Chemical Abundances in the Secondary Star of the Neutron Star Binary Centaurus X-4
Extracted source text · arXiv astro-ph/0504250
% \documentclass[12pt,preprint]{aastex}
\documentclass[preprint2]{aastex}
\newcommand{\myemail}{jonay@ll.iac.es}
%% You can insert a short comment on the title page using the command
%below.
%\slugcomment{Gonz\'alez Hern\'andez et al.}
%\shorttitle{Stellar Abundances in \mbox{Cen X-4}}
%\shortauthors{J. I. Gonz\'alez Hern\'andez et al.}
\begin{document}
%\doublespace
\title{Chemical Abundances in the Secondary Star of\\ the Neutron Star
Binary \mbox{Centaurus X-4}\footnotetext{Based on observations obtained with
UVES at VLT Kueyen 8.2 m telescope in programme 65.H-0447}}
\author{Jonay I. Gonz\'alez Hern\'andez, Rafael Rebolo\altaffilmark{1},\\
Garik Israelian and Jorge Casares}
\affil{Instituto de Astrof{\'\i }sica de Canarias, E-38205 La Laguna,
Tenerife,
SPAIN: \\jonay@iac.es, rrl@iac.es, gil@iac.es, jcv@iac.es}
\and
\author{Keiichi Maeda}
\affil{Department of Earth Science and Astronomy, Graduate School of Arts
and Science, University of Tokyo, Meguro-ku, Tokyo 153-8902, JAPAN:
maeda@esa.c.u-tokyo.ac.jp}
\and
\author{Piercarlo Bonifacio and Paolo Molaro}
\affil{Osservatorio Astronomico di Trieste, via Tiepolo 11, 34131
Trieste, ITALY\\ bonifaci@ts.astro.it, molaro@ts.astro.it}
\altaffiltext{1}{Consejo Superior de Investigaciones Cient{\'\i }ficas,
SPAIN}
\begin{abstract}
Using a high resolution spectrum of the secondary star in the neutron star
binary \mbox{Cen X-4}, we have derived the stellar parameters and veiling
caused by the accretion disk in a consistent way. We have used a $\chi^{2}$
minimization procedure to explore a grid of 1\,500\,000 LTE synthetic
spectra computed for a plausible range of both stellar and veiling
parameters. Adopting the best model parameters found, we have determined
atmospheric abundances of Fe, Ca, Ti, Ni and Al. These element
abundances are super solar ($\mathrm{[Fe/H]}=0.23 \pm 0.10$), but
only the abundance of Ti and Ni appear to be moderately enhanced
($\ge1\sigma$) as compared with the average values of stars of similar iron
content. These element abundances can be explained if the secondary star captured a
significant amount of matter ejected from a spherically symmetric supernova
explosion of a 4 {$M_\odot$} He core progenitor and assuming solar
abundances as primordial abundances in the secondary star. The kinematic
properties of the system indicate that the neutron star received a natal
kick velocity through an aspherical SN and/or an asymmetric neutrino
emission. The former scenario might be ruled out since our model
computations cannot produce acceptable fits to the observed abundances. We
have also examined whether this system could have formed in the Galactic
halo, and our simulations show that this possibility seems unlikely. We also
report a new determination of the Li abundance consistent with previous
studies that is unusually high and close to the cosmic Li abundance in the
Galactic disk.
\end{abstract}
\keywords{stars:neutron---stars:abundances---stars:individual
(\mbox{Cen X-4})---stars:X-rays:low-mass---binaries}
\section{Introduction}
\begin{figure*}[ht!]
\centering
\includegraphics[width=11cm,angle=90]{f1bw.eps}
\caption{Observed spectrum of the secondary star of \mbox{Cen X-4}
(top)
and of a properly broadened template (HD 216803, bottom). \label{fig1}}
\end{figure*}
The optical counterpart of the low mass X-ray binary (LMXB) \mbox{Cen
X-4} (V822 Cen) was firstly identified during an X-ray outburst in 1979
(Kaluzienski et al.\ 1980), ten years after the X-ray outburst detected by
the satellite {\it Vela 5B} during an X-ray outburst in 1969 (Conner et al.\
1969). At that moment, it resembled an $m_{V} \sim 13$ magnitude blue star
that had increased in brightness by $\sim$6 magnitudes from its pre-outburst
value (Canizares et al.\ 1980). About three weeks after the 1979 outburst a
type I X-ray burst was observed, which indicated that the X-ray source is a
neutron star (Matsuoka et al.\ 1980). Two weeks later the optical
counterpart had faded to $m_{V} = 18.2$ mag and spectral observations
revealed the presence of a late-type companion star (K3--K7) in the system
(van Paradijs et al.\ 1980). Later optical studies determined the orbital
period, $P = 0.629$ d and a system center-of-mass velocity with respect to
the Sun of $137 \pm 17$ {${\rm km}\:{\rm s}^{-1}$} (McClintock \& Remillard
1990). The corresponding mass function was then estimated at $f(M) = 0.23$
{$M_\odot$}, which further supported the argument for a neutron star primary
(rather than a black hole, McClintock \& Remillard 1990). More recently,
from measurements of the orbital inclination and binary mass ratio, the
compact object mass has been estimated at $M_{\rm NS} = 0.5$--2.1\
{$M_\odot$} (Shahbaz et al. 1993) and the companion star mass, in the range
$0.04 < M_2 < 0.58$ {$M_\odot$} (Torres et al.\ 2002).
The evolutionary status of this system has been widely discussed in the
literature. Whereas Chevalier et al.\ (1989) and McClintock \& Remillard
(1990) had suggested a scenario in which the secondary star was a ``stripped
giant'' with a very low mass ($M_2\sim0.1$ {$M_\odot$}), Shahbaz et al.\
(1993) favored the case for a subgiant. In both cases the secondary star
would be an evolved star that had spent more than 80 per cent of its life
(i.e., $\sim 10^{9}$--$10^{10}$ yr depending on its initial mass). On the
contrary, the compact object in this system is a neutron star, probably the
remnant of the SN explosion of the initial massive component of the binary.
The ejected material in the SN explosion is composed of heavy nuclei as
products of the hydrostatic and explosive nucleosynthesis in the massive
star.
Recent studies of chemical abundances of secondary stars in low mass
X-ray binaries have opened up a new window of information on the later
stages in the evolution of massive stars. A significant amount of
the ejected matter in any supernova explosion that formed the compact
objects in these systems could be captured by the companion stars. These
stars could therefore show anomalous abundances as signatures of the
explosion. Detailed chemical analysis of the atmospheres of the secondary
stars could constrain many parameters involved in supernova explosion
models, such as the mass cut, the amount of fall-back matter, any possible
mixing, and explosion energies and geometries. Indeed, the chemical analysis
of the secondary stars provides evidence for a supernova event in the
history of the systems Nova Sco 94 (Israelian et al.\ 1999; Brown et al.\
2000; Podsiadlowski et al.\ 2002) and \mbox{A0620--00} (Gonz\'alez
Hern\'andez et al.\ 2004).
In this paper we present the chemical abundance analysis of the secondary
star of the LMXB \mbox{Cen X-4} and discuss these results in the context
of possible evolutionary scenarios.
\section{Observations}
We obtained 20 spectra of \mbox{Cen X-4} with the UV--Visual Echelle
Spectrograph (UVES) at the European Southern Observatory (ESO), {\itshape
Observatorio Cerro Paranal}, using the 8.2 m {Very Large Telescope} (VLT) on
2000 April 25 and 2000 June 9, covering the spectral regions
$\lambda\lambda$4800--5800 \AA\ and $\lambda\lambda$5800--6800 \AA\ at
resolving power $\lambda/\delta\lambda\sim43,000$. Short exposures (718 s)
were chosen in order to minimize possible smearing of spectral lines
associated with the radial velocity change during its orbital motion, which we
estimated to be less than 11 {${\rm km}\:{\rm s}^{-1}$}.
\begin{deluxetable}{ccc}
\tabletypesize{\scriptsize}
\tablecaption{Ranges and steps of model parameters \label{tbl-1}}
\tablewidth{0pt}
\tablehead{
\colhead{Parameter} & \colhead{Range} & \colhead{Step}
}
\startdata
{$T_{\mathrm{eff}}$} & {4000 $\to$ 5500 K} & 100 K\\
{$\log g$} & {2.5 $\to$ 5} & 0.1 \\
{[Fe/H]} & {$-1.$ $\to$ 1} & 0.05 \\
{$f_{4500}$} & {0 $\to$ 2 } & 0.10 \\
{$m_0$} & {0 $\to$ $-$0.00082} &$-$0.00002\\
\enddata
\end{deluxetable}
The spectra were reduced in a standard manner using the UVES reduction
package within the MIDAS environment. The radial velocity for each spectrum
was obtained from the ephemeris reported in Casares et al.\ (2005, in
preparation). The individual spectra were corrected for radial velocity
and combined in order to improve the signal-to-noise ratio. After binning in
wavelength in steps of 0.21 m{\AA} the final spectrum had a signal-to-noise
ratio of 110 in the continuum. This spectrum is shown in Figure 1.
\begin{figure*}[ht!]
\centering
\includegraphics[width=13cm,angle=0]{f2bw.eps}
\caption{Distributions obtained for each parameter using
Monte Carlo simulations. The labels at the top of each bin indicate the
number of simulations consistent with the bin value. The total number
of simulations was 1000.\label{fig2}}
\end{figure*}
\begin{figure*}[ht!]
\centering
\includegraphics[width=11cm,angle=90]{f3bw.eps}
\caption{Best synthetic spectral fits to the UVES spectrum of the
secondary star in the \mbox{Cen X-4} system (bottom panel) and the same
for a template star (properly broadened) shown for comparison (top panel).
Synthetic spectra are computed for solar abundances (dashed line) and best
fit abundance (solid line).\label{fig3}}
\end{figure*}
\begin{figure*}[ht!]
\centering
\includegraphics[width=11cm,angle=90]{f4bw.eps}
\caption{The same as in Figure 3. The spectrum of the template is not
corrected for telluric lines (atm.\ H$_2$O); therefore, these lines
appear broadened in the spectrum (histogram line) and the synthetic
spectrum does not fit (solid line).\label{fig4}}
\end{figure*}
\begin{figure*}[ht!]
\centering
\includegraphics[width=11cm,angle=90]{f5bw.eps}
\caption{The same as in Figure 3.\label{fig5}}
\end{figure*}
\section{Chemical Analysis}
\subsection{Stellar Parameters}
The contribution of the emission from the accretion disk to the total
flux in \mbox{Cen X-4} has been estimated at $\sim$ 25--30\% and 40--50\% in
the {$V$} and {$B$} bands, respectively, whereas it is at least $80$\% at
$\lambda 3800$ \AA (Chevalier et al.\ 1989; Torres et al.\ 2002). As in
Gonz\'alez Hern\'andez et al.\ (2004), we have obtained the veiling,
together with the stellar atmospheric parameters, using synthetic
spectral fits to the high resolution spectrum of the secondary star in
\mbox{Cen X-4}. First, moderately strong and relatively unblended lines of
several elements of interest were identified in the high resolution solar
flux atlas of Kurucz et al.\ (1984). We selected several spectral features
containing in total 29 absorption lines of Fe~{\scshape i} with excitation
potentials between 0.5 and 4.5 eV. In order to compute synthetic spectra for
these features, we adopted the atomic line data from the Vienna Atomic Line
Database (VALD, Piskunov et al. 1995) and used a grid of local thermodynamic
equilibrium (LTE) models of atmospheres provided by Kurucz (1992, private
communication). Synthetic spectra were then computed \mbox{using} the LTE
code MOOG (Sneden 1973). To mi\-ni\-mi\-ze the effects associated with the
errors in the transition probabilities of atomic lines, we adjusted the
oscillator strengths, $\log gf$ values of the selected lines until we
succeeded in re\-pro\-du\-cing the solar atlas of Kurucz et al.\ (1984) with
solar abundances (Anders \& Grevesse 1989).
\begin{deluxetable}{lcccccc}
\tabletypesize{\scriptsize}
\tablecaption{Uncertainties in the abundances of the secondary star in
\mbox{Cen X-4}\label{tbl-2}}
\tablewidth{0pt}
\tablehead{\colhead{Element} & $\mathrm{[E/H]}_{\rm LTE}$ &
\colhead{$\Delta_{\sigma}$}
& \colhead{$\Delta_{T_{\mathrm{eff}}}$} & \colhead{$\Delta_{\log g}$}
& \colhead{$\Delta_{\rm veiling}$} &\colhead{$\Delta_{\rm
TOTAL}$\tablenotemark{\dagger}}}
\startdata
Al & 0.30 & 0.12 & 0.05 & 0.05 & 0.10 & 0.17 \\
Ca & 0.21 & 0.07 & 0.10 & -0.12 & 0.02 & 0.17 \\
Ti & 0.40 & 0.06 & 0.13 & 0.04 & 0.09 & 0.17 \\
Fe & 0.23 & 0.09 & 0.02 & 0.02 & 0.03 & 0.10 \\
Ni & 0.35 & 0.06 & 0.06 & 0.02 & 0.05 & 0.10 \\
Li\tablenotemark{\star} & 3.06 & 0.12 & 0.10 & 0.02 & 0.25 & 0.29 \\
\enddata
\tablenotetext{\star}{\mbox{Li} abundance is expressed as: \\$$\log
\epsilon(\mathrm{Li})_{\rm LTE} =
\log [N(\mathrm{Li})/N(\mathrm{H})]_{\rm LTE} + 12$$}
\tablenotetext{\dagger}{The total error was calculated using the
following
formula: \\$$\Delta_{\rm TOTAL} = \sqrt{\Delta_{\sigma}^2 +
\Delta_{T_{\mathrm{eff}}}^2 + \Delta_{\log g}^2 + \Delta_{\rm
veiling}^2}$$}
\tablecomments{The errors from the dispersion of the best fits to
different
features, $\Delta_{\sigma}$, are estimated using the following formula:
$\Delta_{\sigma} =\sigma/\sqrt{N}$, where $\sigma$ is the standard
deviation of
the measurements. Total errors also take into account the uncertainties
associated with the stellar parameters and the veiling.}
\end{deluxetable}
We generated a grid of synthetic spectra in terms of five free parameters,
three to characterize the star atmospheric model (effective
tem\-pe\-ra\-tu\-re, $T_{\mathrm{eff}}$, surface gravity, $\log g$, and
metallicity, [Fe/H]), and two further parameters to take into account the
effect of the accretion disk emission on the stellar spectrum. This was
assumed to be a linear function of wavelength and thus characterized by
two parameters: veiling at 4500 {\AA}, $f_{4500} = F^{4500}_{\rm
disk}/F^{4500}_{\rm sec}$, and the slope, $m_0$. Note
that the total flux is defined as $F_{\rm total} = F_{\rm disk}+F_{\rm
sec}$, where $F_{\rm disk}$ and $F_{\rm sec}$ are the flux contributions of
the disk and the continuum of the secondary star, respectively. These five
parameters were changed according to the steps and ranges given in Table 1.
A rotational broadening of 44 {${\rm km}\:{\rm s}^{-1}$}, a limb-darkening
$\epsilon = 0.65$ and a fixed value for the microturbulence, $\xi = 2$
{${\rm km}\:{\rm s}^{-1}$}, was adopted.
The observed spectrum was compared with each of the 1\,500\,000 synthetic
spectra in the grid via a $\chi^2$ minimization procedure that provided
the best model fit. Using a {bootstrap Monte-Carlo method} we defined the
1$\sigma$ confidence regions for the five free parameters and established
the most likely values: $T_{\mathrm{eff}} = 4500 \pm 100$ K, $\log g = 3.9
\pm 0.3$, [Fe/H]$ = 0.4 \pm 0.15$, $f_{4500} = 1.85 \pm 0.10$, and $m_0 =
-0.00071 \pm 0.00003$. Confidence regions determined using 1000
realizations are shown in Figure 2. This spectroscopic determination of the
surface gravity is consistent with derived surface gravities from the mass
and radius of the secondary star that lie between $\log g \sim 3.6$--4.0.
\subsection{Stellar Abundances}
Using the derived stellar parameters we ana\-ly\-zed several spectral
regions where we have identified various lines of Fe, Ca, Al, Ti, Ni, and
Li, often blended with Fe lines. Each of these spectral regions was
carefully normalized using a late type star template (HD 216803) properly
broadened with the rotational profile used for the secondary star in
\mbox{Cen X-4} for comparison. We determined the abundances of these
elements by comparing the observed spectrum with a grid of synthetic
spectra through a $\chi^2$ minimization procedure. For these spectral
syntheses we modified the element abundances while the stellar parameters
and the suitable veiling factor for each spectral region were kept fixed
(for further details, see Gonz\'alez Hern\'andez et al.\ 2004).
In Figs\ 3, 4 and 5 we show several spectral regions where we can see the
best model synthesis in comparison with the synthesis using solar
abundances. We also analyzed the broadened spectrum of the template late
type star and found the abundances of all the elements under study to be
close to solar (see Bodaghee et al.\ 2003) except the Li abundance which was
estimated to be $\log \epsilon(\mathrm{Li}) \lesssim 0.7$.
Abundances for all the elements are listed in Table 2 and referred to the
solar values adopted from Anders \& Grevesse (1989). We also give
\mbox{errors}, $\Delta_\sigma$, estimated from the dispersion of the
elemental abundances obtained from the best fits to the various spectral
features. Errors associated with uncertainties in effective temperature,
$\Delta_{T_{\mathrm{eff}}}$, gravity, $\Delta{\log g}$, and veiling,
$\Delta_{\rm veiling}$, are also listed in Table 2.
\begin{figure*}[ht!]
\rotate
\centering
\includegraphics[width=13cm,angle=0]{f6bw.eps}
\caption{Abundances of the secondary star in \mbox{Cen X-4} (wide
crosses) in comparison with the abundances in G and K metal-rich dwarf
stars. Galactic trends of Ca, Ni, and Ti were taken from Bodaghee et al.\
(2003) while Al from Feltzing \& Gustafsson (1998). The size of the cross
indicates the size of error. The dashed-dotted lines indicate solar
abundance values.\label{fig6}}
\end{figure*}
The Li feature is particularly strong in our target resulting in an
abundance of $\log \epsilon(\mathrm{Li})_{\rm LTE}=3.06 \pm 0.29$. We have
also estimated the non-LTE abundance correction for this element, $$\Delta
\log \epsilon(\mathrm{Li})= \log \epsilon(\mathrm{Li})_{\rm NLTE}-\log
\epsilon(\mathrm{Li})_{\rm LTE}$$ from the theoretical LTE and non-LTE
curves of growth in Pavlenko \& Magazz\`u (1996). By using the derived
stellar parameters and LTE Li abundance, we obtain $\log
\epsilon(\mathrm{Li})_{\rm NLTE}= 2.98$ for the secondary star in this
system.
The LiI 6708\AA\ resonance doublet may also form in absorption in
the atmosphere of the accretion disk during outburst decay (Suleimanov \&
Rebolo 1998). Such lines of Li and possibly of other elements could arise
if the disk luminosity is roughly 0.1--0.005 the Eddington luminosity.
However, \mbox{Cen X-4} was observed at quiescence, when its X-ray
luminosity was close to 10$^{-6}$ the Eddington luminosity (Asai et al.\
1996). In addition, these lines would be extremely broadened as a
consequence of the rotation profile of the disk ($\sim500$--1000
{${\rm km}\:{\rm s}^{-1}$}). Therefore, we discard that absorption lines
formed in the accretion disk have any significant influence on the derived
chemical abundances of the secondary star.
In the following section, we will discuss these results in the framework
of the origin and evolutionary scenario of the \mbox{Cen X-4} system.
\begin{deluxetable}{lccccc}
\tabletypesize{\scriptsize}
\tablecaption{Element abundance ratios in the secondary star in
\mbox{Cen X-4} and in the comparison sample\label{tbl-3}}
\tablewidth{0pt}
\tablehead{\colhead{Element} & \colhead{$\mathrm{[E/Fe]}_{{\rm Cen
X}-4}$} &
\colhead{$\Delta^{\tablenotemark{\star}}_{{\rm [E/Fe],Cen X}-4}$} &
\colhead{$\mathrm{[E/Fe]}_{\rm stars}$} & \colhead{$\sigma_{\rm stars}$}
&
\colhead{$\Delta_{\sigma,{\rm stars}}$}}
\startdata
Al & 0.07 & 0.13 & 0.04 & 0.03 & 0.006 \\
Ca & $-$0.02 & 0.18 & $-$0.07 & 0.04 & 0.008 \\
Ti & 0.17 & 0.14 & $-$0.04 & 0.06 & 0.010 \\
Ni & 0.12 & 0.08 & $-$0.02 & 0.05 & 0.010 \\
\enddata
\tablenotetext{\star}{Errors in the element abundance ratios
($\mathrm{[E/Fe]}$) in the secondary star in Cen X-4.}
\tablecomments{$\mathrm{[E/Fe]}_{\rm stars}$ indicates the average values
calculated for stars with iron content in the range $-$0.13 to 0.33
corresponding to 1$\sigma$ in the $\mathrm{[Fe/H]}$ abundance of the
secondary star in Cen X-4. Ca, Ti and Ni for the comparison sample have
been taken from 29 stars in Bodaghee et al.\ (2003) while Al from 28
stars in Feltzing \& Gustafsson (1998). The uncertainty in the average value
of element abundance ratios in the comparison sample is obtained as
$\Delta_{\sigma,{\rm stars}} =\sigma_{\rm stars}/\sqrt{N}$ where
$\sigma_{\rm stars}$ is the standard deviation of the measurements and $N$,
the number of stars.}
\end{deluxetable}
\section{Discussion}
\subsection{Heavy Elements}
The abundances of {\it heavy} elements in the se\-con\-dary star in
\mbox{Cen X-4} are super solar. In Fig.\ 6 these element abundances relative
to iron are shown in comparison with the Galactic abundance trends of these
elements in the relevant metallicity range, taken from Feltzing \&
Gustafsson (1998) and Bodaghee et al. (2003). The error bars in the element
abundance ratios ($\mathrm{[E/Fe]}$, see Table 3) take into account how
individual element abundances depend on the various sources of uncertainty
when dealing with abundance ratios. The $\mathrm{[Ca/Fe]}$ and
$\mathrm{[Al/Fe]}$ ratios of the secondary are consistent with abundances of
stars with similar iron content, while Ni and Ti appear to be moderately
enhanced. In Table 3 we show the element abundance ratios in the secondary
star in \mbox{Cen X-4} and average values in stars with iron content in the
range $-0.13 < \mathrm{[Fe/H]} < 0.33$, the comparison sample, corresponding
to a 1$\sigma$ uncertainty in the iron abundance of the companion star.
Whereas $\mathrm{[Ca/Fe]}$ and $\mathrm{[Al/Fe]}$ ratios are consistent with
average values, the $\mathrm{[Ni/Fe]}$ and $\mathrm{[Ti/Fe]}$ ratios are
1$\sigma$ higher than the average values of the stars in the comparison
sample.
It has been proposed that LMXBs such as \mbox{Cen X-4} begin their lives as
wide binaries and evolve through a common-envelope phase where the companion
spirals into the massive star's envelope (see, for example, Van den Heuvel
1983; de Kool et al.\ 1987; Nelemans \& Van den Heuvel 2001). The helium
core of the massive star continues its evolution until dying as a supernova
leaving a neutron star or a black hole remnant. Part of the ejected mass
in the SN explosion could be captured by the companion star, as has been
found in LMXBs like Nova Sco 94 (Israelian et al.\ 1999) and A0620$-$00
(Gonz\'alez Hern\'andez et al.\ 2004). We have considered this possibility
here in the case of the \mbox{Cen X-4} system.
\subsubsection{Spherical SN explosion}
A binary system like \mbox{Cen X-4} (with neutron star mass $M_{\rm
NS}=0.5-2.5$ {$M_\odot$} and the companion star mass in the range
$M_2=0.04$--0.58 {$M_\odot$}) will survive a spherical SN explosion
if the ejected mass, $\Delta M=M_{\rm He}-M_{\rm NS}\le (M_{\rm He}+M_2)/2$.
Therefore, initial masses for a neutron star and the secondary star of
$\sim1.4$ {$M_\odot$} and $\sim0.8$--1 {$M_\odot$} respectively would
imply a mass of the He core before the SN explosion $M_{\rm He}\le3.6$--3.8
{$M_\odot$}. Using the expressions given by Portegies Zwart et al.\ (1997,
and references therein), we inferred a mass of the progenitor star, $M_{\rm
1}\sim 16$ {$M_\odot$}, and a radius $R_{\rm He}\sim 0.8$ {$R_\odot$} for
the helium core. As in Gonz\'alez Hern\'andez et al.\ (2004), we can also
estimate the amount of the ejected material in a spherical explosion that
could be captured by the companion as coming from a central point since the
pre-SN orbital separation between the binary components would be $a_0 \sim
1.6$--2 {$R_\odot$} (assuming $M_{\rm He}\sim 4$ {$M_\odot$}, $M_{\rm
NS}\sim1.5$ {$M_\odot$} and $M_2\sim0.5$--0.8 {$M_\odot$}). The explosion of
such a massive primary would have taken place only $\sim 7\times10^6$ yr after the formation of the system (Brunish
\& Truran 1982). At that time, the radius of an 0.8 {$M_\odot\; $} secondary
star would be $\sim 1.2$ {$R_\odot\; $} (D'Antona \& Mazzitelli 1994). If we
consider a spherically symmetric supernova explosion, taking into account
the fraction of solid angle subtended by the companion and assuming
different capture efficiency factors, $f_{\rm capture}$, the amount of mass
deposited on the secondary can be computed as $$m_{\rm add}=\Delta M (\pi
R_2^2/4 \pi a_0^2)f_{\rm capture}$$
\begin{deluxetable}{lccccccccc}
\centering
\tabletypesize{\scriptsize}
%\rotate
\tablecolumns{8}
\tablecaption{Spherical Supernova Explosion in \mbox{Cen
X-4}\label{tbl-4}}
\tablewidth{0pc}
\tablehead{ & & & & \multicolumn{4}{c}{${\rm [E/H] \:
EXPECTED}$\tablenotemark{d}}\\
\cline{5-8}\\
& & & & \multicolumn{2}{c}{$M_2=0.5$ {$M_\odot$}} &
\multicolumn{2}{c}{$M_2=0.8$ {$M_\odot$}} \\
\cline{5-6} \cline{7-8} \\
& & & & \multicolumn{2}{c}{$M_{\rm cut}${$(M_\odot)$}} &
\multicolumn{2}{c}{$M_{\rm cut}${$(M_\odot)$}} \\
\cline{5-6} \cline{7-8} \\
ELEMENT & ${\rm [E/H]\:\rm OBS.}$\tablenotemark{a} & ${\rm [E/H]\:\rm
SAMPLE}$\tablenotemark{b} & ${\rm [E/H]}_{\dagger,i}$\tablenotemark{c} &
1.49 & 1.55 & 1.49 & 1.55
}
\startdata
\multicolumn{8}{c}{${\rm [Fe/H]}_{\dagger,i} = 0$} \\
\noalign{\smallskip}
\tableline
\noalign{\smallskip}
Al & 0.30 & 0.25 & 0 & 0.27 & 0.47 & 0.27 & 0.48 \\
Ca & 0.21 & 0.15 & 0 & 0.23 & 0.37 & 0.23 & 0.38 \\
Ti & 0.40 & 0.18 & 0 & 0.48 & 0.19 & 0.49 & 0.20 \\
Fe & 0.23 & 0.23 & 0 & 0.23 & 0.23 & 0.23 & 0.23 \\
Ni & 0.35 & 0.20 & 0 & 0.32 & 0.18 & 0.32 & 0.19 \\
\noalign{\smallskip}
\tableline
\noalign{\smallskip}
\multicolumn{8}{c}{${\rm [Fe/H]}_{\dagger,i}=-0.4$} \\
\noalign{\smallskip}
\tableline
\noalign{\smallskip}
Al & 0.30 & 0.25 & \nodata &\nodata&\nodata&\nodata&\nodata \\
Ca & 0.21 & 0.15 & $-$0.24 & 0.26 & 0.49 & 0.27 & 0.49 \\
Ti & 0.40 & 0.18 & $-$0.11 & 0.65 & 0.26 & 0.66 & 0.26 \\
Fe & 0.23 & 0.23 & $-$0.38 & 0.23 & 0.23 & 0.23 & 0.23 \\
Ni & 0.35 & 0.20 & $-$0.31 & 0.39 & 0.17 & 0.39 & 0.17 \\
\noalign{\smallskip}
\tableline
\noalign{\smallskip}
\multicolumn{8}{c}{${\rm [Fe/H]}_{\dagger,i}=-0.8$} \\
\noalign{\smallskip}
\tableline
\noalign{\smallskip}
Al & 0.30 & 0.25 & \nodata &\nodata&\nodata&\nodata&\nodata \\
Ca & 0.21 & 0.15 & $-$0.71 & 0.24 & 0.50 & 0.23 & 0.51 \\
Ti & 0.40 & 0.18 & $-$0.63 & 0.67 & 0.17 & 0.67 & 0.17 \\
Fe & 0.23 & 0.23 & $-$0.78 & 0.23 & 0.23 & 0.23 & 0.23 \\
Ni & 0.35 & 0.20 & $-$0.65 & 0.41 & 0.14 & 0.41 & 0.15 \\
\noalign{\smallskip}
\tableline
\noalign{\smallskip}
\multicolumn{8}{c}{${\rm [Fe/H]}_{\dagger,i}=-1.4$} \\
\noalign{\smallskip}
\tableline
\noalign{\smallskip}
Al & 0.30 & 0.25 & \nodata &\nodata&\nodata&\nodata&\nodata \\
Ca & 0.21 & 0.15 & $-$1.17 & 0.23 & 0.52 & 0.52 & 0.52 \\
Ti & 0.40 & 0.18 & $-$1.16 & 0.69 & 0.15 & 0.15 & 0.16 \\
Fe & 0.23 & 0.23 & $-$1.41 & 0.23 & 0.23 & 0.23 & 0.23 \\
Ni & 0.35 & 0.20 & $-$1.12 & 0.42 & 0.13 & 0.13 & 0.13 \\
\enddata
\tablenotetext{a}{Observed abundances of the
secondary star in Cen X-4.}
\tablenotetext{b}{Average abundances in stars of
the comparison sample (see also Table 4).}
\tablenotetext{c}{Initial abundances assumed for
the secondary star in \mbox{Cen X-4} from Allende Prieto et al.\ (2004)
and Stephens \& Boesgaard (2002)}
\tablenotetext{d}{Expected abundances of the secondary star}
\tablecomments{Expected abundances in the secondary atmosphere
contaminated with nuclesynthetic products of a 16 {$M_\odot$}
\emph{spherically} symmetric core-collapse explosion model ($M_{\rm He} \sim
4\ M_\odot$) for two different secondary masses, $M_2$. $M_{\rm cut}$ is the
mass cut assumed for each model. ${\rm [Fe/H]}_{\dagger,i}$ is the initial
metallicity of the secondary star. The capture efficiency, $f_{\rm
capture}$, has been modified in each model until matching the expected with
the observed iron abundance. Therefore, the mass captured by the companion,
$m_{\rm add}$, lies between 0.02 ($f_{\rm capture}=0.1$) and 0.11 ($f_{\rm
capture}=0.5$) for the extreme values of $M_2=0.5$ {$M_\odot$}, $M_{\rm
cut}=1.49$ {$M_\odot$} and ${\rm [Fe/H]}_{\dagger,i}=0$, and $M_2=0.8$
{$M_\odot$}, $M_{\rm cut}=1.55$ {$M_\odot$} and ${\rm
[Fe/H]}_{\dagger,i}=-1.4$}
\end{deluxetable}
In Table 4 we present the expected abundances in the atmosphere of the
secondary star after contamination from the progenitor of the compact
object for two plausible secondary masses (0.8 {$M_\odot$} and 0.5
{$M_\odot$}) at the time of the explosion of the primary. We have used a
spherically symmetric core-collapse explosion model of He core of mass
$M_{\rm He}\sim4$ {$M_\odot$} and explosion energy of $\varepsilon =
10^{51}\:{\rm ergs}$. This energy is deposited promptly in the central
region of the progenitor core to generate a strong shock wave. The
subsequent propagation of the shock wave is followed through a
two-dimensional hydrodynamical code, as has been done for aspherical
explosion models by Maeda et al.\ (2002). Nucleosynthesis is solved by
detailed nuclear reaction network calculations including 222 isotopes,
from $^{1}$H up to $^{71}$Ge. In the simulations, the directional
distribution of the input energy at the mass cut is given in a
parametric way to allow asymmetric explosions. One half of the energy is
deposited as kinetic energy, and the corresponding initial velocities
($v_{r}$, $v_{z}$) are imposed as $v_{z} = \alpha z$ and $v_{r} = \beta r$
(here $r$ is the equatorial direction, $z$ the jet direction, and $v_r$ and
$v_z$ the initially imposed velocity components along each direction).
Therefore, the initially imposed velocity is a factor of $\alpha/\beta$
larger in the jet direction than in the equatorial direction. In the
spherical symmetric model, the ratio $\alpha/\beta$ is set to be unity,
therefore the model is similar to canonical 1D explosion calculations (see,
e.g., Umeda et al. 2000). The details of our explosion calculations can be
found in Maeda et al. (2002). Our present models are basically the same as
their models, except for the progenitor's mass and the explosion energy.
In our modeling, we have chosen different capture efficiency
factors, $f_{\rm capture}$, between 0 and 1. However, simulations of type Ia
SNe suggest that the SN blast wave may induce mass-loss in the secondary
instead of matter accretion (Marietta et al. 2000) and therefore, high
values for the capture efficiency may not be appropriated. We also assume that the
matter captured is well-mixed in the convective envelope of the companion
star. The mass of the convective zone of the secondary star was fixed at
$m_{\rm conv} = 0.5$ and $0.6$ {$M_\odot\; $} from the evolutionary tracks
for a 0.5 and 0.8 {$M_\odot\; $} secondary star, respectively, with an age
of $\sim 7\times10^6$ yr (D'Antona \& Mazzitelli 1994).
The mass cut, $M_{\rm cut}$, that divides the ejecta from the collapsing
core, is low enough for a significant fraction of iron to be ejected
and consequently captured by the secondary star. The secondary star could
therefore have had an initial iron abundance lower than its current observed
abundance. In order to check this possibility we have assumed different
initial metallicities for the secondary and primary star in the system,
including low metallicities (given the high galactic latitude,
$b\sim24^{\circ}$, of the system that might suggest a halo origin). We have
taken the initial element abundances from the galactic trends (Allende
Prieto et al.\ 2004; Stephens \& Boesgaard 2002) for four given iron
abundances. We have used nucleosynthetic products of the solar metallicity
explosion model since the abundance pattern of $\alpha$-nuclei can be
similar to the solar abundance (Umeda et al.\ 2000). However, the odd $Z$ to
even $Z$ element ratios become smaller for lower metallicity, and hence we
have excluded $^{27}$Al from low metallicity models.
The expected abundances for the solar abundance model presented in Table
4 can fit very well with the observed abundances in the secondary star for
lower mass cuts. However, these element abundances are strongly sensitive
to the mass cut, and therefore for a slighlty higher value of this parameter
the expected abundances cannot reproduce the observed abundances. This
strong dependence on the mass cut makes the low metallicity models unable to
produce acceptable fits to the observed abundances. In addition, note that
for lower metallicity models the secondary star needs to capture 10--20 per
cent of the mass in its convective zone from the ejecta (i.e., an
$f_{\rm capture}\sim 0.2$--0.7) to achieve the current observed abundances.
Whereas for the solar abundance model by capturing roughly 5\% of its
convective envelope mass the secondary star can increase sufficiently its
element abundances (i.e., an $f_{\rm capture}\sim
0.1$--0.3). According to the chemical abundance pattern in the
secondary star, we find as very unlikely a halo origin. Although, strictly
speaking, we may only rule out an origin in the metal-weak halo. The halo
is also populated by stars that have been accreted from Milky Way satellites
and some of these are of high metallicity. The only known satellites that
are accreting onto the Milky Way and comprise a population as metal-rich as
the Cen X-4 system are Sagittarius (Ibata et al.\ 1994, Bonifacio et al.\
2004) and Canis Major (Martin et al.\ 2004, Sbordone et al.\ 2004).
Sagittarius is in a polar orbit (Ibata et al. 1997) and Canis Major in an
equatorial orbit (Martin et al. 2004). The present orbit of Cen X-4 (see
Gonz\'alez Hern\'andez et al. 2005) is not compatible with that of either of
these galaxies, thus it seems likely that Cen X-4 originated in neither of
them, but a detailed kinematic study is needed to reject this possibility.
\subsubsection{Aspherical SN explosion}
So far we have been considering a spherically symmetric SN explosion, but
the kinematic properties of this system might suggest an aspherical SN
explosion, as has been proposed for Nova Sco 94 (Brown et al.\ 2000). The
maximum ejected mass in a spherical SN explosion, $\Delta M \sim 2.4$
{$M_\odot$}, leads to a system velocity of $\sim$139 ${\rm km}\:{\rm
s}^{-1}$. Gonz\'alez Hern\'andez et al.\ (2005) measured the proper motion
of the system and derived a space velocity of $v_{\rm sys}\gtrsim 300$
${\rm km}\:{\rm s}^{-1}$, i.e., considerably larger than what could be
acquired in the case of a spherical SN explosion. In addition, the Galactic
space velocity components ($U, V, W$) of the system are significantly
different from the mean values that characterize the kinematics of stars
belonging to the thin and the thick disks of the Galaxy. We conclude that
the nascent neutron star must have received an additional \emph{kick}, most
probably at birth. Theorists have been trying to explain these kicks, whose
physical origin remains unclear (see Lai et al.\ 2001 for a review, and
references therein). A \emph{natal} kick could be generated by an asymmetric
explosion due to global hydrodynamic perturbations in the supernova core, or
it could be a result of asymmetric neutrino emission in the presence of
superstrong magnetic fields ($B\gtrsim 10^{15}\: {\rm G}$) in the
proto-neutron star and/or non-standard $\nu$ physics. In the latter cases,
an asymmetry of $\sim$1\% in the neutrino emission is sufficient to account
for a kick of 300 ${\rm km}\:{\rm s}^{-1}$; therefore, a non-spherically
symmetric SN explosion is not necessary. Recently, some authors have started
to explore the asymmetric-collapse kick mechanisms from global hydrodynamic
pertubations. Two and three dimensional studies predict strong neutron star
kicks from this mechanism (Burrows \& Hayes 1996; Janka et al.\ 2004) even
in excess of 1000 km/s. On the contrary, Fryer (2004), on the basis of three
dimensional supernova simulations, found that even the most extreme
asymmetric collapses do not produced final neutron star velocities above 200
${\rm km}\:{\rm s}^{-1}$. Therefore, an aspherical SN explosion cannot be
ruled out and might be the explanation for the high space velocity of the
system.
\begin{deluxetable}{lccccccccc}
\centering
\tabletypesize{\scriptsize}
%\rotate
\tablecolumns{8}
\tablecaption{Aspherical Supernova Explosion in \mbox{Cen
X-4}\label{tbl-5}}
\tablewidth{0pc}
\tablehead{ & & & & \multicolumn{4}{c}{${\rm [E/H] \:
EXPECTED}$\tablenotemark{d}}\\
\cline{5-8}\\
& & & & \multicolumn{2}{c}{$M_2=0.5$ {$M_\odot$}} &
\multicolumn{2}{c}{$M_2=0.8$ {$M_\odot$}} \\
\cline{5-6} \cline{7-8} \\
& & & & \multicolumn{2}{c}{$M_{\rm cut}${$(M_\odot)$}} &
\multicolumn{2}{c}{$M_{\rm cut}${$(M_\odot)$}} \\
\cline{5-6} \cline{7-8} \\
ELEMENT & ${\rm [E/H]\:\rm OBS.}$\tablenotemark{a} & ${\rm [E/H]\:\rm
SAMPLE}$\tablenotemark{b} & ${\rm [E/H]}_{\dagger,i}$\tablenotemark{c} &
1.49
& 1.55 & 1.49 & 1.55}
\startdata
\multicolumn{8}{c}{${\rm [Fe/H]}_{\dagger,i} = 0$} \\
\noalign{\smallskip}
\tableline
\noalign{\smallskip}
Al & 0.30 & 0.25 & 0 & 0.59 & 0.84 & 0.59 & 0.84 \\
Ca & 0.21 & 0.15 & 0 & 0.43 & 0.24 & 0.43 & 0.24 \\
Ti & 0.40 & 0.18 & 0 & 0.21 & 0.20 & 0.21 & 0.20 \\
Fe & 0.23 & 0.23 & 0 & 0.23 & 0.23 & 0.23 & 0.23 \\
Ni & 0.35 & 0.20 & 0 & 0.18 & 0.22 & 0.18 & 0.22 \\
\noalign{\smallskip}
\tableline
\noalign{\smallskip}
\multicolumn{8}{c}{${\rm [Fe/H]}_{\dagger,i}=-0.4$} \\
\noalign{\smallskip}
\tableline
\noalign{\smallskip}
Al & 0.30 & 0.25 & \nodata &\nodata&\nodata&\nodata&\nodata \\
Ca & 0.21 & 0.15 & $-$0.24 & 0.56 & 0.28 & 0.56 & 0.28 \\
Ti & 0.40 & 0.18 & $-$0.11 & 0.28 & 0.26 & 0.28 & 0.26 \\
Fe & 0.23 & 0.23 & $-$0.38 & 0.23 & 0.23 & 0.23 & 0.23 \\
Ni & 0.35 & 0.20 & $-$0.31 & 0.15 & 0.22 & 0.15 & 0.22 \\
\noalign{\smallskip}
\tableline
\noalign{\smallskip}
\multicolumn{8}{c}{${\rm [Fe/H]}_{\dagger,i}=-0.8$} \\
\noalign{\smallskip}
\tableline
\noalign{\smallskip}
Al & 0.30 & 0.25 & \nodata &\nodata&\nodata&\nodata&\nodata \\
Ca & 0.21 & 0.15 & $-$0.71 & 0.59 & 0.25 & 0.59 & 0.25 \\
Ti & 0.40 & 0.18 & $-$0.63 & 0.20 & 0.17 & 0.20 & 0.17 \\
Fe & 0.23 & 0.23 & $-$0.78 & 0.23 & 0.23 & 0.23 & 0.23 \\
Ni & 0.35 & 0.20 & $-$0.65 & 0.12 & 0.21 & 0.12 & 0.21 \\
\noalign{\smallskip}
\tableline
\noalign{\smallskip}
\multicolumn{8}{c}{${\rm [Fe/H]}_{\dagger,i}=-1.4$} \\
\noalign{\smallskip}
\tableline
\noalign{\smallskip}
Al & 0.30 & 0.25 & \nodata &\nodata&\nodata&\nodata&\nodata \\
Ca & 0.21 & 0.15 & $-$1.17 & 0.61 & 0.25 & 0.61 & 0.25 \\
Ti & 0.40 & 0.18 & $-$1.16 & 0.19 & 0.15 & 0.19 & 0.15 \\
Fe & 0.23 & 0.23 & $-$1.41 & 0.23 & 0.23 & 0.23 & 0.23 \\
Ni & 0.35 & 0.20 & $-$1.12 & 0.11 & 0.20 & 0.11 & 0.20 \\
\enddata
\tablenotetext{a}{Observed abundances of the
secondary star in Cen X-4.}
\tablenotetext{b}{Average abundances in stars of
the comparison sample (see also Table 4).}
\tablenotetext{c}{Initial abundances assumed for the secondary star in
\mbox{Cen X-4} from Allende Prieto et al.\ (2004) and Stephens \&
Boesgaard (2002)}
\tablenotetext{d}{Expected abundances of the secondary star}
\tablecomments{Expected abundances in the secondary atmosphere
contaminated with nuclesynthetic products of a 16 {$M_\odot$}
\emph{non-spherically} symmetric core-collapse explosion model ($M_{\rm He}
\sim 4\ M_\odot$) for two different secondary masses, $M_2$. $M_{\rm cut}$
is the mass cut assumed for each model. ${\rm [Fe/H]}_{\dagger,i}$ is the
initial metallicity of the secondary star. The capture efficiency, $f_{\rm
capture}$, has been modified in each model until matching that expected with
the observed iron abundance. Therefore, the mass captured by the companion,
$m_{\rm add}$, lies between 0.04 ($f_{\rm capture}=0.2$) and 0.22 ($f_{\rm
capture}=0.8$) for the extreme values of $M_2=0.5$ {$M_\odot$}, $M_{\rm
cut}=1.49$ {$M_\odot$} and ${\rm [Fe/H]}_{\dagger,i}=0$, and $M_2=0.8$
{$M_\odot$}, $M_{\rm cut}=1.55$ {$M_\odot$} and ${\rm
[Fe/H]}_{\dagger,i}=-1.4$}
\end{deluxetable}
Nucleosynthesis in aspherical explosions has been examined by Maeda et
al.\ (2002), who showed that the chemical composition of the ejecta is
strongly dependent on direction. In particular, the main effect of the
asphericity is that elements produced by the strong $\alpha$-rich
freezeout are greatly enchanced relative to iron (e.g., [Ti/Fe]). This
effect is more pronounced in the jet direction, where the kinetic energy is
higher. Therefore, assuming that the jet is perpendicular to the orbital
plane, i.e., that the secondary star is located in the equatorial plane of
the helium star, elements like Ti, Ni, and Fe are mainly ejected in the jet
direction, while Al, O, Si, S, and Mg are preferentially ejected near the
equatorial plane. Consequently, we should have found a high Al abundance in
the companion star in comparison with the other elements studied but this is
not the case. To compare these effects with our observations of \mbox{Cen
X-4} quantitatively, we compute an aspherical explosion of a $4 M_{\odot}$
He core with the explosion energy $10^{51}$ ergs. The asphericity is induced
by the parametric way (see \S 4.1.1) with the asphericity parameter
$\alpha/\beta = 2$, i.e., the initial shock velocity at the mass cut is 2
times larger along the jet direction than in the equatorial direction. In
table 5 we show the expected element abundances in the secondary star after
pollution from the ejecta of an aspherical SN explosion of a 4 {$M_\odot$}
He core. Contrary to the spherical case, Al is dramatically enhanced in all
the model computations and this effect is even stronger for higher mass
cuts. Therefore, for the solar abundance model, none of the model
computations produces acceptable fits, even for lower mass cuts. Note that
explosive burning in aspherical SN is weak, ejecting little Ti and Ni, in
the equatorial direction. Thus, the mass cut appropiate to the system, i.e.,
$\sim 1.4$ {$M_\odot$}, is not deep enough to eject these elements to
explain the large values of [(Ti, Ni)/H] observed. The mass cut in the
aspherical case is placed at relatively weak burning region as compared with
the spherical model. Therefore, the dependence of the model abundance on
changing the mass cut is different from the spherical model, which also
makes the aspherical model difficult to fit the observed abundances.
Note that all the above results of the aspherical explosion rely on the
assumption that the jet direction is perpendicular to the orbital plane,
therefore to the equatorial direction of the progenitor star since the star
may well co-rotate with the binary orbit. The assumption that the jet is
almost perpendicular to the equatorial direction is justified because recent
core-collapse theories suggest the jets, if formed, propagate along the
progenitor's rotational axis (e.g., Fryer \& Warren 2004 and references
therein). This is also supported by a {\it Hubble Space Telescope} image of
SN 1987A (Wang et al. 2002). The ejecta of SN 1987A is elongated toward the
direction perpendicular to the innermost circumstellar ring, which is
probably along the equatorial direction (e.g., Chevalier \& Dwarkadas
1995).
In the case of low initial metallicity of the secondary star, higher
values for the capture efficiency (i.e., $f_{\rm capture}\sim 0.3$--1) are
required to achieve the observed Fe abundance since this element is mainly
ejected perpendicularly to the orbital plane. Despite this requirement,
these simulations cannot fit the observed abundances even for high mass
cuts.
In summary, a spherically symmetric SN explosion of a 4 {$M_\odot$} He core
leaving an NS gives a better fit than an aspherical model. A relatively
low primary star ($M_1 \lesssim 20$ {$M_\odot$}) seems to leave an NS and to
explode as a normal SN (with an energy of $\varepsilon = 10^{51}\:{\rm
ergs}$, e.g., Nomoto et al.\ 2003). Our result would suggest that a normal
SN would be less aspherical than a hypernova, as proposed for the black hole
in the LMXB Nova Sco 94 (Brown et al.\ 2000). However, we cannot rule out
that the companion star did not capture a significant amount of matter from
the ejecta of a SN explosion, and therefore that the observed abundances
correspond to its primordial abundances.
\begin{figure}[ht!]
\centering
\includegraphics[width=7.5cm,angle=0]{f7bw.eps}
\caption{Li abundance of the secondary star in \mbox{Cen X-4} (filled
circle) in comparison with the abundances of different rotating Pleiades
dwarf stars versus effective temperature from Garc{\'\i }a L\'opez et
al.\ (1994). The sizes of the circles are related to $v \sin i$. Stellar
masses have been assigned following the evolu\-tio\-na\-ry tracks of
D'Antona \& Mazzitelli (1994) for $10^8$ yr.\label{fig7}}
\end{figure}
\subsection{Li abundance}
The companion star has an unexpectedly high $^{7}$Li abundance close to
the cosmic value in the Galactic disk (Mart\'\i{}n et al.\ 1994a),
substantially higher than field late-type main sequence stars of similar
mass. Convective mixing during pre-main sequence and main sequence evolution
is expected to produce significant lithium depletion in the atmospheres of
such stars (see Fig.\ 7), thus a possible explanation for the Li
over-abundance in the secondary is that the \mbox{Cen X-4} system is younger
than the Pleiades cluster whose age is $\sim 120$ Myr (Mart{\'\i}n et al.\
1998; Stauffer et al.\ 1998). One could estimate an upper limit to the age
of the system under the assumption that there is no mechanism able to enrich
the atmosphere of the secondary star with freshly synthesized Li nuclei. In
Fig.\ 8 we show the age of the system for different mass transfer rates and
initial secondary masses by following the Li depletion in low mass stars
according to evolutionary tracks of Siess et al.\ (2000). We have adopted
the NLTE Li abundance and the present value for the initial metallicity of
the system since a possible metallicity enrichment due to a SN explosion
would have taken place early enough (at $\sim$7 Myr) and it would make the
secondary follow the Li depletion as a metal-rich star. In any case, this
metallicity effect is only relevant for masses higher than 0.8 {$M_\odot$}.
Such high Li abundance restricts the age of the system to the range $\sim
10$--70 Myr. This age is quite short compared to the age assumed
in the proposed evolutionary scenarios for the system, where an age of a
few times 10$^9$ yr is required (e.g. Chevalier et al. 1989). These age
constrains are obtained using models that do not consider the impact of high
rotation on the Li depletion. According to Mart\'\i{}n \& Claret (1996) high
rotation may inhibit the rate of Li depletion; therefore, the constraints on
age should be strictly considered as a lower limit. The effect of rotation
on Li depletion has been measured in Pleiades stars with masses in the range
0.6--0.8 {$M_\odot$} (see, for example, Garc\'\i{}a L\'opez et al. 1994).
The kinematic properties of Cen X-4 suggest that it could have been at the
inner regions of the Galactic disk about 100--200 Myr ago (Gonz\'alez
Hern\'andez et al.\ 2005). If the neutron star formed at that time as a
result of an SN explosion, this might explain the high Li abundance of the
system without postulating a Li production mechanism.
For such a young age the effect of the mass transfer in the secondary's
mass is negligible, and initial secondary masses higher than 0.8
{$M_\odot$} can therefore be ruled out. Note that the mass transfer rate
inferred from the observed quiescent X-ray luminosity ($\sim 2.4 \times
10^{32}$ ${\rm erg}\:{\rm s}^{-1}$, Asai et al.\ 1996) is very low $\sim
10^{-6}$ the Eddington rate ($\sim 3 \times 10^{-9}$ $M_\odot\:{\rm
yr}^{-1}$, Yi \& Narayan 1997). On the other hand, Menou \& McClintock
(2001) have suggested a mass transfer rate of $\sim$$3 \times 10^{-10}$
$M_\odot\:{\rm yr}^{-1}$ based on optical--UV emission spectrum of the
system.
\begin{figure}[ht!]
\centering
\includegraphics[width=7.5cm,angle=0]{f8bw.eps}
\caption{Upper limit to the age of the \mbox{Cen X-4} system according
to the Li abundance of the secondary star versus initial stellar masses of
the secondary star, assuming different mass transfer rates. Li depletion
has been followed according to the evolu\-tio\-na\-ry tracks of Siess et
al.\ (2000).\label{fig8}}
\end{figure}
High Li abundances appear to be a common property of late-type
companions in Soft X-ray Transients (SXTs), characterized by strong X-ray
outbursts (ussually lasting for several months) followed by long quiescent
periods of $\sim$10--50 yr. Different models have been proposed to explain
these Li over-abundances in terms of spallation reactions or
$\alpha$--$\alpha$ reactions. Mart\'\i{}n et al.\ (1994b) proposed that Li
production could come from nuclear reactions in the accretion disk during
outburst and being deposited on the companion via a disk wind. They argued
that the observed abundance must correspond closely to the long term average
since the depletion time scale ($\tau_D \gtrsim 10^7$ yr) is much longer
than the recurrence time of outburst ($\sim 30$ yr). Later on, Yi \& Narayan
(1997) investigated the same scenario, in an advection-dominated accretion
flow (ADAF), rather than a Shakura--Sunyaev disk not only during outburst
but also in quiescence, although Li production during outburst may even
dominate over the production in quiescence. This model can explain the
observed Li abundance in \mbox{Cen X-4} only if the mass-accretion rate in
quiescence is $\sim 10^{-3}$ times the Eddington rate and if there is a
``propeller'' effect. More recently, Guessoum \& Kazanas (1999) have
proposed Li production via neutron spallation of CNO elements in the
secondary star. In this scenario, a flux of neutrons produced in the hot
accretion flow around the compact object could escape easily and impinge on
the secondary surface and then produce spallation of heavier elements. A
short exposure to the above flux, i.e., during outbursts, would suffice to
produce the observed enhancement.
A main sequence K star of mass $\sim$$0.1$--$0.8$ {$M_\odot$} would destroy
all its primordial Li abundance in $\lesssim 80$ Myr. Yi \& Narayan (1997,
see also references therein) have discussed in some detail the question of
Li depletion in the atmosphere of the companion star once it has been
deposited/produced there, considering the possibility of a steady state,
in which the enrichment and depletion rates should be equal, as suggested by
Mart\'\i{}n et al.\ (1994b). Mart\'\i{}n et al.\ (1994a) remarked that
isotopic ratios $^{7}$Li/$^{6}$Li as low as 5 are expected and hence would
provide a proof of the spallation scenario. Unfortunately, at such high Li
abundances there is degeneracy between Li abundance and isotopic ratio and
consequently, different Li abundances and different isotopic ratios give
rise to the same line strength. Higher S/N spectra are required to extract
this information from the 6708 \AA\ Li line although the high rotational
velocity of the star might make it very difficult. In addition, Guessoum \&
Kazanas (1999) suggested the possible enhancement of other light elements
such as Be and B that might be analyzed using UV transitions.
\section{Conclusions}
We have obtained a high quality spectrum of the secondary star in
\mbox{Cen X-4} and derived atmospheric parameters and chemical abundances of
several elements. We have applied a technique that provides a determination
of the stellar parameters taking into \mbox{account} any possible veiling
from the accretion disk. We find $T_{\mathrm{eff}} = 4500 \pm
100$ K, $\log g = 3.9 \pm 0.3$, and a disk veiling (defined as $F_{\rm
disk}/F_{\rm total}$) of less than 55 per cent at 5000 {\AA} and
decreasing towards longer wavelengths.
The abundances of Fe, Ca, Ti, Al, and Ni are super solar. The abundance
ratios of each element with respect to Fe were compared with these
ratios in late-type main sequence metal-rich stars. Moderate anomalies
for Ni and Ti have been found. A comparison with element yields from
spherically symmetric supernova explosion models suggests that the secondary
star could have captured part of the ejecta from a supernova that also
originated the compact object in \mbox{Cen X-4} if the primordial abundances
of the secondary star were solar. The observed abundances can be explained
if a progenitor with a $\sim 4$ {$M_\odot\; $} helium core exploded with a
mass cut $\sim 1.5$ {$M_\odot$}, such that a significant amount of iron
could escape from the collapse of the inner layers, and hence the ejecta
could also have enhanced the iron abundance in the secondary star.
The high galactic latitude of the system might suggest that the system
could have originated in the Galactic halo and later enriched by
nucleosynthetic products ejected in the supernova but our model computations
for low metallicities show that this possibility is unlikely. In addition,
the kinematic properties of this system (Gonz\'alez Hern\'andez et al.\
2005) suggest that a natal kick was imparted to the neutron star at birth
due to an asymmetric mass ejection and/or an asymmetry in the neutrino
emission. We have also inspected a non-spherically symmetric SN explosion
model but this model cannot produce acceptable fits to the observed
abundances in the secondary star.
The Li abundance in the secondary star in \mbox{Cen X-4} is dramatically
high in comparison with field late-type main sequence stars, possibly
indicating that this is a young system ($\sim$$10$--$80$ Myr).
Alternatively, if the system is much older a Li production mechanism is
needed to balance the large destruction of Li expected in such low-mass
stars.
\section{Acknowledgements}
This work has made use of the VALD database and IRAF facilities and has
been partially financed by the Spanish Ministry projects AYA2001-1657,
AYA2002-03570 and AYA2002-0036.
\begin{thebibliography}{}
\bibitem[Anders \& Grevesse 1989]{aag89}
Anders, E.,\& Grevesse, N. 1989, \gca, 53, 197
\bibitem[Asai et al 1996]{asa96}
Asai, K., Dotani, T., Mitsuda, K., Hoshi, R., Vaughan, B., Tanaka, Y., \&
Inoue, H. 1996, \pasj, 48, 257
\bibitem[Allende Prieto et al 2004]{all04}
Allende Prieto, C., Barklem, P. S., Lambert, D. L., \& Cunha, K. 2004,
\aap,
420, 183
\bibitem[Bodaghee et al. 2003]{bod03}
Bodaghee, A., Santos, N. C., Israelian, G., \& Mayor, M. 2003, \aap, 404,
715
\bibitem[Bonifacio et al.(2004)]{bon04}
Bonifacio, P., Sbordone, L., Marconi, G., Pasquini, L., \& Hill, V.\
2004,
\aap, 414, 503
\bibitem[Brown et al., 1999]{blb99}
Brown, G. E., Lee, C.-H., \& Bethe, H. A. 1999, NewA, 4, 313
\bibitem[Brunish \& Truran 1982]{bat82}
Brunish, W. M.,\& Truran, J. W. 1982, \apjs, 49, 447
\bibitem[Burrows \& Hades 1996]{bah96}
Burrows, A., \& Hayes, J. 1996, \prl, 76, 352
\bibitem[Canizares et al. 1980]{can80}
Canizares, C. R., McClintock, J. E., \& Grindlay, J. E. 1980, \apj, 236,
55
\bibitem[Chevalier et al. 1989]{che89}
Chevalier, C., Ilovaisky, S. A., van Paradijs, J., Pedersen, H., \& van
der Klis, M. 1989, \aap, 210, 114
\bibitem[Chevalier et al. 1995]{cad95}
Chevalier, R. A., \& Dwarkadas, V. V. 1995, \apj, 452, L45
\bibitem[Conner et al. 1969]{con69}
Conner, J. P., Evans, W. D., \& Belian, R. D. 1969, \apj, 157, 157
\bibitem[D'Antona \& Mazzitelli 1994]{dam94}
D'Antona, F.,\& Mazzitelli, I. 1994, \apjs, 90, 467
\bibitem[Feltzing \& Gustafsson 1998]{fag98}
Feltzing, S.,\& Gustafsson, B. 1998, \aaps, 129, 237
\bibitem[Fryer 2004]{fry04}
Fryer, C. L. 2004, \apj, 601, L175
\bibitem[Fryer \& Warren 2004]{faw04}
Fryer, C. L., \& Warren, M. S. 2004, \apj, 601, 391
\bibitem[Garc{\'\i }a L\'opez et al. 1994]{gar94}
Garc{\'\i }a L\'opez, A., Rebolo, R., \& Mart{\'\i }n, E. L. 1994, \aap,
282, 518
\bibitem[Gessoum \& Kazanas 1999]{ges99}
Gessoum, N., \& Kazanas, D. 1999, \apj, 512, 332
\bibitem[Gonz\'alez Hern\'andez et al. 2004]{gon04}
Gonz\'alez Hern\'andez, J. I., Rebolo, R., Israelian, G., Casares, J.,
Maeder, A., \& Meynet, G. 2004, \apj, 609, 988
\bibitem[Gonz\'alez Hern\'andez et al. 2005]{gon05}
Gonz\'alez Hern\'andez, J. I., Rebolo, R., Pe\~narrubia, J., Casares, J.,
\&
Israelian, G. 2005, submitted.
\bibitem[Ibata et al.(1997)]{iba97}
Ibata, R. A., Wyse, R. F. G., Gilmore, G., Irwin, M. J., \& Suntzeff, N.
B.
1997, \aj, 113, 634
\bibitem[Ibata, Gilmore, \& Irwin(1994)]{ibata94}
Ibata, R. A., Gilmore, G., \& Irwin, M. J. 1994, \nat, 370, 194
\bibitem[Israelian et al. 1999]{isr99}
Israelian, G., Rebolo, R., Basri, G., Casares, J., \& Mart{\'\i }n, E. L.
1999, \nat,
401, 142
\bibitem[Janka et al 2004]{jan04}
Janka, H.-Th., Scheck, L., Kifonidis, K., M\"uller, E., \& Plewa, T.
2004, astro-ph/0408439
\bibitem[Kaluzienski et al 1980]{kal80}
Kaluzienski, L. J., Holt, S. S., \& Swank, J. H. 1980, \apj, 241, 779
\bibitem[de Kool et al., 1987]{koo87}
de Kool, M., van den Heuvel, E. P. J., \& Pylyser, E. 1987, \aap, 183, 47
\bibitem[Kurucz et al., 1984]{kur84}
Kurucz, R. L., Furenild, I., Brault, J., \& Testerman, L. 1984, Solar
Flux Atlas from
296 to 1300 nm, NOAO Atlas 1 (Cambridge: Harvard Univ. Press)
\bibitem[Lai et al 2001]{lai01}
Lai, D., Chernoff, D. F., \& Cordes, J. M. 2001, \apj, 549, 1111
\bibitem[Maeda et al. 2002]{mae02}
Maeda, K., Nakamura, T., Nomoto, K., Mazzali, P. A., Patat, F., \&
Hachisu, I. 2002,
\apj, 565, 405
\bibitem[Marietta et al., 2000]{mat00}
Marietta, E., Burrows, A., \& Fryxell, B. 2000, \apjs, 128, 615
\bibitem[Martin et al.(2004)]{M2004}
Martin, N. F., Ibata, R. A., Bellazzini, M., Irwin, M. J., Lewis, G. F.,
\&
Dehnen, W. 2004, \mnras, 348, 12
\bibitem[Mart{\'\i }n et al. 1994a]{mar94a}
Mart{\'\i }n, E. L., Rebolo, R., Casares, J., \& Charles, P. A. 1994a,
\apj, 435,
791
\bibitem[Mart{\'\i }n et al. 1994b]{mar94b}
Mart{\'\i }n, E. L., Spruit, H. C., \& van Paradijs, J. 1994b, \aap, 291,
L43-L46
\bibitem[Mart{\'\i }n \& Claret 1996]{mac96}
Mart{\'\i }n, \& Claret, A, 1996, \aap, 306, 408
\bibitem[Mart{\'\i }n et al. 1998]{mar98}
Mart\'\i{}n, E. L., Basri, G., Gallegos, J. E., Rebolo, R.,
Zapatero-Osorio, M. R., \& Bejar, V. J. S. 1998, \apj, 499, 61
\bibitem[Matsuoka et al 1980]{mat80}
Matsuoka, M., Inoue, H., Koyama, K., Makishima, K., Murakami, T., Oda,
M., Ogawara, Y., Ohashi, T., Shibazaki, N., Tanaka, Y., \& 10 coauthors
1980, \apj, 240, 137
\bibitem[McClintock \& Remillard 1990]{mar90}
McClintock, J. E., \& Remillard, R. A. 1990, \apj, 350, 386
\bibitem[Menou \& McClintock 2001]{mam01}
Menou, K.,\& McClintock, J. E. 2001, \aap, 557, 304
\bibitem[Nelemans \& Van den Heuvel]{nah01}
Nelemans, G., \& Van den Heuvel, E. P. J. 2001, \aap, 376, 950
\bibitem[Nelemans \& Van den Heuvel]{nah01}
Nomoto, K, Maeda, K, Umeda, H, Ohkubo, T, Deng,
J, \& Mazzali, P. 2003, A Massive Star Odyssey: From Main Sequence to
Supernova, Proceedings of IAU Symposium \#212, held 24-28 June 2001 in
Lanzarote, Canary island, Spain. Edited by Karel van der Hucht, Artemio
Herrero, and Esteban, César. San Francisco: Astronomical Society of the
Pacific, p.395 (astro-ph/0209064)
\bibitem[van Paradijs et al. 1980]{par80}
van Paradijs, J., Verbunt, F., Van der Linden, T., Pedersen, H., \&
Wamsteker, W. 1980, \apj, 241, 161
\bibitem[Pavlenko \& Magazzu 1996]{pam96}
Pavlenko, Ya. V., \& Magazz\`u, A. 1996, \aap, 311, 961
\bibitem[Piskunov et al. 1995]{pis95}
Piskunov, N. E., Kupka, F., Ryabchikova, T. A., Weiss, W. W., \& Jeffery, C.
S. 1995, \aaps, 112, 525
\bibitem[Podsiadlowski et al., 2002]{pod02}
Podsiadlowski, P., Nomoto, K., Maeda, K., Nakamura, T., Mazzali, P., \&
Schmidt, B. 2002, \apj, 567, 491
\bibitem[Portegies Zwart et al. 1997]{poz97}
Portegies Zwart, S. M., Verbunt, F., \& Ergma, E. 1997, \aap, 321, 207
\bibitem[Sbordone et al.(2004)]{SB04}
Sbordone, L., Bonifacio, P., Marconi, G., Zaggia, S., \& Buonanno, R.
2004,
\aap Letters, in press (astro-ph/0412098)
\bibitem[Shahbaz et al. 1993]{sha93}
Shahbaz, T., Naylor, T., \& Charles, P. A. 1993, \mnras, 265, 655
\bibitem[Siess et al. 2000]{sie00}
Siess, L., Dufour, E., \& Forestini, M. 2000, \aap, 358, 593
\bibitem[Sneden 1973]{sne73}
Sneden, C. 1973, PhD Dissertation (Univ. of Texas at Austin)
\bibitem[Stauffer et al. 1998]{mar98}
Stauffer, J. R., Schultz, G., \& Kirkpatrick, J. D. 1998, \apj, 499, 199
\bibitem[Stephens \& Boesgaard 2002]{sab02}
Stephens, A., \& Boesgaard, A. M. 2002, \apj, 123, 1647
\bibitem[Suleimanov \& Rebolo 1998]{sar98}
Suleimanov, V., \& Rebolo, R. 1998, \aap, 330, 999
\bibitem[Torres et al. 2002]{tor02}
Torres, M. A. P., Casares, J., Mart{\'\i}nez-Pais, \& Charles, P. A.
2002,
\mnras, 334, 233
\bibitem[Umeda \& Nomoto 1999]{uan99}
Umeda, H., Nomoto, K., Nakamura, T. 2000, The First Stars. Proceedings of
the MPA/ESO Workshop held at Garching, Germany, 4-6 August 1999. Achim
Weiss, Tom G. Abel, Vanessa Hill (eds.). Springer. (astro-ph/9912248)
\bibitem[Van den Heuvel 1983]{heu83}
Van den Heuvel, E. P. J. 1983, in Accretion-driven Stellar X-ray Sources,
eds. W. H. G. Lewin \& E. P. J. Van den Heuvel, Cambridge University
Press,
p. 303
\bibitem[Wang et al 2002]{wan02}
Wang, L. et al. 2002, \apj, 579, 671
\bibitem[Yi \& Narayan 1997]{yan97}
Yi, I. \& Narayan, R. 1997, \apj, 486, 363
\end{thebibliography}
\end{document}