The evolution of stars paired with supermassive black holes

A star located in the close vicinity of a supermassive black hole (SMBH) in a galactic nucleus or a globular-cluster core could form a close binary with the SMBH, with the star possibly filling its Roche lobe. The evolution of such binary systems is studied assuming that the SMBH mainly accretes matter from the companion star and that the presence of gas in the vicinity of the SMBH does not appreciably influence variations in the star’s orbit. The evolution of the star–SMBH system is mainly determined by the same processes as those determining the evolution of ordinary binaries. The main differences are that the star is subject to an incident flux of hard radiation arising during the accretion of matter by the SMBH, and, in detached systems, the SMBH captures virtually all the wind emitted by its stellar companion, which appreciably influences the evolution of the major axis of the orbit. Moreover, the exchange between the orbital angular momentum and the angular momentum of the overflowing matter may not be entirely standard in such systems. The computations assume that there will be no such exchange of angular momentum if the characteristic timescale for mass transfer is shorter than the thermal time scale of the star. The absorption of external radiation in the stellar envelope was computed using the same formalism applied when computing the opacity of the stellar matter. The numerical simulations show that, with the adopted assumptions, three types of evolution are possible for such a binary system, depending on the masses and the initial separation of the SMBH and star. Type I evolution leads to the complete destruction of the star. Only this type of evolution is realized for low-mass main-sequence (MS) stars, even those with large initial separations from their SMBHs. Massive MS stars will also be destroyed if the initial separation is sufficiently small. However, two other types of evolution are possible for massive stars, with a determining role in the time variations of the parameters of the star–SMBH system being played by the possible growth of the massive star into a red giant during the time it is located in the close vicinity of the SMBH. Type II evolution can be realized for massive MS stars that are initially farther from the SMBH than in the case of disruption. In this case, the massive star fills its Roche lobe during its expansion, but is not fully destroyed; the star retreats inside its Roche lobe after a period of intense mass loss. This type of evolution is characterized by an increase in the orbital period of the system with time. As a result, the remnant of the star (its former core) is preserved as a white dwarf, and can end up at a fairly large distance from the SMBH. Type III evolution can be realized formassiveMSstars that are initially located still farther from their SMBHs, and also for massive stars that are already evolved at the initial time. In these cases, the star moves away from the SMBH without filling its Roche lobe, due to its intense stellar wind. The remnants of such stars can also end up at a fairly large distances from their SMBHs.     

Close binaries containing Supermassive Black Holes

We consider the evolution of binary systems formed by a Supermassive Black Hole (SMBH) residing in the center of a galaxy or a globular cluster and a star in its immediate vicinity. The star is assumed to fill its Roche lobe, and the SMBH accretes primarily the matter of this star. The evolution of such a system is mainly determined by the same processes as for an ordinary binary. The main differences are that the donor star is irradiated by hard radiation emitted during accretion onto the SMBH; in a detached system, nearly all the donor wind is captured by the black hole, which strongly affects the evolution of the semi-major axis; it is not possible for companions of the most massive SMBHs to fill their Roche lobes, since the corresponding orbital separations are smaller than the radius of the last stable orbit in the gravitational field of the SMBH. Moreover, there may not be efficient exchange between the orbital angular momentum and the angular momentum of the overflowing matter in such systems. Our computations assumed that, if the characteristic timescale for mass transfer is smaller than the thermal timescale of the star, no momentum exchange occurs. Absorption of incident external radiation in the stellar envelope was treated using the same formalism that was used when computing the radiative transfer in the stellar atmosphere. Numerical simulations show that Roche-lobe overflow is possible for a broad range of initial system parameters. The evolution of semi-detached systems containing a star and a SMBH nearly always ends with the dynamical disruption of the star. Stars with masses close to the solar mass are destroyed immediately after they fill their Roche lobes. During the accretion of matter of disrupted stars, the SMBH can achieve quasar luminosities. If the SMBH accretes ambient gas as well as gas stripped from stars, the star is subject to additional radiation in the detached phase of its evolution, strengthening its stellar wind. This leads to an increase of the semi-major axis and subsequent decrease of the probability of Roche-lobe overflow during the subsequent evolution of the system.     

An estimate of the probability of capture of a binary star by a supermassive black hole

A simple model for the dynamics of stars located in a sphere with a radius of one-tenth of the central parsec, designed to enable estimation of the probability of capture in the close vicinity (r < 10^?3 pc) of a supermassive black hole (SMBH) is presented. In the case of binary stars, such a capture with a high probability results in the formation of a hyper-velocity star. The population of stars in a sphere of radius <0.1 pc is calculated based on data for the Galactic rotation curve. To simulate the distortion of initially circular orbits of stars, these are subjected to a series of random shock encounters (“kicks”), whose net effect is to “push” these binary systems into the region of potential formation of hyper-velocity stars. The mean crossing time of the border of the close vicinity of the SMBH (r < 10^?3 pc) by the stellar orbit can be used to estimate the probability that a binary system is captured, followed by the possible ejection of a hyper-velocity star.     

The disruption of close binaries in the gravitational field of a supermassive black hole and the formation of hypervelocity stars

The formation of hypervelocity stars due to the dynamical capture of one component of a closebinary system by the gravitational field of a supermassive black hole (SMBH) is modeled. The mass of the black hole was varied between 10⁶ and 10⁹ M _⊙. In the model, the problem was considered first as a three-body problem (stage I) and then as an N-body problem (stage II). In the first stage, the effect of the inclination of the internal close-binary orbit (the motion of the components about the center of mass of the binary system) relative to the plane of the external orbit (the motion of the close binary around the SMBH) on the velocity with which one of the binary components is ejected was assessed. The initial binary orbits were generated randomly, with 10 000 orbits considered for each external orbit with a fixed pericenter distance r _p . Analysis of the results obtained in the first stage of the modeling enables determination of the binary-orbit orientations that are the most favorable for high-velocity ejection, and estimation of the largest possible ejection velocities V _max. The boundaries of the region of stellar disruption derived from the balance of tidal forces and self-gravitation are discussed using V _max-r _p plots, which generalize the results of the first stage of the modeling. Since a point-mass representation does not enable predictions about the survival of stars during close passages by a SMBH, there is the need for a second stage of the modeling, in which the tidal influence of the SMBH is considered. An approach treating a star like a structured finite object containing N bodies (N = 4000) enables the derivation of more accurate limits for the zone of efficient acceleration of hypervelocity stars and the formulation of conditions for the tidal disruption of stars.     

Statistics of Stars with Relativistic Velocities

The scenario for the dynamical capture of a binary system in the neighborhood of a supermassive black hole used byHills in 1988 to predict the existence of hypervelocity stars (~1000 km/s) allows the existence of stars with relativistic velocities attaining (1/3?2/3)c, where c is the speed of light. The increase of the kinetic energy of these stars by more than a factor of 100 is due to the replacement of one component of the binary with a supermassive black hole. This scenario takes candidate of relativisti©velocity stars outside our Galaxy, into intergalactic space, where they could be ejected from merging galaxies populated by supermassive black holes. At present, this is a hypothetical class of stars with anomalous kinematics, but it is already posing a serious challenge for modern astrometry, which, like 300 years ago, is still concerned with the detection of proper motions. While this was related to stars in the solar neighborhood at the time of Halley, is now a problem for studies of the most remote and weakest stars in intergalactic space. Possibilities for detecting such stars must be based on estimates of their abundances, that is, on their statistics. This paper is based on a presentation made at the conference “Modern Astrometry 2017,” dedicated to the memory of K.V. Kuimov (Sternberg Astronomical Institute, Moscow State University, October 23–25, 2017).

     

The nature hypervelocity stars

We list and analyze the main currently known mechanisms for accelerating the space motions of stars. A high space velocity of a star can be a consequence of its formation in the early stages of the evolution of a massive galaxy, when it was spheroidal and non-stationary, so that stars were born with velocities close to the escape velocity for the galaxy. Another possibility is that the star arrived from another galaxy with a velocity that is high for our Galaxy. The decay of unstable close multiple stars or supernova explosions in close binaries can also provide velocities of up to several hundreds of km/s to main-sequence stars and velocities of up to ∼1000 km/s to degenerate stars, neutron stars, and stellar-mass black holes. The merger of components of a binary system containing two neutron stars or a neutron star and a black hole due to gravitational-wave radiation can accelerate the nascent black hole to a velocity∼1000 km/s. Hypervelocity relativistic stars can be born due to asymmetric neutrino ejection during a supernova explosion. Stars can be efficiently accelerated by single and binary supermassive black holes (with masses from several millions to several billions of solar masses) in the nuclei of galaxies. Thanks to their gravitational field and fast orbital motion (in the case of binary objects), supermassive black holes are able to accelerate even main-sequence stars to relativistic velocities.     

Fast magnetic reconnection and particle acceleration in the non-equilibrium magnetosphere of a relativistic star

A scenario for hard impulsive flares due to magnetic reconnection and particle acceleration in cosmic plasma is proposed. The properties of fast reconnection in an appreciably non-equilibrium nagnetosphere of a compact relativistic object, such as a neutron star, magnetar, or white dwarf, are discussed. Such a magnetosphere could form as the result of the action of a relativistic shock on the strong magnetic field of the star. An analytical solution is presented for the generalized, two-dimensional structure, shape, and boundaries of the magnetosphere, together with the magnitudes of the direct and reverse currents in the reconnecting current layer. The uncompensated magnetic force acting on the reverse current is determined. The characteristic parameters of the non-equilibrium magnetospheres of compact stellar objects are estimated. The excess magnetic energy of the magnetosphere is comparable to the mechanical energy carried by the shock at the time of impact. The possible acceleration of particles to gigantic energies is discussed.     

Formation and evolution of Ae and Be stars

Several scenarios for the formation of accretion and decretion disks in single and binary Ae and Be stars are proposed. It is shown that, in order for a rapidly rotating main-sequence Be star to lose mass via a disk, the star’s rotation must be quasi-rigid-body. Estimates show that such rotation can be maintained by the star’s magnetic field, which is probably a relict field. The evolution of single Be main-sequence stars is numerically simulated allowing for mass loss via the stellar wind and rotational mass loss assuming rigid-body rotation. The stellar wind is the factor that determines the maximum mass of Be stars, which is close to 30M _⊙. The evolution of Be stars in close binaries is analyzed in the approximation adopted in our scenario. Long gamma-ray bursts can be obtained as a result of the collapse of rapidly rotating oxygen—neon degenerate dwarfs—the accreting companions of Be stars—into neutron stars.     

Wolf-Rayet stars and relativistic objects: Distinctions between the mass distributions in close binary systems

The observed properties of Wolf-Rayet stars and relativistic objects in close binary systems are analyzed. The final masses M _CO ^f for the carbon-oxygen cores of WR stars in WR + O binaries are calculated taking into account the radial loss of matter via stellar wind, which depends on the mass of the star. The analysis includes new data on the clumpy structure of WR winds, which appreciably decreases the required mass-loss rates $\dot M_{WR}$ for the WR stars. The masses M _CO ^f lie in the range (1–2)M _⊙–(20–44)M _⊙ and have a continuous distribution. The masses of the relativistic objects M _x are 1–20M _⊙ and have a bimodal distribution: the mean masses for neutron stars and black holes are 1.35 ± 0.15M _⊙ and 8–10M _⊙, respectively, with a gap from 2–4M _⊙ in which no neutron stars or black holes are observed in close binaries. The mean final CO-core mass is $\overline M _{CO}^f = 7.4 - 10.3M_ \odot$ , close to the mean mass for the black holes. This suggests that it is not only the mass of the progenitor that determines the nature of the relativistic object, but other parameters as well-rotation, magnetic field, etc. One SB1R Wolf-Rayet binary and 11 suspected WR + C binaries that may have low-mass companions (main-sequence or subgiant M-A stars) are identified; these could be the progenitors of low-mass X-ray binaries with neutron stars and black holes.     

Evolution of close binaries and gamma-ray bursts

We analyze the late stages of evolution of massive (M ₀ ? 8 M _⊙) close binaries, from the point of view of possible mechanisms for the generation of gamma-ray bursts. It is assumed that a gamma-ray burst requires the formation of a massive (～1 M _⊙), compact (R ? 10 km) accretion disk around a Kerr black hole or neutron star. Such Kerr black holes are produced by core collapses of Wolf-Rayet stars in very close binaries, as well as by mergers of neutron stars and black holes or two neutron stars in binaries. The required accretion disks can also form around neutron stars that were formed via the collapse of ONeMg white dwarfs. We estimate the Galactic rate of events resulting in the formation of rapidly rotating relativistic objects. The computations were carried out using the “Scenario Machine.”     

Population synthesis of rapidly rotating main-sequence stars with companions

Usingthe “Scenario Machine” (a specialized numerical code formodeling the evolution of large ensembles of binary systems), we have studied the physical properties of rapidly rotating main-sequence binary stars (Be stars) with white-dwarf companions and their abundance in the Galaxy. The calculations are the first to take into account the cooling of the compact object and the effect of synchronization of the rotation on the evolution of Be stars in close binaries. The synchronization time scale can be shorter than the main-sequence lifetime of a Be star formed during the first mass transfer. This strongly influences the distribution of orbital periods for binary Be stars. In particular, it can explain the observed deficit of short-period Be binaries. According to our computations, the number of binary systems in the Galaxy containing a Be star and white dwarf is large: 70–80% of all Be stars in binaries should have degenerate dwarf companions. Based on our calculations, we conclude that the compact components in these systems have high surface temperatures. Despite their high surface temperatures, the detection of white dwarfs in such systems is hampered by the fact that the entire orbit of the white dwarf is embedded in the dense circumstellar envelope of the primary, and all the extreme-UV and soft X-ray emission of the compact object is absorbed by the Be star’s envelope. It may be possible to detect the white dwarfs via observations of helium emission lines of Be stars of not very early spectral types. The ultraviolet continuum energies of these stars are not sufficient to produce helium line emission. We also discuss numerical results for Be stars with other evolved companions, such as helium stars and neutron stars, and suggest an explanation for the absence of Be-black-hole binaries.     

The chemical evolution of globular clusters

The sequence of events determining the initial stages of star formation is analyzed in framework of the self-enrichment scenario. The computations are based on a single-zone chemo-dynamical model. It is shown that the first episode of star formation was characterized by an initial mass function shifted toward massive stars (M ≥ 8M_⊙). We argue that the transition to a star formation with a normal (Salpeter) initial mass function was due to more efficient radiative cooling of the proto-globular cluster gas after its enrichment to a metallicity of Z ～ 0.02 Z_⊙ in agreement with those observed in globular clusters.     

Signs of magnetic accretion in X-ray pulsars

The spin-down mechanism of accreting neutron stars is discussedwith an application to one of the best studied X-ray pulsars GX301-2. We show that the maximum possible spin-down torque applied to a neutron star from the accretion flow can be evaluated as K _sd ^(t) = ??²/(r _m r _cor)^3/2. The spin-down rate of the neutron star in GX301-2 can be explained provided the magnetospheric radius of the neutron star is smaller than its canonical value. We calculate the magnetospheric radius considering the mass-transfer in the binary system in the frame of the magnetic accretion scenario suggested by V.F. Shvartsman. The spin-down rate of the neutron star expected within this approach is in a good agreement with that derived from observations of GX301-2.     

High-velocity stars as a result of encounters between stars and massive binary black holes in galactic nuclei

Numerical simulations of the motions of stars in the gravitational fields of binary black holes with various component mass ratios have been carried out. Two models are considered: (1) the two-body problem with two fixed centers; (2) the general three-body problem. The first model is applicable only over short times Δt ? T, where T is the period of the binary system. The second model is applicable at all times except for during close encounters of stars with one of the binary components, r ≤ 0.00002 pc, where r is the distance from the star to the nearer black hole. In very close passages, relativistic corrections must be taken into account. Estimates of the probability of formation of high-velocity stars as a result of such interactions are obtained. It is shown that this mechanism is not suitable for the nucleus of our Galaxy due to the probable absence of a second massive black hole in the central region of the Galaxy.     

The origin of the rotational and spatial velocities of radio pulsars

We analyze possible origins of the observed high rotational and spatial velocities of radio pulsars. In particular, these can be understood if all radio pulsars originate in close binary systems with orbital periods of 0.1–100 days, with the neutron star being formed by a type Ib,c supernova. The high spatial velocities of pulsars (v _p up to 1000 km/s) reflect the high Keplerian velocities of the components of these binaries, while their short periods of rotation (P _p < 4 s) are due to the rapid rotation of the presupernova helium-star components with masses of 2.5–10 M_⊙, which is synchronous with their orbital rotation. Single massive stars or components in wide binaries are likely to produce only slowly rotating (P _p > 4 s) neutron stars or black holes, which cannot be radio pulsars. As a result, the rate of formation of radio pulsars should be a factor of a few lower than the rate of type II and type Ib,c supernovae estimated from observations. This scenario for the formation of radio pulsars is supported by (i) the bimodal spatial velocity distribution of radio pulsars; (ii) the coincidence of the observed spatial velocities of radio pulsars with the orbital velocities of the components of close binaries with nondegenerate helium presupernovae; (iii) the correlation between the orbital and rotational periods for 22 observed radio pulsars in binaries with elliptical orbits; and (iv) the similarity of the observed rate of formation of radio pulsars and the rate of type Ib,c supernovae.     

A study of the evolution of the close binaries Cyg X-3, IC 10 X-1, NGC 300 X-1, SS 433, and M33 X-7 using the “scenario machine”

Evolutionary tracks for the X-ray binaries Cyg X-3, IC 10 X-1, NGC 300 X-1, SS 433, and M33 X-7 are computed using the Scenario Machine code. The compact objects in IC 10 X-1, NGC 300 X-1, and M33 X-7 are the most massive stellar-mass black-hole candidates. Cyg X-3, IC 10 X-1, and NGC 300 X-1 are the only currently known Wolf-Rayet stars with degenerate companions. SS 433 is the only known superaccretor in the Milky Way. Therefore, the stars studied provide excellent laboratories for testing scenarios for the evolution of binaries under extreme conditions. The classical evolutionary scenario is consistent with modern observational data. During the evolution of these binaries, hypernova explosions accompanied by the collapse of stellar cores with large angular momenta can occur, leading to long gamma-ray bursts. At the end of their evolution, Cyg X-3, IC 10 X-1, NGC 300 X-1, and SS 433 may form binary relativistic objects, which will subsequently merge due to the radiation of gravitational waves. The gravitational waves emitted during mergers of relativistic stars should be detectable by existing and future gravitational-wave antennas. In the course of its future evolution, M33 X-7 will pass through a Thorne-?ytkow stage. The formation of a Thorne-?ytkow object can also be accompanied by gravitational-wave radiation.     

Late stages of the evolution of close compact binaries: Type I supernovae,gamma-ray bursts,and supersoft X-ray sources

We consider the evolution of close binaries resulting in the most intensive explosive phenomena in the stellar Universe—Type Ia supernovae and gamma-ray bursts. For Type Ia supernovae, which represent thermonuclear explosions of carbon-oxygen dwarfs whose masses reach the Chandrasekhar limit during the accretion of matter from the donor star, we derive the conditions for the accumulation of the limiting mass by the degenerate dwarf in the close binary. Accretion onto the degenerate dwarf can be accompanied by supersoft X-ray radiation with luminosity 1–10⁴ L _⊙. Gamma-ray bursts are believe to accompany the formation and rapid evolution of compact accretion-decretion disks during the formation of relativistic objects—black holes and neutron stars. The rapid (～1 M _⊙/s) accretion of matter from these disks onto the central compact relativistic star results in an energy release of ～0.1 M _⊙ c ² ～ 10⁵³ erg in the form of gamma-rays and neutrinos over a time of 0.1–1000 s. Such disks can form via the collapse of the rapidly rotating cores of Type Ib, Ic supernovae, which are components in extremely close binaries, or alternately due to the collapse of accreting oxygen-neon degenerate dwarfs with the Chandrasekhar mass into neutron stars, or the merging of neutron stars with neutron stars or black holes in close binaries. We present numerical models of the evolution of some close binaries that result in Type Ia supernovae, and also estimate the rates of these supernovae (～0.003/year) and of gamma-ray bursts (～10^?4/year) in our Galaxy for various evolutionary scenarios. The collimation of the gamma-ray burst radiation within an opening angle of several degrees “matches” the latter estimate with the observed rate of these events, ～10^?7–10^?8/year calculated for a galaxy with the mass of our Galaxy.     

Studies of classical barium stars

Using atmosphere models based on high-resolution spectra, we have derived the abundances of chemical elements in the atmospheres of seven classical barium stars and compared them with the elemental abundances of moderate barium stars and normal red giants. The behavior of elements up to the iron peak is the same in all three groups of giants, providing evidence that they have a common origin. The dependence of the anomalous abundances of s-process elements on stellar mass and metallicity is qualitatively similar for all three groups, probably indicating that a substantial role is played by the evolutionary phase of the stars. We conclude that the barium-star phenomenon and the overabundances of s-process elements in barium stars cannot be explained as a consequence of binarity alone. The extent to which the s-process elements are overabundant is affected by the mass, metallicity, and evolutionary phase of the given star, and any of these parameters may prove to be important in a specific object.     

The probability of forming hyperveloicty stars in the Galaxy

The probability of forming a Galactic hypervelocity star is estimated for the scenario of Hills, which describes the dynamical capture of one component of a binary star by the gravitational field of the supermassive black hole in the Galactic center, leading to the ejection of the other component. Ten thousand initial orientations of the binary orbits were considered, and the semi-major axes of the binary orbits were varied in a wide range from 11.3 R _⊙ to 425 R _⊙. Two series of computations were carried out, in which the mass of the supermassive black hole was taken to be 10⁶ M _⊙ and 3.4 × 10⁶ M _⊙. Numerical simulations of encounters of the binary and black hole in the framework of the three-body and N-body problems are used to localize regions favorable for the formation of hypervelocity stars. The motion of the ejected star in the regular field of the Galaxy is calculated, and the conditions under which the star escapes the Galaxy defined. The probability of escaping the Galaxy is caluclated as a function of various parameters the initial separation of the binary components and the distance of the binary from the black hole. On average, the probability of forming a hypervelocity star is higher for closer encounters and more tightly bound binary pairs.

     

The spin-down mechanism of the X-ray pulsar 4U 2206+54

Observations of the X-ray pulsar 4U 2206+54 obtrained over 15 years show that its period, which is now 5555 ± 9 s, is increasing dramatically. This behavior is difficult to explain using traditional scenarios for the spin evolution of compact stars. The observed spin-down rate of the neutron star in 4U 2206+54 is in good agreement with the value expected in a magnetic-accretion scenario, taking into account that, under certain conditions, the magnetic field of the accretion stream can affect the geometry and type of flow. The neutron star in this case accretes material from a dense gaseous slab with small angular momentum, which is kept in equilibrium by the magnetic field of the flow itself. A magnetic-accretion scenario can be realized in 4U 2206+54 if the magnetic-field strength at the surface of the optical counterpart to the neutron star is higher than 70 G. The magnetic field at the surface of the neutron star is 4 × 10¹² G in this scenario, in agreement with estimates based on an analysis of X-ray spectra of the pulsar.