2 The history of Vela X-1

Vela X-1 was first detected by a rocket-borne experiment (Chodil et al. 1967) and subsequent observations (see, for example, Giacconi et al. 1972), suggested that the source was highly variable. Using data from the OSO-7 satellite, Ulmer et al. (1972) demonstrated evidence for intensity variations which were interpreted as eclipses with a period of around 9 days. An optical counterpart, GP Vel / HD 77581 (a B0.5 giant with m_v = 6.8) was identified by Brucato & Kristian (1972) and Hiltner et al. (1972), based on its ultra-violet excess and radial velocity variations.

An X-ray pulse period of 283 s was subsequently discovered using the SAS-3 satellite (Rappaport & McClintock 1975; McClintock et al. 1976). Timing observations of these pulses by Rappaport et al. (1976) allowed the radial velocity semi-amplitude of the X-ray component to be measured as $K_{\rm x} = 273 \pm 9$ km s^-1, and the eccentricity of the system, $e \sim 0.1$. Later work, using Hakucho and Tenma data (Deeter et al. 1987), gave a value of $a_{\rm x} \sin i = 113.0\pm0.4$ light seconds for the projected semi-major axis. Using values for the orbital period P and the eccentricity of the orbit e, the corresponding $K_{\rm x}$ value may be calculated according to

$$K_{\rm x} = \frac{2\pi}{P} \frac{a_{\rm x} \sin i}{(1-e^2)^{1/2}}$$ (9)

as $K_{\rm x} = 276 \pm 1$ km s^-1. The most recent, and accurate, values determined from X-ray pulse timing analysis are those obtained with BATSE on board CGRO reported by Bildsten et al. (1997). They obtain a value $a_{\rm x} \sin i = 113.89 \pm 0.13$ light seconds which corresponds to $K_{\rm x} = 278.1 \pm 0.3$ km s^-1 when combined with their accurate values for the orbital period, $P = 8.964368 \pm 0.000040$ d, and eccentricity, $e = 0.0898 \pm 0.0012$.

The X-ray eclipse duration appears to be quite variable, and somewhat energy dependent. For example, Forman et al. (1973) obtained a value for the half angle of the eclipse of $\theta_{\rm e} = 38^{\circ}\pm1^{\circ}$ using the Uhuru satellite and Charles et al. (1976) obtained a value of $\theta_{\rm e} = 39\hbox{$.\!\!^\circ$ }8\pm0\hbox{$.\!\!^\circ$ }4$ using the Copernicus satellite, whereas Watson & Griffiths (1977) quote a value of $\theta_{\rm e} = 33\hbox{$.\!\!^\circ$ }8\pm1\hbox{$.\!\!^\circ$ }3$ obtained using Ariel V data. However the earlier two experiments were at much softer energies than Watson & Griffiths observed, and softer X-rays are much more likely to be absorbed by circumstellar material, thus extending the observed eclipse time.

 

 

Table 1: Measured amplitudes of the radial velocity curve of GP Vel.

$K_{\rm o}$ / km s-1	Reference
37-45	Wallerstein (1974)
$26\pm0.7$	Zuiderwijk (1974)
$20\pm1$	van Paradijs et al. (1976)
$21.75\pm1.15$	van Paradijs et al. (1977)
$21.8\pm 1.2$	Rappaport & Joss (1983)
18.0-28.2	(95% conf. range) van Kerkwijk et al. (1995)
$17.8\pm1.6$	Stickland et al. (1997)
22	(Correction to the above) Barziv et al. (2001)
$21.7\pm1.6$	Barziv et al. (2001)
$22.6\pm1.5$	this paper

Early determinations of $K_{\rm o}$ were made by Wallerstein (1974) and Zuiderwijk et al. (1974), see Table 1. As regular X-ray pulsations had yet to be discovered at this point, assumptions had to be made about the mass of the optical component in order to estimate the mass of the compact object. Zuiderwijk et al. (1974) obtained a value of $M_{\rm x}>2.5\pm0.3~M_{\odot}$, and suggested that such a large mass coupled with the lack of regular pulsations indicated that the compact object was a black hole. The discovery of regular X-ray pulsations provided a means of determining the masses of both components directly, and also ruled out the possibility of the compact object being a black hole.

Van Paradijs et al. (1976) combined their $K_{\rm o}$ value obtained from 26 coudé spectrograms (Table 1) with the $K_{\rm x}$ value of Rappaport & McClintock (1975). Using X-ray eclipse data, they determined that $i > 74^{\circ}$, and thus arrived at a mass of $1.6 \pm 0.3~M_{\odot}$ for the neutron star. Van Paradijs et al. (1977) subsequently refined their $K_{\rm o}$value using yet more photographic spectra (Table 1) and Rappaport & Joss (1983) revised $K_{\rm o}$ further (Table 1) obtaining a neutron star mass estimate of $1.85^{+0.35}_{-0.30}~M_{\odot}$ by combining data from a number of sources, including Watson & Griffiths (1977), and Rappaport et al. (1980), and performing a Monte Carlo analysis to estimate the uncertainties.

More recently, van Kerkwijk et al. (1995) made further optical observations of GP Vel, and discovered strong deviations from a pure Keplerian velocity curve, which were auto-correlated within a single night, but not from one night to another. It was suggested that the variable gravitational force exerted by the neutron star as it travels around its eccentric orbit excites short-lived oscillations on the surface of the optical component which affect the measured radial velocity. From their $K_{\rm o}$ value (Table 1) van  Kerkwijk et al. (1995) obtained $M_{\rm x}=1.9^{+0.7}_{-0.5}~M_{\odot}$. A significantly lower value for $K_{\rm o}$ (Table 1) was obtained from observations using the IUE satellite by Stickland et al. (1997). However, Barziv et al. (2001) report that the analysis of these IUE data was subject to an error and a correct analysis yields a value consistent with those previously measured (Table 1) thus solving the discrepancy.

The most recent measurement of the optical radial velocity curve of GP Vel (Barziv et al. 2001) made use of 183 spectra obtained over a nine month campaign in order to try to average out the deviations reported by van Kerkwijk et al. (1995). Although they were quite successful in averaging out these excursions, they were left with different, phase-locked deviations in the radial velocity curve. Despite this they determined an accurate $K_{\rm o}$ value (Table 1) and set a limit on the neutron star mass of $M_x \sin^3 i = 1.78 \pm 0.15~M_{\odot}$.