arXiv subjects

Dimitris Stamatellos

Publications and source records attributed to Dimitris Stamatellos.

43 records · Page 3Linked to original sources

arXiv2008

The properties of brown dwarfs and low-mass hydrogen-burning stars formed by disc fragmentation

We suggest that a high proportion of brown dwarfs are formed by gravitational fragmentation of massive extended discs around Sun-like stars. Such discs should arise frequently, but should be observed infrequently, precisely because they fragment rapidly. By performing an ensemble of radiation-hydrodynamic simulations, we show that such discs fragment within a few thousand years, and produce mainlybrown dwarf (BDs) stars, but also planetary mass (PM) stars and very low-mass hydrogen-burning (HB) stars. Most of the the PM stars and BDs are ejected by mutual interactions. We analyse the statistical properties of these stars, and compare them with observations. After a few hundred thousand years the Sun-like primary is typically left with a close low-mass HB companion, and two much wider companions: a low-mass HB star and a BD star, or a BD-BD binary. There is a BD desert extending out to at least ~100 AU; this is because BDs tend to be formed further out than low-mass HB stars, and then they tend to be scattered even further out, or even into the field. BDs form with discs of a few Mj and radii of a few tens of AU, and they are more likely to retain these discs if they remain bound to the primary star. Binaries form by pairing of the newly-formed stars in the disc, giving a low-mass binary fraction of ~0.16. These binaries include close and wide BD/BD binaries and BD/PM binaries. BDs that remain as companions to Sun-like stars are more likely to be in BD/BD binaries than are BDs ejected into the field. Disc fragmentation is a robust mechanism; even if only a small fraction of Sun-like stars host the required massive extended discs,this mechanism can produce all the PM stars observed, most of the BD stars, and a significant proportion of the very low-mass HB stars.

astro-ph↗

arXiv2008

The formation of brown dwarfs and low-mass stars by disc fragmentation

We suggest that a high proportion of brown dwarfs are formed by gravitational fragmentation of massive, extended discs around Sun-like stars. We argue that such discs should arise frequently, but should be observed infrequently, precisely because they fragment rapidly. By performing an ensemble of radiation-hydrodynamic simulations, we show that such discs typically fragment within a few thousand years to produce mainly brown dwarfs (including planetary-mass brown dwarfs) and low-mass hydrogen-burning stars. Subsequently most of the brown dwarfs are ejected by mutual interactions. We analyse the properties of these objects that form by disc fragmentation, and compare them with observations.

astro-ph↗

arXiv2008

Introducing a Hybrid Method of Radiative Transfer in Smoothed Particle Hydrodynamics

We present a new method of incorporating radiative transfer into Smoothed Particle Hydrodynamics (SPH). There have been many recent attempts at radiative transfer in SPH (Stamatellos et al 2005, 2005, Mayer et al 2007, Whitehouse and Bate 2006), however these are becoming increasingly complex, with some methods requiring the photosphere to be mapped (which is often of non-trivial geometric shape), and extra conditions to be applied there (matching atmospheres as in Cai et al (2008), or specifying cooling at the photosphere as in Mayer et al (2007)). The method of identifying the photosphere is usually a significant addition to the total simulation runtime, and often requires extra free parameters, the changing of which will affect the final results. Our method is not affected by such concerns, as the photosphere is constructed implicitly by the algorithm without the need for extra free parameters. The algorithm used is a synergy of two current formalisms for radiative effects: a) the polytropic cooling formalism proposed by Stamatellos et al (2007), and b) flux-limited diffusion, used by many authors to simulate radiation transport in the optically thick regime (e.g. Mayer et al 2007). We present several tests of this method: (1) The evolution of a 0.07 solar mass protoplanetary disc around a 0.5 solarmass star (Pickett et al 2003, Mejia et al 2005, Boley et al 2006, Cai et al 2008); (2) The collapse of a non-rotating 1 solar mass molecular cloud (Masunaga & Inutsuka 2000, Stamatellos et al 2007); (3) The thermal relaxation of temperature fluctuations in an static homogeneous sphere (Masunaga et al 1998, Spiegel 1957, Stamatellos et al 2007)

astro-ph↗

arXiv2008

Can giant planets form by gravitational fragmentation of discs?

Gravitational fragmentation has been proposed as a mechanism for the formation of giant planets in close orbits around solar-type stars. However, it is debatable whether this mechanism can function in the inner regions (R<40 AU) of real discs. We use a newly developed method for treating the energy equation and the equation of state, which accounts for radiative transfer effects in SPH simulations of circumstellar discs. The different chemical and internal states of hydrogen and the properties of dust at different densities and temperatures (ice coated dust grains at low temperatures, ice melting, dust sublimation) are all taken into account by the new method.We present radiative hydrodynamic simulations of the inner regions of massive circumstellar discs and examine two cases: (i) a disc irradiated by a cool background radiation field (T_bgr=10K)and (ii) a disc heated by radiation from its central star (T_bgr~1/R). In neither case does the disc fragment: in the former because it cannot cool fast enough and in the latter because it is not gravitationally unstable. Our results (a) corroborate previous numerical results using different treatments for the hydrodynamics and the radiative transfer, and (b) confirm our own earlier analytic predictions. We conclude that disc fragmentation is unlikely to be able to produce giant planets around solar-type stars at radii <40 AU.

astro-ph↗

arXiv2007

Brown dwarf formation by gravitational fragmentation of massive, extended protostellar discs

We suggest that low-mass hydrogen-burning stars like the Sun should sometimes form with massive extended discs; and we show, by means of radiation hydrodynamic simulations, that the outer parts of such discs (R>100 AU) are likely to fragment on a dynamical timescale (10^3 to $10^4 yr), forming low-mass companions: principally brown dwarfs (BDs), but also very low-mass hydrogen-burning stars and planetary-mass objects. A few of the BDs formed in this way remain attached to the primary star, orbiting at large radii. The majority are released into the field, by interactions amongst themselves; in so doing they acquire only a low velocity dispersion (<2 km/s), and therefore they usually retain small discs, capable of registering an infrared excess and sustaining accretion. Some BDs form close BD/BD binaries, and these binaries can survive ejection into the field. This BD formation mechanism appears to avoid some of the problems associated with the `embryo ejection' scenario, and to answer some of the questions not yet answered by the `turbulent fragmentation' scenario.

astro-ph↗

arXiv2005

Monte Carlo radiative transfer in SPH density fields

We combine a Monte Carlo radiative transfer code with an SPH code, so that -- assuming thermal equilibrium -- we can calculate dust-temperature fields, spectral energy distributions, and isophotal maps, for the individual time-frames generated by an SPH simulation. On large scales, the radiative transfer cells (RT cells) are borrowed from the tree structure built by the SPH code, and are chosen so that their size -- and hence the resolution of the calculated temperature field -- is comparable with the resolution of the density field. We refer collectively to these cubic RT cells as the 'global grid'. The code is tested and found to treat externally illuminated dust configurations very well. However, when there are embedded discrete sources, i.e. stars, these produce very steep local temperature gradients which can only be modelled properly if -- in the immediate vicinity of, and centred on, each embedded star -- we supplement the global grid with a 'star grid' of closely spaced concentric RT cells.

astro-ph↗

arXiv2005

Radiative equilibrium in Monte Carlo radiative transfer using frequency distribution adjustment

The Monte Carlo method is a powerful tool for performing radiative equilibrium calculations, even in complex geometries. The main drawback of the standard Monte Carlo radiative equilibrium methods is that they require iteration, which makes them numerically very demanding. Bjorkman & Wood recently proposed a frequency distribution adjustment scheme, which allows radiative equilibrium Monte Carlo calculations to be performed without iteration, by choosing the frequency of each re-emitted photon such that it corrects for the incorrect spectrum of the previously re-emitted photons. Although the method appears to yield correct results, we argue that its theoretical basis is not completely transparent, and that it is not completely clear whether this technique is an exact rigorous method, or whether it is just a good and convenient approximation. We critically study the general problem of how an already sampled distribution can be adjusted to a new distribution by adding data points sampled from an adjustment distribution. We show that this adjustment is not always possible, and that it depends on the shape of the original and desired distributions, as well as on the relative number of data points that can be added. Applying this theorem to radiative equilibrium Monte Carlo calculations, we provide a firm theoretical basis for the frequency distribution adjustment method of Bjorkman & Wood, and we demonstrate that this method provides the correct frequency distribution through the additional requirement of radiative equilibrium. We discuss the advantages and limitations of this approach, and show that it can easily be combined with the presence of additional heating sources and the concept of photon weighting. However, the method may fail if small dust grains are included... (abridged)

astro-ph↗