Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 355, 1041-1048 (2000)

Previous Section Next Section Title Page Table of Contents

5. Discussion

5.1. Orbital variations and binary parameters

There are basically two different models to explain the optical orbital variations in supersoft X-ray binaries: (1) the optical light is produced by reprocessing of the white dwarf emission on the accretion disk, and the orbital variations are due to the varying aspect of the illuminated, heated, and flared accretion disk (Schandl et al. 1997); (2) these variations are due to the changing aspect of the irradiated secondary (van Teeseling et al. 1998). An irradiated secondary, if dominating the optical light, would produce a smooth, sinusoidal light curve. This indeed is what we observe on three out of the five nights. However, the irregular brightness variations and minima of different depths as well as the asymmetric minima during the two other nights are exactly what the flared accretion disk would produce when the accretion rate varies and the interaction of the accretion stream with the disk causes either a varying amount of "spray", e.g. matter splashing at the stream-disk impact site (Schandl et al. 1997) or a varying accretion disk rim height (Meyer-Hofmeister et al. 1997). Based on our available data we cannot distinguish between these two alternatives.

The origin of the HeII emission line is still a mystery. If it came from the illuminated side of the donor, its radial velocity amplitude would have to be 300-400 km/s, much higher than the observed value of 115 km/s. Therefore, it is generally assumed that the HeII emission line originates near to the accretor. Indeed, the expected velocity amplitude of the accretor (for our best period) is in the range 100-150 km/s.

If we assume that the measured HeII emission-line velocity is roughly similar to the motion of the accretor, we may derive an estimate of the donor mass. Combining the orbital period and the velocity amplitude results in a mass function of f(M) = 0.023 [FORMULA] 0.014 [FORMULA]. If we assume the donor to be a main-sequence star filling its Roche lobe, the resulting mass for the donor is restricted to a narrow range of 0.32-0.37 [FORMULA], nearly independent of the inclination (Fig. 7).

[FIGURE] Fig. 7. Possible range of binary component masses based on our mass function of f(M)=0.023 [FORMULA]. Solid lines mark various inclinations of the binary plane, and dotted lines visualize constant mass ratio q = [FORMULA]. The dashed line indicates where the donor fills its Roche lobe (using the mass-radius relationship of Patterson 1984). The shaded area represents our best-estimate parameter space for the component masses (see text).

The low mass of the donor implies that it does not contribute to the total light, since the apparent magnitude of a 0.35 [FORMULA] star at the LMC distance is [FORMULA] mag. Thus, the optical emission as well as its orbital modulation must be caused differently, e.g. by the (irradiated) accretion disk. Also, it is very unlikely that the Balmer absorption lines arise in the illuminated secondary, unless the illumination increases the donor emission by 10 magnitudes. Since the white dwarf itself is also an unlikely source, we therefore tentatively assign the Balmer absorption lines to either the accretion disk or a possible wind. Substantially better data are needed to establish the velocity variations of the Balmer absorption lines, and thus shed more light on their nature.

The lack of detectable (at our accuracy) [FORMULA] color variation through the orbital cycle of RX J0537.7-7034 suggests that the inner, bluer (hotter) part of the accretion disk are never occulted (during the 18 Jan. 1998 B band observation). This implies that the inclination is smaller than 70o-74o. Since RX J0537.7-7034 has shown luminous supersoft X-ray emission for the period of approximately one year, the white dwarf mass should be large enough to allow hydrogen burning. Previous investigations suggest a minimum mass of 0.4 [FORMULA] (Sienkewicz 1980). This in turn implies (using Fig. 7) that the inclination should be [FORMULA]45 degrees. We therefore tentatively assume [FORMULA] 45o-70o in the following.

Since the inclination cannot be larger than 90 degrees, Fig. 7 also implies that the white dwarf mass is certainly smaller than 0.9 [FORMULA]. With the above best-estimate inclination range the mass of the white dwarf is [FORMULA] 0.8 [FORMULA].

5.2. On the nature of RX J0537.7-7034

We have determined a period of [FORMULA]3.5 hrs for the supersoft X-ray source RX J0537.7-7034. Since this is a period in the range typically populated by (low-luminosity) cataclysmic variables it is worth mentioning here that the systemic velocity of RX J0537.7-7034 of [FORMULA]350 km/s together with the characteristic spectrum of a SSS leaves no doubt that RX J0537.7-7034 is a supersoft X-ray source in the Large Magellanic Cloud, and not a galactic foreground cataclysmic variable. The lack of any other variable object within the X-ray error circle together with the improved X-ray position and the good positional match clearly suggest that the optical object studied here is indeed the optical counterpart of RX J0537.7-7034.

Many properties of RX J0537.7-7034 are very similar to those of SMC 13, a supersoft X-ray source in the Small Magellanic Cloud.. SMC 13 [FORMULA] 1E 0035.4-7230 [FORMULA] RX J0037.3-7214 (1RXS J003723.2-721415), has a 4.1 hrs orbital period (Schmidtke et al. 1996, Crampton et al. 1997, van Teeseling et al. 1998). In SMC 13 the Balmer absorption line system moves in phase with the HeII emission lines, but with a larger amplitude. Based on the interpretation of the light curve as partially eclipsing (which constrains the inclination to 70o [FORMULA] 78o) and the HeII radial velocity data, Crampton et al. (1997) derive a donor mass of 0.4-0.5 [FORMULA] and a mass of the accreting object of 1.3-1.5 [FORMULA]. In contrast, van Teeseling et al. (1998) interpret the orbital modulation by the varying aspect of the illuminated secondary, and derive an inclination of 20o [FORMULA] 50o. Also, fitting non-LTE models to the BeppoSAX X-ray spectrum, Kahabka et al. (1999) derive a white dwarf mass of 0.6-0.7 [FORMULA] consistent with the inclination range of van Teeseling et al. (1998). Thus, RX J0537.7-7034 shares the following characteristics with SMC 13: (i) the short orbital period, (ii) an almost identical optical spectrum, (iii) the existence of Balmer absorption lines, (iv) the different velocities of HeII emission and Balmer absorption lines, (v) the low visual absolute magnitude (see Table 4). (vi) the luminous, soft X-ray spectrum, and (vii) possibly a low white dwarf mass. But there is also one clear difference: RX J0537.7-7034 shows a factor of 7 X-ray variability over the past 8 years while 1E 0035.4-7230 has been completely constant (e.g. Kahabka et al. 1999).


Table 4. Comparison of brightness and colours of SSS. We assume here 18.5 mag as the distance modulus for the LMC (Panagia et al. 1991), 18.8 mag for the SMC and reddening E(B-V)=0.065 and E(B-V)=0.043 for the LMC and SMC, respectively.
1) References: (1) Crampton et al. (1987), (2) Smale et al. (1988), (3) Schmidtke et al. (1993), (4) Orio et al. (1997), (5) this work, (6) Schmidtke et al. (1996), (7) Crampton et al. (1997), (8) van Teeseling et al. (1998), (9) Schmidtke & Cowley (1996), (10) Teeseling et al. (1996), (11) Diaz & Steiner (1989)
2) During the phase as a supersoft X-ray source.

With [FORMULA] hrs RX J0537.7-7034 is the shortest-period binary among the SSS. This implies that the standard scenario of SSS, in which the donor is assumed to be more massive than the accreting white dwarf to ensure high mass transfer rates on a thermal timescale (van den Heuvel et al. 1992), is not applicable for this system. Under the assumption of Roche-lobe filling and a minimum white dwarf mass for steady-state burning of [FORMULA]0.5-0.6 [FORMULA] (Fujimoto 1982, Sion & Starrfield 1994, Cassisi et al. 1998), a mass ratio [FORMULA] implies that [FORMULA] hours always. In addition, with a donor mass of about 0.35 [FORMULA] it seems impossible to have a less massive white dwarf with H burning. Thus, RX J0537.7-7034 clearly does not fit this standard scenario.

Similar arguments, though not as strong (because the component parameters are less constrained), have been put forward for SMC 13 (Crampton et al. 1997). Thus, there are now two sources which require a new donor scenario. Alternative scenarios are either symbiotic systems (see Kahabka & van den Heuvel 1997), classical novae (e.g. Ögelman et al. 1993), the systems that Kahabka and Ergma (1997) define "SMC 13" type, and wind-driven binaries (van Teeseling & King 1998). Symbiotics, involving giant donors, have much too long orbital periods to be applicable to RX J0537.7-7034. A classical post-nova interpretation (similar to e.g. GQ Mus; Ögelman et al. 1993) is also unlikely since the LMC was continuously monitored to search for novae in the last 50 years, and a novae would have been detected in the optical passband. Thus, there are two serious alternative explanations for RX J0537.7-7034:

  1. SMC 13 systems: SMC 13 is interpreted by Kahabka & Ergma (1997) as a cataclysmic variable with a low mass white dwarf (0.6-0.7 [FORMULA] white dwarf) having a thick helium buffer layer. Such a system must have been a classical nova in which, after a number of outbursts, the white dwarf was heated up. As a consequence the flashes have become very mild, without actual mass loss, and the system is proposed to presently be in a phase of residual hydrogen burning after a mild shell flash. Van Teeseling & King (1998) have argued on statistical grounds that such a phase should last longer than 100 yrs. In order for the residual burning to last long enough (at least about 20 years since the Einstein discovery; Seward & Mitchell 1981), a low CO abundance of the burning matter and a hot, low-mass (0.6-0.7 [FORMULA]) white dwarf are required. Sion & Starrfield (1994) have demonstrated that hot white dwarfs with masses as low as 0.5 [FORMULA] can stably burn hydrogen in a steady state.

  2. wind-driven binaries: Van Teeseling & King (1998) have shown that the strong X-ray flux in supersoft sources should excite a strong wind ([FORMULA] [FORMULA]/yr) from the irradiated companion which in short-period binaries would be able to drive Roche lobe overflow at a rate comparable to [FORMULA]. This may self-consistently sustain stable recurrent or steady-state hydrogen burning on the white dwarf. In binaries with a low-mass companion, the angular momentum loss by the wind may dominate the binary evolution, and cause the period to increase with time. One important open question of this scenario is how a system could eventually enter such a self-consistent wind-driven phase. Van Teeseling & King (1998) suggested either a preceding SMC 13 phase, or a late helium shell flash of the cooling white dwarf after the binary has already come into contact as a cataclysmic variable.

Because of the X-ray variability of RX J0537.7-7034 its secure classification into one of the above two scenarios seems difficult. One could expect that the SMC 13 systems should display a rather smooth X-ray light curve, unless a mild flash is occurring. Such a flash should be accompanied by some optical brightening, however. Also, if the X-ray turn-off of RX J0537.7-7034 were to be explained by an only short phase ([FORMULA]1 year as observed) of burning, the white dwarf mass must be high ([FORMULA]1.1 [FORMULA]), contrary to our findings (see Fig. 7). Finally, the residual burning (between the H flashes) should be connected to a temperature increase (Sion & Starrfield 1994), while the observations tend to suggest the opposite behaviour (see the hardness ratio values in Table 3 which are softest [FORMULA] coolest during the last detection).

Whether or not the wind-driven supersoft source scenario is valid, remains to be seen: (i) The drop in X-ray intensity in 1993 is very probably not caused by increased absorption in the wind, otherwise the sensitive hardness ratio HR1 would have changed substantially; (ii) since it is difficult to enter the self-excited wind phase, any repeated on-states would be hard to explain - assuming that the observed X-ray decline corresponds to the sudden end of this wind-driven phase; (iii) this phase would have lasted less than two years (Fig. 6) - was it just an unsuccessful attempt to enter the wind-driven phase? Overall, the wind-driven supersoft source scenario seems to provide a possible explanation for RX J0537.7-7034, though it remains unclear what the observational consequences would be.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 2000

Online publication: March 21, 2000