# Gravitational Lens VLBI using the Arecibo Telescope

Porcas (MPIfR, Germany), Garrett (JIVE, Netherlands), Patnaik (Canada) & Nair (RRI, India) have attempted to make a definitive'' high-resolution study of the gravitational lens system, 2016+112, using an 18-station Global VLBI Array run at 5 GHz, including the Arecibo 305-m telescope.

This gravitational lens system has always been something of a puzzle. There are 3 image regions'' (A, B & C) spread over an area of 4 arcsec.  Previous VLBI observations (e.g., Garrett et al., 1996, MNRAS, 279, L7) reveal that A and B are both images of a single radio source, produced by bending and delay of the ray paths as they traverse the intense gravitational field of a massive galaxy a few arcsec to the south. Region C breaks up into at least 4 distinct components, aligned along a gently curving arc of length 200 milliarcsec. It is presumed that this region contains the merging images'' of (almost) another 2 images of the same source, making this system a (nearly) 4-image quad'' system.  However, the details of the relationship between images A and B, and their partial counterparts in region C remain unclear. With these VLBI observations (another observation at 1.6 GHz was made on February 25th) this team will shed some light on the matter.

Figure 1: The c-band fringe amplitude for 2016+112 on the
Arecibo-VLBA_Saint Croix baseline. The bandwidth is 16 MHz, the
integration time per point is 1 min, and the two circular polarizations
are averaged. Note the beating between the three image regions.

All 18 stations (VLBA, phased-VLA, Arecibo and 6 stations of the EVN) worked successfully during the observations. The calibration source shows fringes on all 152 baselines, including all 17 baselines to Arecibo for the 2.25 hr for which the source was trackable there (Fig.~\ref{fig117}). There are fringes on the gravitational lens source(s) too, but these are harder to see for two reasons. Firstly, the total source flux density is around 10 times weaker than the calibrator -- ~70 mJy. Secondly, this emission is distributed amongst the 3 image regions, which causes very rapid beating in the visibility function on the long, transatlantic baselines -- and bandwidth smearing too if averaging across the 16-MHz bandwidth.  Thus, to preserve the total 4 arcsec field of view for the mapping process, the data must be kept unaveraged in time (1-s samples) and bandwidth (500-kHz samples) -- and this is not so easily viewed!

Due to the very weak signal expected on long baselines (even those to the highly-sensitive Arecibo), the observations were made in phase-reference'' mode, so that the instrumental residuals'' (clock errors, phase errors, etc.) could be determined from frequent observations of the calibrator, 2029+121.  A duty cycle of 6 or 7 min was used -- a compromise which ensured that Arecibo did not spend most of the cycle driving between the program source and the calibrator, whilst hopefully fast enough to track the wandering instrumental phase throughout the observation.

Figure 2: A very preliminary 5-GHz Global VLBI map of 2016+112, region C.

A first crude mas-resolution image of the gravitational lens has been made from unedited, not-properly-calibrated, blind-phase-referenced data, and is the start of a more laborious process of calibration and editing refinement. Figure 2 is a crude representation of region C, in which one caN see the 4 distinct components of region C - a thin band of emission, stretched by the strong gravitational field of the galaxy lens.