1 Introduction

PSR B0950+08 is the fourth in the list of the first radio pulsars discovered in 1968 (e.g., Bell Burnell 1998). Since that time it has been extensively studied in a wide radio frequency range, from 0.102 to 10.55 GHz. The pulsar parameters (period P and its derivative $\dot P$, age $\tau$, magnetic field B, dispersion measure $D\!M$, spin-down luminosity $\dot E$, distance d, position $\alpha$, $\delta$, proper motion $\mu_\alpha$, $\mu_\delta$, and parallax $\pi$) are shown in Table 1. So far this is the oldest pulsar among ordinary pulsars detected outside the radio range.

 

Table 1: Parameters of PSR B0950+08 (from Taylor et al. 1993, unless specified otherwise).

$$\begin{tabular}{ccccccccccc} \hline\hline \multicolumn{5}{c}{... ...\vrule height 2.5ex width 0ex depth 0ex\\ \hline \end{tabular}$$

a The position at the epoch of the Subaru observations, MJD 51930.	d The numbers in parentheses are uncertainties height 2.5ex width 0ex depth 0ex
b See Brisken et al. (2002).	referring to the last significant digit quoted,
c Parallax based distance.	e.g., $1.234(56)=1.234\pm 0.056$.

The X-ray counterpart of PSR B0950+08 was first found by Seward & Wang (1988) and Cordova et al. (1989) with the Einstein observatory. Later, it was observed with the ROSAT by Manning & Willmore (1994), who analyzed the soft X-ray spectrum in (0.08-2.4) keV range. The blackbody (hereafter BB) spectral fit gave a temperature $T_{\rm BB} = (2.1 \pm 0.6) \times 10^6$ K (90% confidence limits) and a very small radius of the emitting region $R_{\rm BB} \le 20$ m. The pulsar distance was assumed to be of about 130 pc based on early radio parallax measurements by Gwinn et al. (1986) with the VLBI. The power law (hereafter PL, $F_\nu\propto\nu^{-\alpha_\nu}$) spectral fit was also acceptable, but, owing to the poor count statistics for this faint X-ray object, it resulted in a very uncertain spectral index ${\rm -0.2 \le \alpha_{\nu} \le 3.1}$. An evidence of the pulsed X-ray emission with the pulsar period was reported by Wang & Halpern (1997) using archival ASCA observations in (0.5-5) keV range. However, later the authors recalled their ASCA measurements of PSR B0950+08 since they discovered that the pulsar flux in ASCA data is heavily contaminated by a brighter X-ray source, a previously unknown Seyfert galaxy, only $1\farcm5$ away from the pulsar (see a short Note in Wang et al. 1998). In accordance with NS cooling theories, the temperature of the whole surface of this relatively old NS should be less than 10⁵ K (e.g., Yakovlev et al. 1999). This is much less than observed by the ROSAT. The high temperature and small emitting area inferred from the BB fit can be explained by the presence of small hot polar caps at the NS surface produced by the impact of relativistic particles from the pulsar magnetosphere, see, e.g., a discussion in Wang & Halpern (1997) and references therein.

An optical counterpart of PSR B0950+08 was suggested by Pavlov et al. (1996) based on observations of the pulsar field with the HST/FOC with the long-pass F130LP filter ( $\lambda\lambda = 2310{-}4530$ Å). A faint ( $m_{\rm F130LP}=27$ $.\!\!^{\rm m}$1) point-like object was found with the 1 $.\!\!^{\prime\prime}$85 offset from the pulsar radio position. The offset was later revised and decreased to $\simeq$1 $^{\prime \prime }$ by Pavlov (2000). If this is a pulsar, it is the faintest pulsar ever detected in the optical. For comparison, the visual magnitude of the young Crab pulsar, which is about ten times more distant but $1.75\times 10^4$ times younger, is 16 $\hbox{$.\!\!^{\rm m}$ }65$ (e.g., see the review by Mignani et al. 2000).

Assuming that the detected optical object is the pulsar, Pavlov et al. (1996) showed that the extension of the ROSAT BB fit into the UV-optical range gives a flux lower than observed by 3-4 orders of magnitude. This excludes thermal radiation from the pulsar polar caps as a possible source of the optical radiation. The assumption that the detected optical flux is due to thermal emission from the entire surface of a NS with a BB radius $R_{\rm BB}=13$ km yielded the surface temperature $T_{\rm BB} \sim 7 \times 10^4$ K at ${\rm d=130}$ pc. This is still too high for the $\sim$ $1.75 \times10^7$ yr cooling NS and can be only explained by some reheating mechanism operating inside the star (e.g., Miralles et al. 1998). On the other hand, the optical flux would be in agreement with the extension of the PL X-ray fit if the index ${\rm \alpha_{\nu}}$ lies within the 0.26-0.35 range. It is in agreement with the ROSAT data but needs a confirmation by deeper observations in X-rays and by the detection of the counterpart in other optical bands.

Possible detection of PSR B0950+08 in the R band with the 6 m telescope BTA has been reported by Sokolov et al. (1998) and Kurt et al. (2000). An object with $R=25\hbox{$.\!\!^{\rm m}$ }4\pm0.3$ was marginally ( $S/N \simeq 3$) detected in poor seeing conditions. If confirmed, this detection suggests that the pulsar may be much brighter in the optical and may have a very steep increase of the spectrum towards longer wavelengths than one would expect from the detection in the near UV.

In this paper we report the observations of the PSR B0950+08 field in the B band with the Subaru telescope. We analyze our data together with the available optical-UV data from the HST, and with the X-ray data from the ROSAT, making use of the recent much more accurate radio measurements of the pulsar proper motion, parallax, and distance with the VLBA by Brisken et al. (2002). Observations and data reduction are described in Sect. 2. In Sect. 3 we present the astrometrical referencing and photometry, and in Sect. 4 we discuss the results and their implications.

 

 

Table 2: Log of observations of the pulsar in the B band.

Exposure	UT	Duration	Airmass	Seeing
number	21 Jan. 2001	s		arcsec
1	12:41	600	1.037	0.65
3	13:21	600	1.085	0.66
4	13:54	600	1.129	0.69
5	14:08	600	1.164	0.72
6	14:22	600	1.205	0.73
8	14:50	600	1.313	0.75