Context. Type-II migration of giant planets has a speed proportional to the disc’s viscosity for values of the α viscosity parameter larger than 10−4. Previous studies based on two-dimensional simulations, have shown that, at even lower viscosities, migration can be very chaotic and is often characterised by phases of fast migration. The reason is that vortices appear in low-viscosity discs due to the Rossby-wave instability at the edges of the gap opened by the planet. Migration is then determined by vortex-planet interactions. Aims. Our goal is to study giant planet migration in low-viscosity discs with 3D simulations. In 3D, vortices are more complex than the simple vertical extension of their 2D counterparts; their impact on planet migration is therefore not obvious. Methods. We performed numerical simulations using two grid-based codes: FARGOCA for three-dimensional simulations and FARGO-ADSG for the two dimensional case. Two-dimensional simulations were used mainly for preliminary tests to check the impact of self-gravity on vortex formation and on vortex-disc dynamics. After selecting disc masses for which self-gravity is not important at the planet location, three-dimensional simulations without self-gravity can be safely used. We have considered an adiabatic equation of state with exponential damping of temperature perturbations in order to avoid the development of the vertical shear instability. In our nominal simulation, we set α = 0 so that only numerical viscosity is present. We then performed simulations with non-zero α values to assess the threshold of prescribed viscosity below which the new migration processes appear. Results. We show that for α ≲ 10−5 two migration modes are possible, which differ from classical Type-II migration in the sense that they are not proportional to the disc’s viscosity. The first occurs when the gap opened by the planet is not very deep. This occurs in 3D simulations and/or when a big vortex forms at the outer edge of the planetary gap, diffusing material into the gap. The de-saturation of co-orbital and co-rotation resonances keeps the planet’s eccentricity low. Inward planet migration then occurs as long as the disc can refill the gap left behind by the migrating planet, either due to diffusion caused by the presence of the vortex or to the inward migration of the vortex itself due to its interaction with the disc. We call this type of migration ‘vortex-driven migration’, which differs from ‘vortex-induced’ migration described in Lin & Papaloizou (2010, MNRAS, 405, 1473, and 2011a, MNRAS, 415, 1445). This migration is very slow and cannot continue indefinitely because eventually the vortex dissolves. The second migration mode occurs when the gap is deep so that the planet’s eccentricity grows to a value e ~ 0.2 due to inefficient eccentricity damping by co-rotation resonances. Once the planet is on an eccentric orbit, gas can pass through the gap and planet migration unlocks from the disc’s viscous evolution. This second, faster migration mode appears to be typical of two-dimensional models in discs with slower damping of temperature perturbations. Conclusions. Vortex-driven migration in low-viscosity discs can be very slow and eventually reverses and stops, offering an interesting mechanism to explain the existence of the cold-Jupiter population, even if these planets originally started growing at the disc’s snowline.

Context. Planets in accretion disks can excite spiral shocks, and-if massive enough-open gaps in their vicinity. Both of these effects can influence the overall disk thermal structure. Aims. We model planets of different masses and semimajor axes in disks of various viscosities and accretion rates to examine their impact on disk thermodynamics and highlight the mutable, non-axisymmetric nature of icelines in systems with massive planets. Methods. We conduct a parameter study using numerical hydrodynamics simulations where we treat viscous heating, thermal cooling and stellar irradiation as additional source terms in the energy equation, with some runs including radiative diffusion. Our parameter space consists of a grid containing different combinations of planet and disk parameters.Results. Both gap opening and shock heating can displace the iceline, with the effects being amplified for massive planets in optically thick disks. The gap region can split an initially hot (T > 170 K) disk into a hot inner disk and a hot ring just outside of the planet's location, while shock heating can reshape the originally axisymmetric iceline into water-poor islands along spirals. We also find that radiative diffusion does not alter the picture significantly in this context. Conclusions. Shock heating and gap opening by a planet can effectively heat up optically thick disks and in general move and/or reshape the water iceline. This can affect the gap structure and migration torques. It can also produce azimuthal features that follow the trajectory of spiral arms, creating hot zones, "islands" of vapor and ice around spirals which could affect the accretion or growth of icy aggregates.

Recent ALMA observations revealed concentric annular structures in several young class-II objects. In an attempt to produce the rings and gaps in some of these systems, they have been modeled numerically with a single embedded planet assuming a locally isothermal equation of state. This is often justified by observations targeting the irradiation-dominated outer regions of disks (approximately 100 au). We test this assumption by conducting hydrodynamics simulations of embedded planets in thin locally isothermal and radiative disks that mimic the systems HD 163296 and AS 209 in order to examine the effect of including the energy equation in a seemingly locally isothermal environment as far as planet-disk interaction is concerned. We find that modeling such disks with an ideal equation of state makes a difference in terms of the number of produced rings and the spiral arm contrast in the disk. Locally isothermal disks produce sharper annular or azimuthal features and overestimate a single planet's gap-opening capabilities by producing multiple gaps. In contrast, planets in radiative disks carve a single gap for typical disk parameters. Consequently, for accurate modeling of planets with semimajor axes up to about 100 au, radiative effects should be taken into account even in seemingly locally isothermal disks. In addition, for the case of AS 209, we find that the primary gap is significantly different between locally isothermal and radiative models. Our results suggest that multiple planets are required to explain the ring-rich structures in such systems.

Context. Several observations of protoplanetary disks reveal non-axisymmetric features, which are often interpreted as vortices. Numerical modeling has repeatedly shown that gap-opening planets are capable of producing large and long-lasting vortices at their outer gap edge, making massive planets popular candidates as the source of such features. Aims. We explore the lifetime of vortices generated by Jupiter-sized planets as a function of the thermal relaxation timescale, the level of turbulence, and the effect of disk self-gravity. Methods. We conduct 2D numerical simulations using the hydrodynamics codes PLUTO and FARGO, scanning through several physical and numerical parameters. Vortex properties are automatically extracted from thousands of simulation snapshots. Results. We find that vortices that spawn at the outer gap edge can survive for about 100–3000 planetary orbits, with the shortest lifetimes occurring for moderately efficient dissipation and cooling. However, we also observe a different regime of long-lasting vortices with lifetimes of at least 15 000 orbits for very low viscosity and very short thermal relaxation timescales. Disk self-gravity significantly shortens the lifetime of regular vortices but still allows long-lived ones to survive. Conclusions. Our results suggest that the cooling timescale plays an important role in vortex formation and lifetime and that planet-generated vortices should be observable at large distances from the star for typical thermal relaxation timescales and low turbulence levels.

Context: Recent observations with the Atacama Large Millimeter Array (ALMA) have shown that the large dust aggregates observed at millimeter wavelengths settle to the midplane into a remarkably thin layer. This sets strong limits on the strength of the turbulence and other gas motions in these disks. Aims: We intend to find out if the geometric thinness of these layers is evidence against the vertical shear instability (VSI) operating in these disks. We aim to verify if a dust layer consisting of large enough dust aggregates could remain geometrically thin enough to be consistent with the latest observations of these dust layers, even if the disk is unstable to the VSI. If this is falsified, then the observed flatness of these dust layers proves that these disks are stable against the VSI, even out to the large radii at which these dust layers are observed. Methods: We performed hydrodynamic simulations of a protoplanetary disk with a locally isothermal equation of state, and let the VSI fully develop. We sprinkled dust particles with a given grain size at random positions near the midplane and followed their motion as they got stirred up by the VSI, assuming no feedback onto the gas. We repeated the experiment for different grain sizes and determined for which grain size the layer becomes thin enough to be consistent with ALMA observations. We then verified if, with these grain sizes, it is still possible (given the constraints of dust opacity and gravitational stability) to generate a moderately optically thick layer at millimeter wavelengths, as observations appear to indicate. Results: We found that even very large dust aggregates with Stokes numbers close to unity get stirred up to relatively large heights above the midplane by the VSI, which is in conflict with the observed geometric thinness. For grains so large that the Stokes number exceeds unity, the layer can be made to remain thin, but we show that it is hard to make dust layers optically thick at ALMA wavelengths (e.g., τ1.3mm 1) with such large dust aggregates. Conclusions: We conclude that protoplanetary disks with geometrically thin midplane dust layers cannot be VSI unstable, at least not down to the disk midplane. Explanations for the inhibition of the VSI out to several hundreds of au include a high dust-to-gas ratio of the midplane layer, a modest background turbulence, and/or a reduced dust-to-gas ratio of the small dust grains that are responsible for the radiative cooling of the disk. A reduction of small grains by a factor of between 10 and 100 is sufficient to quench the VSI. Such a reduction is plausible in dust growth models, and still consistent with observations at optical and infrared wavelengths.