Astron. Astrophys. 325, 383-390 (1997)

1. Introduction

The application of Very-Long-Baseline Interferometry (VLBI) to the determination of positions of objects on the sky, and telescopes on the Earth (i.e., astrometry and geodesy) routinely provides precisions at the milli-arcsecond (mas) and centimeter levels, respectively. VLBI astrometry has allowed a celestial reference frame to be built on the stable positions of extragalactic radio sources determined with mas accuracy (Ma et al. 1990); this technique is now also accurate enough for geodetic determinations which are geophysically significant (see, for example, Herring et al. 1986; Robertson 1991; Haas et al. 1995). Its inherent resolution makes the VLBI technique ideal for the study of the kinematics of the inner parts of radio sources. The standard theory of extragalactic radio sources (relativistic jet theory; e.g., Blandford & Königl 1979) assumes that emission from quasars and active nuclei of galaxies comes from a central engine where energetic nuclear phenomena take place; this engine is believed to be stationary on the sky at the micro-arcsecond (µas) level.

VLBI position determinations are limited by errors in calibrating the contributions of the atmosphere and of source structure to propagation delays (Clark et al. 1989), and not by instrumental errors or the precision of the measurements. The interferometric quantities of astrometric interest are the phase delay, group delay, and phase delay rate (Shapiro 1976). The ratio of the statistical errors in the estimates of the group delay to those in the phase delay is about and is given approximately by the ratio of the central radio frequency (, a few GHz) observed to the rms spread of the totality of radio frequencies observed (some tens to hundreds of MHz) (see, for example, Thompson et al. 1986). Despite its high precision, the phase delay is not used routinely because of the difficulty in eliminating the " ambiguities". Its use requires knowledge of the changes between successive observations of the contributions to the delay from the propagation medium, the geometry, and the instrumentation, with an uncertainty well under 1/ .

However, the difficulty is substantially reduced in difference astrometry. The difference between the phase delays for two sources close to one another on the sky is affected far less than the undifferenced delays by variations, for example, in the neutral atmosphere. The angular separation between the two sources is a crucial factor for the successful application of this technique. The radio sources 1038+528 A and B (Owen et al. 1978, 1980) are 33" apart and therefore lie within the primary beam of most radio telescopes at cm wavelengths. This pair is an excellent one for the application of differential astrometric techniques; the achievable accuracy in estimating the change in the angular separation between these sources at different epochs is limited primarily by the structure of the sources, i.e. by the uncertainty in locating stable reference features in the brightness distributions of the sources.

Simultaneous observations of the radio sources 1038+528 A and B at 3.6 cm and 13 cm have spanned nearly a decade; their purpose has been to monitor the position of the core of one of them (1038+528 A, z=0.678) with respect to that of the other (1038+528 B, z=2.296). Observations at other wavelenghts have been reported by Marcaide et al. (1985) and Elósegui (1991). The observations of this pair made in March 1981 (1981.2) and May 1983 (1983.4) at 3.6 cm and 13 cm are described elsewhere (Marcaide & Shapiro 1983, 1984; Marcaide et al. 1985; Elósegui 1991; Marcaide et al. 1994, hereafter MES 1994). As a result of their analyses, a spectral dependence of the position of the observed core of 1038+528 A was discovered and later confirmed, and was explained in terms of opacity effects (Marcaide & Shapiro 1983, 1984). In addition, an unexpectedly large change in the angular separation between the A and B quasars from 1981.2 to 1983.4 was noted (MES 1994). To discriminate among several hypotheses proposed to explain this unexpected change, new observations were made in June 1990 (1990.5). We report here the analysis of the data from this third epoch, and propose a scenario which reproduces the astrometric results for all three epochs.

© European Southern Observatory (ESO) 1997

Online publication: May 5, 1998

helpdesk.link@springer.de