A VLBI investigation of the discrepant image flux-density ratio in the gravitational lens JVAS B0218+357

Abstract

B0218+357 is a doubly-imaged gravitational lens with the smallest known image angular separation (334 mas) amongst the galactic-scale lens systems. Besides the two compact images with flat-spectra, A and B, it has a steep-spectrum Einstein ring of a similar diameter. This work focuses on the observed systematic decline in the image flux ratio (A/B) with decreasing frequency (from ~ 4 at 15 GHz to ~ 2 at 1.7 GHz), which is not in compliance with the achromatic behaviour of gravitational lensing. In my work, I present various scenarios that can lead the observed image flux-densities to differ from those predicted by the lens model. These scenarios were tested based on multi-frequency and phase-reference VLBI observations of B0218+357.<br /> One of the possible explanations is a frequency-dependent source structure, combined with the relative image-magnification which changes significantly over the extent of the structure. I have shown that the centroids of the image brightness distributions do not reveal sufficiently large frequency-dependent shifts to account for the changing image flux ratios. A detailed examination of the interaction of the image flux density distributions with the magnification from lens models can also not explain the effect.<br /> A frequency-dependent source size could in principle interact with mass substructure in the lens but a simple model of this cannot reproduce the observed frequency trend of the image flux ratio. Refractive scattering by a screen only partially covering image A might reproduce the observed anomalous frequency-dependent flux density ratio, but there is little evidence that image A suffers strong scattering compared with image B. However, free-free absorption in a dense molecular cloud along the path of image A, known to be present from radio spectroscopic observations, can indeed reproduce the difference between the radio spectra of images A and B. I have shown that assuming an HII region present in front of image A, there is an excellent agreement between the predicted free-free absorption curve and the observed spectrum of image A, thus providing a plausible explanation for the image flux ratio anomaly seen in B0218+357.