Astron. Astrophys. 331, 821-828 (1998)

1. Introduction

In addition to about 15 radio pulsars associated with supernova remnants (Gaensler & Johnston 1995), several radio-silent isolated neutron star (NS) candidates within supernova remnants (SNRs) have been observed with the X-ray observatories HEAO -1, Einstein, EXOSAT, ROSAT, and ASCA (see Caraveo et al. 1996 for a review). These objects have not been detected outside the X-ray range, and their X-ray spectra resemble blackbody (BB) spectra with temperatures of a few million kelvin. If they are indeed thermally emitting NSs, the analysis of their radiation provides an opportunity to study thermal evolution of NSs of ages  yr, which is important for elucidating the properties of the superdense matter in NS interiors.

One of the most convincing examples of such objects is the point-like source 1E1207.4-5209 within the barrel-shaped radio, X-ray, and optical SNR PKS 1209-52 (also known as G269.5+10.0). From the analysis of radio and optical observations of this SNR, Roger et al. (1988) estimated its age  yr, with an uncertainty of a factor of 3, and concluded that the remnant is in an adiabatic expansion phase. The distance to PKS 1209-52 is not well determined - estimates in the range 1.1-3.9 kpc were suggested (Milne 1979; Mills 1983). Estimates of the interstellar hydrogen column density from the radio and optical data yield (see Roger et al. 1983 and Kellet et al. 1987 for references), consistent with a distance  kpc.

After the first X-ray detection of PKS 1209-52 with HEAO -1 (Tuohy et al. 1979), the point source 1E1207.4-5209 was discovered with the Einstein observatory (Helfand & Becker 1984), off-center the diameter SNR. Matsui et al. (1988) concluded that its spectrum can be interpreted as a BB spectrum with an apparent temperature  K (assuming , as inferred from the data obtained with HEAO -1). The Einstein High Resolution Imager (HRI) observations showed a lack of diffuse X-ray emission around 1E1207.4-5209  (later confirmed by the ROSAT HRI observations), which greatly simplifies the analysis of the point source radiation. From the analysis of observations made with the EXOSAT Position Sensitive Detector, Kellett et al. (1987) estimated the BB temperature of the central object  K (at ), which implies an emitting area of a radius  km, where . Applying the Raymond & Smith (1977) line-emission model, Kellet et al. estimated also an average SNR temperature,  K, and the hydrogen column towards the SNR, , consistent with the values obtained from radio and optical data.

Observations of PKS 1209-52 and its central source with ROSAT and ASCA have further supported the NS hypothesis for 1E1207.4-5209. Mereghetti et al. (1996; hereafter MBC96) showed that the ROSAT data of 1E1207.4-5209  can be interpreted as blackbody emission of  K from an area with radius  km. The authors noticed that this temperature is too high to be explained in the framework of a cooling NS with the age of PKS 1209-52 ( yr). The hydrogen column inferred from the BB fit, , is 3-4 times lower than the estimate given by Kellett et al. (1987). From observations at 4.8 GHz, MBC96 found an upper limit of  mJy on the radio flux from 1E1207.4-5209. They also set a deep limit of for an optical counterpart in the Einstein HRI error circle, which supports the hypothesis that 1E1207.4-5209  is indeed an isolated NS.

Vasisht et al. (1997; hereafter V97) have recently analyzed ASCA observations of 1E1207.4-5209. They found that each of the three spectra obtained with the ASCA detectors can be fitted by a BB spectrum consistent with that obtained from the analysis of the ROSAT data. The hydrogen column is poorly restricted in the ASCA spectral fits due to low detector sensitivities at photon energies below 0.5 keV. Using a fit with a Raymond-Smith model at fixed cosmic abundances to the ROSAT data of the SNR shell, V97 estimated the remnant temperature  K and the foreground column density . They attributed the difference in between the NS and SNR fits to either separate lines of sight or the large column density that the shell X-rays can encounter in the SNR postshocked gas.

Based on the results of the blackbody analysis, both MBC96 and V97 concluded that 1E1207.4-5209 is an isolated NS with hot spots on its surface, aligned with the magnetic poles. V97 suggested that the spots are heated either by dissipative heating in the NS interior or by the bombardment of polar caps by relativistic particles from the NS magnetosphere if 1E1207.4-5209 is an active pulsar. However, the former hypothesis can hardly explain the small sizes of the hot spots, even with allowance for large anisotropy of thermal conductivity of the magnetized NS crust. The latter heating mechanism is also in doubt because of the absence of radio and -ray emission from 1E1207.4-5209. Moreover, both the ROSAT and ASCA data did not reveal pulsations, which may not be consistent with the presence of the hot spots unless the magnetic and rotation axes are coaligned.

Although it looks very plausible that 1E1207.4-5209 is a thermally emitting isolated NS, the BB interpretation adopted by previous authors leaves several controvertible points. Since thermal radiation emitted by NS atmospheres may significantly differ from the BB radiation, it is natural to employ more realistic models of NS radiation to resolve the inconsistencies following from the simplified BB interpretation. Here we present a combined analysis of the ROSAT and ASCA data based on NS atmosphere models (Pavlov et al. 1995, and references therein). These models have been applied successfully to the interpretation of the soft X-ray radiation from, e.g., the Vela pulsar (Page et al. 1996) and the brightest millisecond pulsar, J0437-4715 (Zavlin & Pavlov 1998; Pavlov & Zavlin 1997). These examples show that fitting soft X-ray pulsar spectra with hydrogen or helium atmosphere models results, as a rule, in lower effective temperatures and greater emitting areas (or smaller distances) than those obtained from the BB fits.

We show in Sect. 2 that, indeed, the model atmosphere fits of the X-ray radiation from 1E1207.4-5209 yield an NS surface temperature compatible with NS cooling models, and they do not require hot spots on the NS surface. Moreover, the hydrogen column density inferred from this interpretation is in excellent agreement with that obtained from our fits of the SNR X-ray radiation as well as with independent estimates of for stars in the vicinity of 1E1207.4-5209 (Sect. 3). These results warrant application of the same approach to other similar objects and enable one to obtain reliable estimates of surface temperatures of NSs of different ages.

© European Southern Observatory (ESO) 1998

Online publication: March 3, 1998
helpdesk.link@springer.de