Aims. Migration of dense gaseous clumps that form in young protostellar disks via gravitational fragmentation is investigated to determine the likelihood of giant planet formation. Methods. High-resolution numerical hydrodynamics simulations in the thin-disk limit are employed to compute the formation and long-term evolution of a gravitationally unstable protostellar disk around a solar-mass star.Results. We show that gaseous clumps that form in the outer regions of the disk (¿100 AU) through disk fragmentation are often perturbed by other clumps or disk structures, such as spiral arms, and migrate toward the central star on timescales from a few 10 3 to few 10 4 yr. The migration timescale is slowest when stellar motion in response to the disk gravity is considered. When approaching the star, the clumps first gain mass (up to several tens of MJup), but then quickly lose most of their diffuse envelopes through tidal torques. Part of the clump envelope can be accreted on the central star causing an FU-Ori-type accretion and luminosity outburst. The tidal mass loss helps the clumps to significantly slow down or even halt their inward migration at a distance of a few tens of AU from the protostar. The resulting clumps are heavily truncated both in mass and size compared to their wider-orbit counterparts, keeping only a dense and hot nucleus. During the inward migration, the temperature in the clump interiors may exceed the molecular hydrogen dissociation limit (2000 K) and the central region of the clump can collapse into a gas giant protoplanet. Moreover, migrating clumps may experience close encounters with other clumps, resulting in the ejection of the least massive (planetary-mass) clumps from the disk. We argue that FU-Orionis-type luminosity outbursts may be the end product of disk fragmentation and clump inward migration, ushering the formation of giant protoplanets on tens-of-au orbits in systems such as HR 8799.

We present self-consistent calculations coupling numerical hydrodynamics simulations of collapsing pre-stellar cores and stellar evolution models of accreting objects. We analyse the main impact of consistent accretion history on the evolution and lithium depletion of young low-mass stars and brown dwarfs. These consistent models confirm the generation of a luminosity spread in Herzsprung-Russell diagrams at ages ∼ 1-10 Myr. They also confirm that early accretion can produce objects with abnormal Li depletion, as found in a previous study that was based on arbitrary accretion rates. The results strengthen that objects with anomalously high level of Li depletion in young clusters should be extremely rare. We also find that early phases of burst accretion can produce coeval models of similar mass with a range of different Li surface abundances, and in particular with Li-excess compared to the predictions of non-accreting counterparts. This result is due to a subtle competition between the effect of burst accretion and its impact on the central stellar temperature, the growth of the stellar radiative core and the accretion of fresh Li from the accretion disk. Only consistent models could reveal such a subtle combination of effects. This new result could explain the recent, puzzling observations of Li-excess of fast rotators in the young cluster NGC 2264. Present self-consistent accreting models are available in electronic form.

Aims. The pre-main-sequence evolution of low-mass stars and brown dwarfs is studied numerically starting from the formation of a protostellar/proto-brown dwarf seed and taking into account the mass accretion onto the central object during the initial several Myr of evolution. Methods. The stellar evolution was computed using the STELLAR evolution code developed by Yorke & Bodenheimer with recent modifications by Hosokawa et al. The mass accretion rates were taken from numerical hydrodynamics models of Vorobyov & Basu computing the circumstellar disk evolution starting from the gravitational collapse of pre-stellar cloud cores of various mass and angular momentum. The resulting stellar evolution tracks were compared with the isochrones and isomasses calculated using non-accreting models.Results. We find that mass accretion in the initial several Myr of protostellar evolution can have a strong effect on the subsequent evolution of young stars and brown dwarfs. The disagreement between accreting and non-accreting models in terms of the total stellar luminosity L * , stellar radius R * and effective temperature T eff depends on the thermal efficiency of accretion, i.e., on the fraction of accretion energy absorbed by the central object. The largest mismatch is found for the cold accretion case, in which essentially all accretion energy is radiated away. The relative deviations in L * and R * in this case can reach 50% for 1.0-Myr-old objects and remain notable even for 10-Myr-old objects. In the hot and hybrid accretion cases, in which a constant fraction of accretion energy is absorbed, the disagreement between accreting and non-accreting models becomes less pronounced, but still remains notable for 1.0-Myr-old objects. These disagreements may lead to the wrong age estimate for objects of (sub-)solar mass when using the isochrones based on non-accreting models, as was also previously noted by Baraffe et al. and Hosokawa et al. We find that objects with strong luminosity bursts exhibit notable excursions in the L * -T eff diagram, but the character of these excursions is distinct for hybrid/hot and cold accretion scenarios. In particular, the cold accretion scenario predicts peak luminosities that are greater than those of known FU-Orionis-type outbursts, which implies that cold accretion is physically less realistic. Conclusions. Mass accretion during the early stages of star and brown dwarf evolution is an important factor, but its effect depends on the details of how accretion energy is distributed within the star.

Context. Episodic accretion is an important process in the evolution of young stars and their environment. The observed strong luminosity bursts of young stellar objects likely have a long lasting (i.e. longer than the burst duration) impact on the chemical evolution of the disk and envelope of young stars. Aims. We aim to investigate the observational signatures of the chemical evolution in the post-burst phase for embedded sources. With such signatures it is possible to identify targets that experienced a recent luminosity burst. Methods. We present a new model for the chemistry of episodic accretion based on the two dimensional, radiation thermo-chemical disk code PRODIMO (PROtoplanetary DIsk MOdel). We have extended PRODIMO with a proper treatment for envelope structures. For a representative Class I model, we calculated the chemical abundances in the post-burst phase and produced synthetic observables such as intensity maps and radial intensity profiles. Results. During a burst, many chemical species, such as CO, sublimate from the dust surfaces. As the burst ends they freeze out again (post-burst phase). This freeze-out happens from inside-out due to the radial density gradient in the disk and envelope structure. This inside-out freeze-out produces clear observational signatures in spectral line emission, such as rings and distinct features in the slope of radial intensity profiles. We fitted synthetic C 18 O J = 2−1 observations with single and two component fits and find that post-burst images are much better matched by the latter. Comparing the quality of such fits therefore allows identification of post-burst targets in a model-independent way. Conclusions. Our models confirm that it is possible to identify post-burst objects from spatially resolved CO observations. However, to derive proper statistics, such as the strength and frequencies of bursts, from observations it is important to consider the inclination and structure of the target and dust properties, as these have a significant impact on the freeze-out timescale.

Episodic accretion-driven outbursts are an extreme manifestation of accretion variability. It has been proposed that the development of gravitational instabilities in the proto-circumstellar medium of massive young stellar objects (MYSOs) can lead to such luminous bursts, when clumps of fragmented accretion discs migrate onto the star. We simulate the early evolution of MYSOs formed by the gravitational collapse of rotating 100 M pre-stellar cores and analyze the characteristics of the bursts that episodically accompany their strongly time-variable protostellar lightcurve. We predict that MYSOs spend ≈ 10 3 yr (≈ 1.7%) of their modelled early 60 kyr experiencing eruptive phases, during which the peak luminosity exceeds the quiescent pre-burst values by factors from 2.5 to more than 40. Throughout these short time periods, they can acquire a substential fraction (up to ≈ 50 %) of their zero-age-main sequence mass. Our findings show that fainter bursts are more common than brighter ones. We discuss our results in the context of the known bursting MYSOs, e.g. NGC6334I-MM1 and S255IR-NIRS3, and propose that these monitored bursts are part of a long-time ongoing series of eruptions, which might, in the future, be followed by other luminous flares.