The Compton Gamma Ray Observatory (EGRET high-energy instrument) detected roughly 70 blazars in the course of its sky survey and pointed observations. In support of this effort, we undertook a massive campaign to monitor the parsec-scale structure of 42 gamma-ray blazars over a nearly four-year period. Our VLBA observations, mostly at 22 and 43 GHz, provided images with resolutions as fine as about 0.15 milliarcseconds. Some objects were monitored more regularly than others, and of course there was the usual gap of about six months when the NRAO scheduling committee decided that we did not need to monitor so many objects so frequently. They were wrong: We found that gamma-ray blazars are very fast, with an average apparent superluminal speed of bright knots (moving away from the core, which is at one end of the jet) of 10-12c (for a Hubble constant of 65 and either a low value of the cosmic deceleration parameter - 0 or 0.1 - or the new cosmology of Omega (matter) ~ 0.3 and Omego (Lambda) ~ 0.7). So, at the ultra-fine resolution of our images, features were difficult to follow from one observational epoch to the next. Nevertheless, we were able to identify many knots across epochs in order to determine their proper motions and times of ejection.

You can view and, if you wish, download the images on our gamma-ray blazar image page.

Three journal papers came out of this project. The first paper, a massive compendium of images and analysis, was published in the Astrophysical Journal Supplement Series in June 2001, by Svetlana Jorstad, Alan Marscher, Ann Wehrle (JPL), John Mattox (Francis Marion U., formerly at Boston U.), and Steven Bloom (Hampden-Sydney College, formerly at Boston U. and IPAC/JPL). Here is the abstract, which summarizes our findings.:

We present the results of a program to monitor the structure of the radio emission in 42 gamma-ray bright blazars (31 quasars and 11 BL Lac objects) with the VLBA at 43, 22, and occasionally 15 and 8.4 GHz, over the period from November 1993 to July 1997. We determine proper motions in 33 sources and find that the apparent superluminal motions in gamma-ray sources are much faster than for the general population of bright compact radio sources. This follows the strong dependence of the gamma-ray flux on the level of relativistic beaming for both external-radiation Compton and synchrotron self-Compton emission. There is a positive correlation (correlation coefficient r=0.45) between the flux density of the VLBI core and the gamma-ray flux and a moderate correlation (partial correlation coefficient r=0.31) between gamma-ray apparent luminosity and superluminal velocities of jet components, as expected if the gamma-ray emission originates in a very compact region of the relativistic jet and is highly beamed. In 43% of the sources the jet bends by more than 20 degrees on parsec scales, which is consistent with amplification by projection effects of modest actual changes in position angle. In 27 of the sources in the sample there is at least one non-core component that appears to be stationary during our observations. Different characteristics of stationary features close to and farther from the core lead us to suggest two different classes of stationary components: those within about 2 milliarcseconds (mas) of the core, probably associated with standing hydrodynamical compressions, and those farther down the jet, which tend to be associated with bends in the jet.

The second paper, by Svetlana Jorstad, Alan Marscher, Ann Wehrle, John Mattox, and Margo Aller and Hugh Aller (U. Michigan), compared the times of ejections of superluminal knots and of variations in the total and polarized radio flux density with epochs of high gamma-ray flux. Here is the abstract:

We examine the coincidence of times of high gamma-ray flux and ejections of superluminal components from the core in EGRET blazars based on a Very Long Baseline Array (VLBA) monitoring program at 22 and 43 GHz from 1993 November to 1997 July. In 23 cases of gamma-ray flares for which sufficient VLBA data exist, 10 of the flares (in eight objects) fall within 1-sigma uncertainties of the extrapolated epoch of zero separation from the core of a superluminal radio component. In each of two sources (0528+134 and 1730-130), two successive gamma-ray flares were followed by the appearance of new superluminal components. We carried out statistical simulations that show that if the number of coincidences is 10, the radio and -ray events are associated with each other at greater than 99.999% confidence. Our analysis of the observed behavior, including variability of the polarized radio flux, of the sources before, during, and after the gamma-ray flares suggests that the gamma-ray events occur in the superluminal radio knots. This implies that the gamma-ray flares are caused by inverse Compton scattering by relativistic electrons in the parsec-scale regions of the jet rather than closer to the central engine.

The third paper, by Alan Marscher, Svetlana Jorstad, John Mattox, and Ann Wehrle, presents a VLBA image for each blazar, either one that includes the polarized intensity and electric-vector position angles (EVPAs) or, for those not observed in 1997 when the polarization observations were carried out, the best total-intensity image over all the epochs. We searched for correlations between polarization properties and other characteristics of the source, but without much success. Here is the abstract:

We present Very Long Baseline Array images of 42 gamma-ray bright blazars, including 36 with polarization vectors, obtained during the course of a multiepoch monitoring program. Each object was observed at either 43 or 22 GHz, with some objects observed at both frequencies and/or at 15 or 8.4 GHz. The morphologies are varied, with some of the blazars displaying long, thin jets, others short, broad jets, and still others containing cores with only very weak features that are probably knots in faint jets. The polarization of the cores ranges from less than 1% to 8.6%, with electric vector position angles (EVPAs) that are split between nearly parallel and nearly transverse to the jet axis. The polarization of knots in the jets covers a much broader range, from less than 2% to tens of percent. The EVPA of the brightest compact feature in each jet ranges from 0° to 80° from the jet position angle, with roughly half measuring less than 20°. The distribution is consistent with intrinsically oblique magnetic fields whose observed directions are altered by relativistic aberration.