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

4. Analysis and discussion

Since there is a substantial time interval between the (1983) IRAS observations and the other data, core variability could be a possible explanation for the FIR excess. However, long-term radio data from literature show that the core variations do not exceed 10% and therefore are not large enough to explain the total excess. In addition, all variability references report fairly low core flux densities for 1983 (IRAS epoch), which does not help to explain a FIR excess. Although the IRAS detections could be spurious, since the sources are close to the detection limit, or due to a confusing source, the issue remains why no spurious radio galaxy detections appear. Below we examine all possible explanations for the FIR excess observed in these quasars.

4.1. Non-thermal FIR

The main goal of this research was to find out if beamed non-thermal radiation could be responsible for the FIR excess. Adopting single component core models it is evident from the plots in Fig. 2 that beaming cannot be responsible, except for 3C 207 applying the powerlaw fit. However, we do know that beaming operates in 3C 47 and 3C 334, both superluminal objects. Both the FIR beaming and the radio core R-parameter depend on the Doppler factor to some power. In terms of R-parameter strength the strongest component of nonthermal FIR radiation would be expected from 3C 207, followed by 3C 334 and 3C 47. This order is not reflected in the total 60µm luminosities: the largest FIR excess, normalized with respect to the 5 GHz flux density, originates in the quasar with the smallest core fraction. While the reported superluminal velocities roughly scale with the infrared excess in 3C 334 and 3C 47, a measurement is still lacking in 3C 207. Moreover, these tests have little statistical significance.

The formula derived by Hoekstra et al. (1997) permits an estimate of the relative amount of beamed radiation. Hoekstra et al. use a relation depending on Q and to estimate the amount of beamed 60µm emission. Here Q is the observed 5GHz core fraction, and is a measure for the value of Q at which beaming becomes significant. They determined the value of from a large sample of FIR-detected blazars and quasars, including the three discussed here. This yields a maximum nonthermal FIR contribution of 15%, for 3C 207, and considerably less for 3C 47 and 3C 334.

The Hoekstra et al. (1997) analysis postulates a direct (single component) connection between the nonthermal radio and FIR emission. It is quite likely that the real situation is more complicated. As mentioned above, Brown et al. (1989) demonstrated the presence of two nonthermal core components in blazars. One component is fairly quiescent and dominates the radio to mm region. A second, ultracompact, component is prevalent in the submm regime. This component becomes self-absorbed at wavelengths longer than 3mm, and displays strong variability (flares), sometimes within days. Our data do not constrain the submm region well: it is likely that such variable submm components are being missed in our analysis. The possibility that such a component has pushed the total (thermal plus nonthermal) FIR into IRAS detection cannot as yet be ruled out. In fact, as inferred from the 100 GHz data, our 3C 47 observations may have caught such a submm component. In 3C 334 this might also be the case, considering that our Q-band upper limit is lower than the predicted value if we include the 3mm point in our fit.

We stress that since the IRAS detections are just above the detection limit, not the total 60µm emission has to be accounted for, but only a substantial nonthermal component lifting the total 60µm flux into IRAS detection. For 3C 47 and 3C 334, superluminal and hence beamed objects, the possibility of an additional, variable submm component is considered likely. Full sampling of the cm-mm-submm-FIR spectral range is needed to confirm our suspicion.

4.2. Thermal FIR

If most or all 60µm emission is identified with thermal radiation, the question arises why 3C 47, 3C 207 and 3C 334 are more luminous than other quasars and radio galaxies. There are several mechanisms that can produce thermal FIR in AGN: cold cirrus and warm starburst heated dust in the host galaxies, and furthermore warm AGN-related dust. Models for these components are for instance described by Rowan-Robinson & Crawford (1989). In order to distinguish between these, more IR data are needed to perform a detailed analysis of the FIR spectral energy distribution. Multiple component fitting is necessary to isolate the various dust components. However, if these mechanisms are responsible for the FIR excess, unification is difficult, since it states that radio galaxies and lobe-dominated quasars are basically the same objects and thus should have similar dust composition.

If we wish to maintain the unification concept we have to postulate optically thick dust emission at m in combination with aspect effects. If the FIR excess originates from an optically thick torus shielding the optical QSO in a plane perpendicular to the radio axis, models by Pier & Krolik (1992) yield aspect dependent anisotropies of factors up to about ten. At longer FIR wavelengths the anisotropies should become zero (optical thin dust): data at 100µm or longer wavelengths are needed to investigate optical thickness effects.

© European Southern Observatory (ESO) 1998

Online publication: June 2, 1998

helpdesk.link@springer.de