Quads to doubles: a probe of substructure

Over 70 galaxy strong gravitational lenses have been discovered. These occur when light from a faraway source, such as a high redshift quasar, gets bent so much by the matter of a galaxy that it passes en route to earth that multiple images are produced. Lensing depends only on the bending of the path of the light by matter, so it measures the dark matter content of the galaxies in a way that dynamical estimates (which require equilibrium assumptions) cannot. With the CASTLES and CLASS surveys, we are beginning to get extremely useful statistical samples of these intervening galaxies, which have average redshifts around 0.7.
(figure by Campbell, Datta, Gorelik, Herrmann, Troyani, source page here)
A "standard model" of galaxy structure is emerging from analysis of specific lenses which have many images and the sampe of known lenses as a whole. In particular, galaxies which lens seem to have density ~ r^-2 and some flattening (ellipticity).

Puzzle

Within this picture a puzzle has emerged over the last few years. Most lensed systems have either two or four images (doubles and quads respectively).
Here are two examples:


the "quad" B1422, a redshift 0.3 galaxy (source, CASTLES)	the "double" B2114, another redshift 0.3 galaxy (source, CASTLES)

The number of images found depends upon the properties of the galaxy and the position of the source, if the light from the source passes by the galaxy in certain regions of space a quad will arise, while in other regions a double will arise. Below is a picture of regions near a galaxy which will produce quads or doubles. The galaxy is taken to have the average amount of ellipticity or flattening.

Regions behind a galaxy and image multiplicities:

A source behind regions in RED produces a "double"

A source behind regions in YELLOW produces a "quad"

The galaxy is centered in the middle of the plot and is flattened in the vertical direction, with ellipticity 0.3.

Using this standard model and taking into account biases, one can predict the relative numbers of quads to doubles found in a homogeneous sample, such as the CLASS survey, which was completed recently. The expected number is about 1 quad for every 4 doubles, while the actual observed number is closer to 2 to 5, ie about twice as many quads as expected.

A way out: substructure predicted by hierarchical structure formation

The Cold Dark Matter paradigm of structure formation begins with small objects clumping together to form larger ones which in turn clump together to form even larger ones. Thus, near and within any structure there will be smaller mass objects, which can also lens. Thus instead of considering an isolated galaxy, one should include the effects of lensing from the combined system. Below is an example of what happens to multiplicites when the galaxy given above has another galaxy "nearby".

Regions behind a galaxy and image multiplicities:

A source behind regions in RED produces a "double"

A source behind regions in YELLOW produces a "quad"

A source behind regions in GREEN produces a "five"

The position of the satellite and the galaxy are given in blue. The galaxy above now has another galaxy "nearby" (i.e. somewhere between us and its source), whose mass is about 2/3 of that of the larger galaxy.

Solution?

Adding up the contributions of having a nearby galaxy can increase the number of quads significantly, getting close to the doubling needed in order to match the data (J. Cohn and C. Kochanek, astro-ph.0306171).
This research was supported in part by NSF-AST-0205935. Opinions stated here do not reflect those of NSF.