I. Iben et al., ON THE EVOLUTION OF LOW-MASS X-RAY BINARIES UNDER THE INFLUENCE OF A DONOR STELLAR WIND-INDUCED BY X-RAYS FROM THE ACCRETOR, The Astrophysical journal, 486(2), 1997, pp. 955-986
In a low-mass X-ray binary (LMXB), an intense stellar wind from the ma
ss donor may be a consequence of the absorption of X-rays from the mas
s-accreting neutron star or black hole, and such a wind could change t
he evolution of these binaries dramatically compared with the evolutio
n of cataclysmic variables (CVs), which are close binaries in which th
e accretor is a white dwarf. An analytical study and numerical models
show that, in the closest and brightest LMXBs, a relativistic companio
n can capture up to similar to 10% of the mass lost in the induced ste
llar wind (ISW) from the main-sequence or subgiant donor, and this is
enough to keep the X-ray luminosity of a typical LMXB on the level of
L-X similar to 5000 L. and to accelerate the rotation of an old neutro
n star with a low magnetic field into the millisecond-period range. A
self-sustained ISW may exist even if the donor does not fill its Roche
lobe, but the system can be bright (L-X > 100 L.) only if the radius
of the donor is a substantial fraction (greater than or similar to 0.8
) of the Roche lobe radius. A lower limit on the Roche lobe filling fa
ctor follows from the circumstance that both the rate (E) over dot(win
d) at which work must be done to lift wind matter off the donor and th
e rate (E) over dot(abs) at which the donor absorbs X-ray energy are p
roportional to (M) over dot(ISW) (the ISW mass-loss rate) and from the
requirement that (E) over dot(wind) < (E) over dot(abs) in order for
energy to be conserved. The observed number (similar to 100) of bright
LMXBs in our Galaxy can be understood as the product of a relatively
short lifetime (a few x 10(7) yr) and a small theoretical birthrate (s
imilar to 2 x 10(-6)-8 x 10(-6) yr(-1)), which is comparable to semiem
pirical estimates of the birthrate of LMXBs and millisecond pulsars (s
imilar to 2 x 10(-6) yr(-1)). The theoretical lifetime is similar to 1
0-60 times shorter than when the ISW is not taken into account, and th
e theoretical birthrate is similar to 3-6 times smaller, because of th
e fact that the ISW acts to expand the orbit and reduce the number of
systems that can evolve through an X-ray bright stage under the influe
nce of a magnetic stellar wind (MSW) when the donor is a main-sequence
star (CV-like LMXBs), or under the influence of nuclear evolution whe
n the donor is a subgiant or giant with a degenerate helium core (Algo
l-like LMXBs) of mass in the range M-He = 0.13-0.45 M.. The observed c
oncentration of LMXBs in the 3-24 hr orbital period range corresponds
to a similar concentration in the CV distribution and could be interpr
eted as evidence that the MSW in LMXBs operates at a strength not too
different from its strength in CVs. For 0.3-1 M. main-sequence donors,
if the radius of the donor is larger than similar to 70% of the Roche
lobe radius, the tendency of the ISW to force orbital expansion can b
alance the braking influence of the MSW and prevent an LMXB with a mai
n-sequence donor from evolving to periods less than similar to 3 hr. W
hen a main-sequence donor becomes completely convective (donor mass si
milar to 0.1-0.3 M., depending on the mass-loss rate) and the MSW shut
s off, orbital angular momentum loss due to gravitational wave radiati
on (GWR) is unable to counter the tendency toward expansion, and this
may explain the apparent absence of short-period (P-orb < 3 hr) LMXBs
with main-sequence donors. This contrasts with the CV family in which
the number of systems in the Galaxy with P-orb similar to 1.3-2 hr (wi
th donor mass less than or equal to 0.3 M. and with evolution driven o
nly by GWR) is larger by a factor of similar to 100 than the number of
systems with P-orb > 3 hr. In Algol-like LMXBs in the Galactic disk,
the timescale for the evaporation (caused by the ISW) of the donor wit
h a low-mass, degenerate helium core can be smaller than the timescale
for the radial expansion of the donor owing to nuclear evolution, and
the donor may never fill its Roche lobe. However, if progenitor binar
ies are initially wide enough, the donor may escape evaporation as a m
ain-sequence star, and significant mass transfer may not occur until t
he secondary evolves into a giant with a degenerate helium core of lar
ge mass and fills its Roche lobe. In globular clusters, as a result of
capture and exchange reactions, semidetached Algol-like LMXBs can be
formed in which the donor can fill its Roche lobe even when its degene
rate helium core is of small mass, and Roche lobe mediated mass transf
er driven by the nuclear evolution of the donor can dominate over capt
ure from the ISW. The numerical models formally imply the possible pre
sence in the Galaxy of similar to 10(4) dim (L-X similar to 1-100 L.),
long-period LMXBs or radio pulsars with low-mass (similar to 0.05 M.)
companions. Since there are few, if any, known observational counterp
arts of these systems, it is necessary to invoke a mechanism or mechan
isms to destroy their formal progenitors. Possible destruction mechani
sms include: (1) evaporation driven by the radiation from the rapidly
rotating pulsar into which the accretor has been transformed by accret
ion during the bright LMXB phase, and (2) a dynamical instability aris
ing when the donor is almost completely convective and fills its Roche
lobe. In the case of dynamical disruption, the donor may be transform
ed into the envelope of a Thorne-Zytkow (1975) object with a neutron s
tar or black hole core or into a planet-forming disk around the neutro
n star or black hole. A few short-period (P-orb < 3 hr) LMXBs do exist
, and, in them, the donor may be a helium white dwarf of mass less tha
n similar to 0.09 M.. An ISW operating before the donor fills its Roch
e lobe may be responsible for reducing the mass of the white dwarf fro
m an initial value of greater than or equal to 0.13 M. to a value of l
ess than or equal to 0.09 M., thus permitting stable mass exchange (at
a rate smaller than the Eddington limiting rate) and evolution to lon
ger periods to occur after the donor fills its Roche lobe. Another sce
nario relies on the collapse of a massive oxygen-neon white dwarf, whi
ch has accreted from a Roche lobe filling helium white dwarf. Problems
that must be explored further in order to acquire a better understand
ing of the evolution of LMXBs include the formation of a corona around
an irradiated low-mass main-sequence or degenerate dwarf star, accret
ion of ISW matter by a neutron star or black hole companion, the effec
t of an ISW on the MSW, formation of millisecond pulsars, co