Astron. Astrophys. 334, 799-804 (1998)

3. Results and spectral fits

3.1. Core cm-mm spectra

The core spectra for 3C 47, 3C 207 and 3C 334 are plotted in Fig. 1, using data from Tables 3, 4, 5, and 6. Our 1995 data, which will be used for the spectral energy distribution (SED) fits, are plotted as open circles. The unusually high IRAS 60µm points appear on the right in each figure as filled circles. Literature data are plotted as stars, the 1982/1983 VLA archive data as filled triangles. All errors are within the sizes of the points. The following remarks can be made for the individual objects.

Fig. 1. Rest frame core spectra for the IRAS detected quasars, including literature and 1983 VLA data. Open circles represent our 1995 data, filled triangles 1983 VLA data, and stars literature data. The filled circles represent the total (thermal plus nonthermal) IRAS m flux densities. All error bars fall within the size of the points, except for the IRAS observations.

3.1.1. 3C 47

An archival VLA 5 GHz dataset (observed by Ekers in June 1982, C-array) appeared in good agreement with our observations: no evidence for core variability is seen. Also the literature data show no sign of cm variability over the last twenty years. Since this quasar seems to have a stable core, we decided to include the 1.4 GHz datapoint (1985 observation, VLA A-array) from Fernini et al. (1991) in our fitting procedure.

3.1.2. 3C 207

For this source we retrieved VLA archive observations at C- and U-band, made by Wardle in March 1982 (A-array) and May 1983 (C-array) respectively. The core flux densities are in good agreement with our 1995 VLA data. Also, most literature data show no large discrepancies with respect to our data. However, Hough (1986) reports changes of 100 - 200 mJy in the early 1980's. Comparison of transatlantic VLBI with VLA data at 5 GHz is also indicative of some level of variability in the 3C 207 nucleus. Since the core shows no strong variability, we added the (1992 VLA A-array) 20 cm flux density value from Bogers et al. (1994).

3.1.3. 3C 334

An extensive archival VLA data set was available for this quasar, observed by Perley in January 1983 (C-array). Observations were carried out in C-, U-, and K-band. 3C 334 displays substantial core variability with respect to our 1995 data, as seen from Figs. 1 and 3. Fig. 3 displays 5 GHz core flux density values, measured at resolution over the last two decades (data from Table 2 in Hough et al. 1992).

3.2. Core spectral energy distributions

To obtain a good estimate of the (beamed) nonthermal 60µm flux density, we fitted curves to the nearly simultaneous 1995 cm and mm data. As mentioned above, for 3C 47 and 3C 207 we added the 20cm (L-band) points from Fernini et al. (1991) and Bogers et al. (1994) respectively. Upper limits are excluded from the fitting procedures. The fitted curves can subsequently be extrapolated into the far-infrared. However, the shape of the curve is a priori unknown. We therefore used two different models. First we fitted a parabola, which is an empirical result for blazar spectra (Landau et al., 1986). In the light of the unification theory and provided we are dealing with single components, these parabolic fits can be used to fit core spectra for double-lobed quasars. It should be noted however that Brown et al. (1989) found evidence for double components in their blazar sample. We will come back to this issue in Sect. 4.1.

Second we used the theoretical spectrum of a single self-absorbed synchrotron source, which has a powerlaw at frequencies larger than the turnover frequency. For this fit we excluded all the frequencies below the turnover frequency. Both fits were done using least-squares methods. The resulting fits, in the QSR rest frames, are shown in Fig. 2. With the exception of 3C 47, the parabolae fit remarkably well, but it should be kept in mind that their high frequency parts are not well constrained.

Fig. 2. Rest frame fits for the IRAS detected quasars, using only our 1995 data and additional 20cm points for 3C 47 and 3C 207. The straight lines represent powerlaw fits to the high frequency points, the parabola fits draw from an empirical result for blazar spectra. The dashed parabola in Fig. a includes the 3mm point, while the solid parabola does not. See text for details on the fitting procedure.

Fig. 3. Variability of 3C 334 over the last two decades. Data points taken from Table 2 in Hough et al. (1992)

Given the upper limits at 1cm and 7mm, a parabola does not yield a good fit to the 3C 47 data when we include the 3mm point (dashed parabola in Fig. 2a). Leaving out this 3mm point, a single parabola does fit the 3C 47 data remarkably well (solid parabola in Fig. 2a). A secondary (sub)mm component may be causing the rather high 3mm point. In the case of 3C 334, the parabola does fit well (solid parabola in Fig. 2c), but the Q-band upper limit is below our fit. If we exclude the 3mm point from the fit procedure, the parabola does not change significantly. The dashed parabola is a fit including the Q-band upper limit and excluding the 3mm point.

As can be seen immediately, only in 3C 207 the IRAS flux density may suffer from substantial nonthermal contamination, adopting the high frequency powerlaw fit. For the other two quasars the far-infrared point is far above the extrapolated nonthermal value, leaving a factor of 100 to account for. Note that especially 3C 47 stands out: the IRAS point is even higher than the radio peak flux density. The extrapolated values of the FIR flux density for different models are compared with the observed values in Table 7. is the IRAS flux density from Hes et al. (1995), are the extrapolated flux densities and is the FIR excess defined as the difference between observations and models at 60µm observed wavelength. All values are in mJy/beam.

Table 7. Comparison of measured total and extrapolated nonthermal FIR flux densities, in mJy.

© European Southern Observatory (ESO) 1998

Online publication: June 2, 1998

helpdesk.link@springer.de