ALBERT

Description: The effects of a giant impact on Uranus with respect to the axis tilt of Uranus and its satellites are discussed. The simulations of possible giant impacts were carried out using Cray supercomputers. The technique used is called smooth particle hydrodynamics (SPH). In this technique, the material in the proto-Uranus planet and in the impactor is divided into a large number of particles which can overlap one another so that local averages over these particles determine density and pressure in the problem, and the particles themselves have their own temperatures and internal energies. During the course of the simulation, these particles move around under the influence of the forces acting on them: gravity and pressure gradients. The results of model simulations are presented.

Description: The break-up of Periodic Comet Shoemaker-Levy 9 into multiple pieces following its grazing encounter with Jupiter in July 1992 can be used to study tidally-induced fracture in comets. This spectacular event allows us not only to set limits on the size, strength and density of Shoemaker-Levy 9 itself, but provides invaluable guidance to numerical modeling of such encounters. In an extensive treatment of tidal breakup which assumed self-gravitating, homogeneous, perfectly elastic bodies, Dobrovolskis derived simple analytical expressions for the tidally-induced surface and central stresses. Both can be cast in such a way that Poisson's ratio is the only material dependent constant entering these expressions. Whether both surface and central failure must be initiated as a criterion for breakup, or either one of them is sufficient, remains a subject of disagreement. To resolve this debate, we model the details of cometary breakup using a three-dimensional Smooth Particle Hydrodynamics (SPH) code modified to simulate fracture in small solid objects. At lower stresses associated with brittle failure, we use a rate-dependent strength based on the nucleation of incipient flaws whose number density is given by a Weibull distribution.

Description: Impact phenomena shaped our solar system. As usual for most solar system processes, the scales are far different than we can address directly in the laboratory. Impact velocities are often much higher than we can achieve, sizes are often vastly larger, and most impacts take place in an environment where the only gravitational force is the mutual pull of the impactors. The Smooth Particle Hydrodynamics (SPH) technique has been applied in the past to the simulations of giant impacts. In these simulations, the colliding objects were so massive (at least a sizeable fraction of the Earth's mass) that material strength was negligible compared to gravity. This assumption can no longer be made when the bodies are much smaller. To this end, we have developed a 3D SPH code that includes a strength model to which we have added a von Mises yielding relation for stresses beyond the Hugoniot Elastic Limit. At the lower stresses associated with brittle failure, we use a rate-dependent strength based on the nucleation of incipient flaws whose number density is given by a Weibull distribution. Following Grady and Kipp and Melosh et al., we introduce a state variable D ('damage'), 0 less than D less than 1, which expresses the local reduction in strength due to crack growth under tensile loading. Unfortunately for the hydrodynamics, Grady and Kipp's model predicts which fragments are the most probable ones and not the ones that are really formed. This means, for example, that if a given laboratory experiment is modeled, the fragment distribution obtained from the Grady-Kipp theory would be equivalent to a ensemble average over many realizations of the experiment. On the other hand, the hydrodynamics itself is explicit and evolves not an ensemble average but very specific fragments. Hence, there is a clear incompatibility with the deterministic nature of the hydrodynamics equations and the statistical approach of the Grady-Kipp dynamical fracture model. We remedy these shortcomings by making the incipient flaw distribution explicit, i.e., particles carry activation strains which are distributed at random with a probability of occurrence given by the Weibull distribution. If the local principal axis strain exceeds this limit, damage starts to grow. By growing explicit cracks together with statistical cracks (damage) at the sub-particle scale, we ensure that material strength and fragmentation is independent of model resolution. We tested our scheme by simulating laboratory impact experiments on basalt spheres.

Description: We present three-dimensional Newtonian simulations of the coalescence of two neutron stars, using a smoothed particle hydrodynamics (SPH) code. We begin the simulations with the two stars in a hard, circular binary, and have them spiral together as angular momentum is lost through gravitational radiation at the rate predicted by modeling the system as two point masses. We model the neutron stars as hard polytropes (gamma = 2.4) of equal mass, and investigate the effect of the initial spin of the two stars on the coalescence. The process of coalescence, from initial contact to the formation of an axially symmetric object, takes only a few orbital periods. Some of the material from the two neutron stars is shed, forming a thick disk around the central, coalesced object. The mass of this disk is dependent on the initial neutron star spins; higher spin rates result in greater mass loss and thus more massive disks. For spin rates that are most likely to be applicable to real systems, the central coalesced object has a mass of 2.4 solar mass, which is tantalizingly close to the maximum mass allowed by any neutron star equation of state for an object that is supported in part by rotation. Using a realistic nuclear equation of state, we estimate the temperature of the material after the coalescence. We find that the central object is at a temperature of approximately 10 MeV, while the disk is heated by shocks to a temperature of 2 to 4 MeV.

Description: We simulated encounters between a neutron star and primordial and tidal-capture binaries. In the case of encounters involving a tidal-capture binary, comprising a white dwarf and a main-sequence star, we find that most exchange encounters will produce a single merged object with the white dwarf and neutron star engulfed in a common envelope of gas donated by the main-sequence primary of the original binary. A small fraction of exchanges induce a merger of the white dwarf and main-sequence star, with this object being unbound to the neutron star, and the two objects having a large relative speed at infinity. For encounters involving a primordial binary, fewer encounters require the inclusion of hydrodynamical effects. Those involving collisions or close encounters tend to produce a binary comprised of the two merged stars (now forming one star) and the third star. The binaries produced typically have large enough separations to prevent the formation of a single merged object until subsequent stellar evolution of one of the components causes it to fill its Roche lobe. Clean exchanges produce binaries with large eccentricities; they are typically sufficiently wide to avoid circularization.

Description: Numerical simulations have been used to study high velocity two-body impacts. In this paper, a two-dimensional Largrangian finite difference hydro-code and a three-dimensional smooth particle hydro-code (SPH) are described and initial results reported. These codes can be, and have been, used to make specific predictions about particular objects in our solar system. But more significantly, they allow us to explore a broad range of collisional events. Certain parameters (size, time) can be studied only over a very restricted range within the laboratory; other parameters (initial spin, low gravity, exotic structure or composition) are difficult to study at all experimentally. The outcomes of numerical simulations lead to a more general and accurate understanding of impacts in their many forms.

Description: The dynamic mass exchange process in doubly degenerate binaries was investigated using a three-dimensional numerical simulation of the evolution of a doubly degenerate binary system in which the primary is a 1.2-solar-mass white dwarf and the Roche lobe filling secondary is a 0.9-solar-mass dwarf. The results show that, in a little more than two orbital periods, the secondary is completely destroyed and transformed into a thick disk orbiting about the primary. Since only a very small fraction of the mass (0.0063 solar mass) escapes the system, the evolution of the binary results in the formation of a massive object. This object is composed of three parts, the initial white dwarf primary, a very hot pressure-supported spherical envelope, and a rotationally supported outer disk. The evolution of the system can be understood in terms of a simple analytical model where it is shown that the angular momentum carried by the mass during the transfer and stored in the disk determines the evolution of the system.

Description: We simulated encounters involving binaries of two eccentricities: e = 0 (i.e., circular binaries) and e = 0.5. In both cases the binary contained a point mass of 1.4 solar masses (i.e., a neutron star) and a 0.8 solar masses main-sequence star modeled as a polytrope. The semimajor axes of both binaries were set to 60 solar radii (0.28 AU). We considered intruders of three masses: 1.4 solar masses (a neutron star), 0.8 solar masses (a main-sequence star or a higher mass white dwarf), and 0.64 solar masses (a more typical mass white dwarf). Our strategy was to perform a large number (40,000) of encounters using a three-body code, then to rerun a small number of cases with a three-dimensional smoothed particle hydrodynamics (SPH) code to determine the importance of hydrodynamical effects. Using the results of the three-body runs, we computed the exchange across sections, sigma(sub ex). From the results of the SPH runs, we computed the cross sections for clean exchange, denoted by sigma(sub cx); the formation of a triple system, denoted by sigma(sub trp); and the formation of a merged binary with an object formed from the merger of two of the stars left in orbit around the third star, denoted by sigma(sub mb). For encounters between either binary and a 1.4 solar masses neutron star, sigma(sub cx) approx. 0.7 sigma(sub ex) and sigma(sub mb) + sigma(sub trp) approx. 0.3 sigma(sub ex). For encounters between either binary and the 0.8 solar masses main-sequence star, sigma(sub cx) approx. 0.50 sigma(sub ex) and sigma(sub mb) + sigma(sub trp) approx. 1.0 sigma(sub ex). If the main sequence star is replaced by a main-sequence star of the same mass, we have sigma(sub cx) approx. 0.5 sigma(sub ex) and sigma(sub mb) + sigma(sub trp) approx. 1.6 sigma(sub ex). Although the exchange cross section is a sensitive function of intruder mass, we see that the cross section to produce merged binaries is roughly independent of intruder mass. The merged binaries produced have semi-major axes much larger than either those of the original binaries or those of binaries produced in clean exchanges. Coupled with their lower kick velocities, received from the encounters, their larger size will enhance their cross section, shortening the waiting time to a subsequent encounter with another single star.

Description: A 3D smooth-particle hydrodynamic code with 7000 particles is used to simulate collisions between lower main sequence (MS) stars whose masses differ by a factor of five. Collisions between point-mass intruders and MS stars are simulated to study how the physics of the encounters depends on the finite radius of the smaller MS stars. It is found that the two MS stars become gravitationally bound in a physical collision if their relative velocity at infinity is less than a critical velocity Vd. Vd decreases from 1000 km/s in a head-on collision to 150 km/s in a grazing one. If the less massive star is replaced by a point mass, Vd remains the same for grazing collision and tidal encounters but drops to about 600 km/s in head-on collisions. The collision and coalescence of two MS stars resets the nuclear clock of the coalesced star, so it contracts to the MS. Tidal capture of binary stars in globular clusters can only occur in encounters in which the closest approach of the two stars to their center of mass is less than 2.0 times the sum of their radii.

Description: Encounters between a 1.4 solar mass neutron star and a 0.8 solar mass red giant (RG) and between a 1.4 solar mass neutron star (NS) and an 0.8 solar mass main-sequence (MS) star have been successfully simulated. In the case of encounters involving an RG, bound systems are produced when the separation at periastron passage R(MIN) is less than about 2.5 R(RG). At least 70 percent of these bound systems are composed of the RG core and NS forming a binary engulfed in a common envelope of what remains of the former RG envelope. Once the envelope is ejected, a tight white dwarf-NS binary remains. For MS stars, encounters with NSs will produce bound systems when R(MIN) is less than about 3.5 R(MS). Some 50 percent of these systems will be single objects with the NS engulfed in a thick disk of gas almost as massive as the original MS star. The ultimate fate of such systems is unclear.