values, we derive that the disk is 17% of the star mass, meaning that it could be prone to gravitational instabilities. This result would strongly support the hypothesis that the two spiral arms are generated by gravitational instabilities.

Observations with the Atacama Large Millimeter/Submillimeter Array (ALMA) have dramatically improved our understanding of the site of exoplanet formation: protoplanetary disks. However, many basic properties of these disks are not well understood. The most fundamental of these is the total disk mass, which sets the mass budget for planet formation. Disks with sufficiently high masses can excite gravitational instability and drive spiral arms that are detectable with ALMA. Although spirals have been detected in ALMA observations of the dust, their association with gravitational instability, and high disk masses, is far from clear. Here we report a prediction for kinematic evidence of gravitational instability. Using hydrodynamics simulations coupled with radiative transfer calculations, we show that a disk undergoing such instability has clear kinematic signatures in molecular line observations across the entire disk azimuth and radius, which are independent of viewing angle. If these signatures are detected, it will provide the clearest evidence for the occurrence of gravitational instability in planet-forming disks, and provide a crucial way to measure disk masses.

Context. Most disks observed at high angular resolution show signs of substructures, such as rings, gaps, arcs, and cavities, in both the gas and the dust. To understand the physical mechanisms responsible for these structures, knowledge about the gas surface density is essential. This, in turn, requires information on the gas temperature. Aims. The aim of this work is to constrain the gas temperature as well as the gas surface densities inside and outside the millimeter-dust cavities of two transition disks: LkCa15 and HD 169142, which have dust cavities of 68 AU and 25 AU, respectively. Methods. We use some of the few existing ALMA observations of the J = 6-5 transition of 13CO together with archival J = 2−1 data of 12CO, 13CO, and C18O. The ratio of the 13CO J = 6−5 to the J = 2−1 transition is used to constrain the temperature and is compared with that found from peak brightness temperatures of optically thick lines. The spectra are used to resolve the innermost disk regions to a spatial resolution better than that of the beam of the observations. Furthermore, we use the thermochemical code DALI to model the temperature and density structure of a typical transition disk as well as the emitting regions of the CO isotopologs. Results. The 13CO J = 6−5 and J = 2−1 transitions peak inside the dust cavity in both disks, indicating that gas is present in the dust cavities. The kinematically derived radial profiles show that the gas is detected down to 10 and 5-10 AU, much farther in than the dust cavities in the LkCa15 and HD 169142 disks, respectively. For LkCa15, the steep increase toward the star in the 13CO J = 6−5 transition, in contrast to the J = 2−1 line, shows that the gas is too warm to be traced by the J = 2−1 line and that molecular excitation is important for analyzing the line emission. Quantitatively, the 6−5/2−1 line ratio constrains the gas temperature in the emitting layers inside the dust cavity to be up to 65 K, warmer than in the outer disk, which is at 20-30 K. For HD 169142, the lines are optically thick, complicating a line ratio analysis. In this case, the peak brightness temperature constrains the gas in the dust cavity of HD 169142 to be 170 K, whereas that in the outer disk is only 100 K. The data indicate a vertical structure in which the 13CO 6-5 line emits from a higher layer than the 2-1 line in both disks, consistent with exploratory thermochemical DALI models. Such models also show that a more luminous central star, a lower abundance of polycyclic aromatic hydrocarbons, and the absence of a dusty inner disk increase the temperature of the emitting layers and hence the line ratio in the gas cavity. The gas column density in the LkCa15 dust cavity drops by a factor of >2 compared to the outer disk, with an additional drop of an order of magnitude inside the gas cavity at 10 AU. In the case of HD 169142, the gas column density drops by a factor of 200–500 inside the gas cavity. Conclusions. The gas temperatures inside the dust cavities steeply increase toward the star and reach temperatures of up to 65 K (LkCa15) and 170 K (HD 169142) on scales of ~15–30 AU, whereas the temperature gradients of the emitting layers in the outer disks are shallow, with typical temperatures of 20-30 and 100 K, respectively. The deep drop in gas column density inside the HD 169142 gas cavity at <10 AU could be due to a massive companion of several MJ, whereas the broad dust-depleted gas region from 10 to 68 AU for LkCa15 may imply several lower mass planets. This work demonstrates that knowledge of the gas temperature is important for determining the gas surface density and thus whether planets, and if so what kinds of planets, are most likely to be carving the dust cavities.