Astron. Astrophys. 360, 447-456 (2000)

3. Results

3.1. Spatial analysis

As already mentioned, no useful spatial information on the X-ray structure of NGC 4303 can be obtained from the PSPC data due to the spatial resolution of the detector, and moreover to the 17´ off-axis position of the source in the FOV.

In contrast, the more detailed HRI image reveals a number of X-ray sources distributed over the galactic disk (Fig. 4). In comparison to the numbered sources in Fig. 1 the closer view allows to distinguish more details. For example, source no. 9 in Fig. 1 splits into three X-ray spots (labeled A-C in Fig. 3). The most luminous source coincides with the center of NGC 4303 and dominates in the soft X-rays. The count rates and fluxes derived for sources from the HRI are listed in Table 4. The corresponding areas are plotted in Fig. 3. To determine the fluxes we used the energy conversion factor (ECF) from the ROSAT Call For Proposals documentation. The ECF determines the ratio between count rates and unabsorbed source flux in the ROSAT band for given spectral parameters. For the disk sources A-F we assume a 0.3 keV Raymond-Smith model (Raymond & Smith 1977) with an absorption column density of 310 cm^-2. For the nucleus a power law with =2.6 and a column density of 310 cm^-2 (see Sect. 3.2 and Table 5) is applied as spectral model.

Fig. 3. The inner 4´ of the pointed HRI observation. The labeled sources (A-F) are the ones spatially coinciding with the optical disk of NGC 4303 (see text and Table 4 for details). The image is centered on the coordinates of the pointed observation. The distances on the axes are relative to the central X-ray source with the coordinates (2000) = ; (2000) = . North is up, east is to the left.

Fig. 4a and b. Overlay of the HRI X-ray contours onto the optical image of NGC 4303 a taken from Frei et al. (1996) and the H image b taken from MR92. The contour levels are at 5, 7, 10, 15, 20, and 30 above the mean background with = 1.510 cts s-1 arcsec-2. The images are centered on the coordinates of the pointed observation. The axis scales are relative to the central X-ray maximum at coordinates = ; = . North is up, east is to the left.

Table 4. X-ray sources spatially coinciding with the optical galaxy NGC 4303. The source designations refer to Fig. 3.
Notes:
a) Energy conversion factor ECF: disk sources A-F, ECF=2.410 cts cm² erg^-1; nucleus, ECF=510 cts cm² erg^-1

Table 5. Best spectral fits to the PSPC data of NGC 4303 and the derived model parameters. For the determination of the X-ray luminosity a distance of 16.1 Mpc is assumed.
Notes:
Col. (1) - Spectral models: BS = thermal Bremsstrahlung, RS = Raymond-Smith, PO = power law.
Col. (2) - Column density in units of 10²⁰ cm^-2.
Col. (3) - Plasma temperature in units of keV.
Col. (4) - Photon index.
Col. (5) - Metallicity in units of .
Col. (6) - Scaling factor: for BS in units of (10^-18/(4))dV, = electron and ion densities (cm^-3); for RS in units of (10^-19/(4))dV, = electron and H densities (cm^-3); for PO in units of 10^-4 photons keV^-1 cm^-2 s^-1 at 1 keV.
Col. (7) - Reduced .
Col. (8) - Degrees of freedom.
Col. (9) - Unabsorbed X-ray flux in units of 10^{- 12} erg cm^-2 s^-1. Values in brackets give the contribution of the thermal component.
Col. (10) - X-ray luminosity in units of 10⁴⁰ erg s^-1. Values in brackets give the contribution of the thermal component.

The contours of sources B, C, D, and F in Fig. 3 are all located within the optical arm structure and coincide with bright H emission regions within the spiral arms (Fig. 4). In addition, source E is embedded in the faint outer part of the southwestern spiral arm. MR92 distinguished 79 H  II regions in NGC 4303, mainly in the spiral arms. The X-ray contour overlay over the H image in Fig. 4 reveals that the X-ray sources B, C, D, and F coincide with H  II regions, while sources A and E are located near such regions.

Gas dynamical models of barred galaxies (Englmaier & Gerhard 1997) show strong gas accumulation at the tips of the bars due to corotation of the bar structure with the disk what should lead to enhanced star formation. H  I observations as well as the existence of prominent H features strikingly support the outcome of these models. The X-ray contours B and F seem to arise from these regions. The X-ray maximum D is connected with another interesting feature of NGC 4303: in the eastern part the galactic arm seems to be deformed to a boomerang-like bow where source D lies at the bend but without any significant brightening in H.

The lower X-ray contours of the nucleus indicate a possible extended source. Recent high-resolution UV observations of the central region with the Hubble Space Telescope reveal a spiral-shaped structure of massive young (2-3 Myr) star-forming regions with an outer radius of 225 pc (CA99). This structure cannot be resolved by HRI. Due to the low age of the star clusters almost no thermal X-ray emission is expected at the galactic nucleus (see Sect. 4). The extended X-ray contours may originate from additional sources at distances of about 1 kpc around the nucleus.

No X-ray emission has been detected from the possible interaction companions NGC 4303 A and NGC 4292.

3.2. Properties of the X-ray spectrum

Since the HRI maxima are separated by only 50", and because of the reasons mentioned in Sect. 2.2, the PSPC observations do not allow to study the spectra of the X-ray components of NGC 4303 individually. ROSAT PSPC detected 50524 backgroung-subtracted source counts from NGC 4303 in a total integration time of 8135 sec. The spectra of the source and the background are shown in Fig. 5.

Fig. 5. Background-subtracted ROSAT PSPC spectrum of NGC 4303 in the energy range of 0.1-2.4 keV. The 505 photons were binned to get a signal-to-noise ratio of 6. The dashed-dotted line represents the contribution of the background (see text).

We fitted the spectrum with several single-component models, as Bremsstrahlung (BS), Raymond-Smith model (RS), and a power law (PO), and a combined RS-PO model. The results are listed in Table 5.

A single power-law model implies the assumption, that the active nucleus of NGC 4303 dominates the X-ray emission. Furthermore, the sources detected by the HRI in the galactic disk would also have to be described with the same power law. The photon index in this model is =3.20.2. The emission of an AGN in the ROSAT energy band is best described by a power law with a photon index of 2.4; nevertheless some cases have been observed with 3 (MCG -5-23-16: Mulchaey et al. 1993; Mkn 335: Turner et al. 1993). High-mass X-ray binaries (HMXB) found in young star-forming regions in the spiral arms have a similar spectral shape in the 0.1-2.4 keV energy range with a photon index of 2.7 (Mavromatakis 1993). The column density of the absorbing component amounts to 5.710 cm^-2, which is by a factor of 3 higher than the Galactic foreground H  I column density (Dickey & Lockman 1990, DL90). Nevertheless, self-absorption within NGC 4303 must be expected, and small-scale deviations from the observed Galactic value by DL90cannot be ruled out and may result in a higher absorption from the Milky Way. The resulting 0.1-2.4 keV X-ray luminosity amounts to 1.310 erg s^-1. The flux portion from the sources outside the nuclear region as observed with the HRI amounts to 1.410 erg s^-1 in the case of a single power-law emission model with =3.2 using the corresponding ECF of 110 cts cm² erg^-1. Assuming a mean X-ray luminosity of 10³⁷ erg s^-1 for an HMXB, as observed in the Milky Way (Fabbiano et al. 1982; Watson 1990) would require an unlikely high number of 1400 of these systems to produce the observed X-ray flux. The ratio of OB stars to HMXBs is assumed to be 500 (Fabbiano et al. 1982). This means that a total number of 710 OB stars would be required to account for the HMXB X-ray flux in NGC 4303. Even if we consider to have 10⁵ OB stars in NGC 4303, as observed e.g. in Mkn 297 (Benvenuti et al. 1979), it is still a factor of 7 higher than expected. Moreover, this is the required number only for the disk sources and would involve almost 1.510 in massive stars with a Salpeter IMF and, by this, would require a moderately high SFR of about 15  yr^-1 in the disk. On the other hand, the corresponding supernova type II (SN  II ) rate (0.1 yr^-1) should contribute to the X-ray emission via hot gas.

The single component models BS and RS show similar results. Consequently, we only achieve an adequate fit of RS with very low metallicity, e.g. the portion of emission lines to the spectrum is very small. In contrast, it is expected that emission lines of highly ionized elements, like Fe and Mg, should play an important role in the X-ray spectrum of SNe  II in starburst regions because of the nucleosynthesis of massive SN  II progenitor stars (Woosley & Weaver 1985). In both models the column density is about 310 cm^-2 and the plasma temperature is 0.6 keV. For the BS model we get a total X-ray luminosity of 410 erg s^-1, for the RS model it is 3.510 erg s^-1. RS models with different higher metallicities yield unacceptable fits.

The fit of the X-ray spectrum with a two-component model (RS+PO) is only slightly better than the one-component fits. Nevertheless, from the points mentioned above and the physical picture discussed in Sect. 4 this model serves as the best explanation for the observed soft X-ray emission. Hydrogen column density (=3.310 cm^-2) and power-law spectral index (=2.6) lie within the expected range (as discussed for the single power law above). The plasma temperature of 0.3 keV fits with the observed values of other galaxies (e.g. NGC 253: Forbes et al. 1999; NGC 1808: Junkes et al. 1995). The total 0.1-2.4 keV luminosity for this model amounts to 4.710 erg s^-1 with 13% contribution from the RS component. The spectral fit together with the residuals is plotted in Fig. 6.

Fig. 6. Fit of the observed ROSAT PSPC spectrum of NGC 4303 with the two-component model RS+PO (solid line). The parameter values for this model are listed in Table 5. The long-dashed line shows the spectrum of the thermal model component, the short-dashed line represents the power-law component. The residuals of the fit are plotted in the lower box.

© European Southern Observatory (ESO) 2000

Online publication: August 17, 2000
helpdesk.link@springer.de