From three epochs of data, one each in 1981.2, 1983.4, and 1990.5, we have measured at 3.6 cm significant changes of up to 170 as in the angular separation of the reference point chosen in quasar 1038+528 A from the one chosen in quasar 1038+528 B. We interpret this change as being due primarily to the motion [ as yr-1, v c] of the reference point in quasar B relative to its core. For this interpretation, we conclude that there is no "bona fide" proper motion of the quasar A, the upper bound set on any such motion being about 10 as yr-1 (v c).
The null result on proper motion obtained from adding the 3.6 cm data from the third epoch of observations (1990.5) are consistent to within less than 1.5 standard deviations with the proper motion of 1038+528 A of as yr-1 [corresponding to, v= c; km s-1 Mpc-1 ; ] deduced from the 1981.2 and 1983.4 3.6 cm observations (MES 1994).
We detected no change in the relative positions of the quasars from our observations at 13 cm, where the interferometric beam size is nearly four times larger (see Table 3 for contributions to the astrometric uncertainty). However, comparison of the observations at 3.6 cm and 13 cm shows that the position of the observed core in quasar A is frequency dependent (Marcaide & Shapiro 1983, 1984). Such an effect has since been reported for other sources, e.g. 3C345 (Biretta et al. 1986) and 3C395 (Lara et al. 1996).
At cosmological distances the standard theory for extragalactic radio sources predicts stable cores at the as level. So far, astrometric studies have not reached this level of precision. Bartel et al. (1986) obtained bounds on the proper motion of 3C345 (with respect to NRAO150) equal to as yr-1 and as yr-1. The bound on the proper motion of quasar A presented here, as yr-1, is the lowest upper limit set to date, despite much lower flux densities in this source and in the reference source compared to the 3C345-NRAO150 pair, because of the formers' smaller angular separation. In our case, the accuracy of the angular separation estimate is limited by the signal-to-noise ratio for the maps and the interferometric resolution, both of which affect the determination of the location of reference features. This limitation is independent of the relative separation of the sources. For sources with larger angular separations between them, other effects usually limit the precision (Guirado et al. 1994).
The results from the third epoch of observations in 1990.5 allow us to reject the hypothesis of a possible "bona fide" proper motion of as yr-1 for the core of quasar A, proposed (MES 1994) as one possible explanation for the change in position between 1981.2 and 1983.4, and favors the alternative model, proposed by those authors, based on opacity changes in the jet emission, where the displacements between 1981.2 and 1983.4 are related to core "jitter" effects in the quasar A, or the alternative of no motion.
The flat-spectrum quasar 1038+528 B displays expansion in the separation between its two main components at a rate near c. In the context of relativistic beaming models, such motions will be observed in sources with jets that move in directions not very close to perpendicular to the plane of the sky, and hence would be expected to display weak cores and jet-to-counter-jet flux ratios not greatly exceeding unity.
Based on its spectral index (Marcaide & Shapiro 1984), we assume that the centre of mass of quasar B is near the westmost component in the 3.6 cm maps. The astrometric analysis presented here used the eastmost component as the reference point, because it had the highest flux density. The selection of a weaker component as a reference point in the astrometric analysis would have had disadvantages: the uncertainty in the location of the reference point would have been larger; in addition, the core of quasar B might exhibit a "jitter" phenomenon, such as might be present in the core of quasar A.
Although our scenario successfully reproduces the kinematics of this quasar pair, we consider it important to continue to monitor these sources at the same frequencies and to start monitoring at new frequencies, with the highest feasible angular resolution. Such monitoring should allow further investigations, with unprecedented detail, of the physical conditions, and their variability, in a region very close to the central engine of a quasar.
© European Southern Observatory (ESO) 1997
Online publication: May 5, 1998