Dark Matter: Strong Gravitational Lensing

Science Goals / Science Portfolio / Dark Matter / Dark Energy / Solar System / Transient / Milky Way

Credit: X-ray: NASA/CXC/M.Markevitch et al.
Optical: NASA/STScI; Magellan/U.Arizona/D.Clowe et al.
Lensing Map: NASA/STScI; ESO WFI; Magellan/U.Arizona/D.Clowe et al.


Strong gravitational lensing happens when there is so much mass contrast in the lens that the light rays from a distant source form multiple images. This was first seen in a quasar lensed by a galaxy in 1979. More commonly, the huge dark matter concentrations in clusters of galaxies create typical bending angles of 30 arcseconds, and multiple highly distorted images of a fortuitously aligned background source galaxy.

Illustration of strong lensing.

Download a lensing simulation [MPEG 0.8MB] (or Full resolution version [MPEG 7.3MB] LENS0.mpg [0.8 MB] Full resolution version [10.4 MB]

Two simulations of strong lensing by a massive cluster of galaxies. On the left, all the dark matter is clumped around individual cluster galaxies (orange), causing a particular distortion of the background galaxies (white and blue). On the right, the same amount of mass is more smoothly distributed over the cluster, causing a very different distortion pattern. Clicking on each image will bring up MPEG movies (800 kB) showing the evolution of the distortion as the clusters move against the background over half billion years. A full description of the simulation process is available. Courtesy J. A. Tyson, UC Davis.

The Physics of Dark Matter via Strong Lensing

The LSST will obtain thosands of images of each patch of sky covering 20,000 square degrees, integrating to 26.5 AB mag in 6 bands covering uv to IR (ugrizy). Photometric redshifts will be available for the ~3 billion detected galaxies. The data set will provide deep multicolor photometry and variability monitoring. One of the many strengths of the LSST will be its ability to use strong gravitational lensing to study dark matter distributions on galaxy and cluster scales.

Mass distribution in the cluster CL0024. This detailed dark matter distribution can be used to constrain theories of dark matter. Strong lensing of a background galaxy was inverted to yield a model for the mass distribution.

Cluster scale Lenses

The all-visible sky survey by LSST will enable sensitive probes of the dark matter distribution on a wide range of physical scales. The unprecedented combination of depth and area provided by the LSST will be exploited to find rare alignments, such as clusters in which background sources are lensed into multiple images. By sampling the gravitational potential at several radii in these systems, the LSST imaging will allow accurate, high-angular resolution reconstructions of cluster mass distributions. Our simulations predict that the LSST dataset will provide at least an order of magnitude increase in the number of such systems known. Other rare lensed image configurations will provide important insights into cosmography and source astrophysics. These include multiply-imaged supernovae, multiple-plane lensing, and unusual strong lenses with higher-order catastrophes.


Redshift distribution of lenses (left) and sources (right) that produce giant arcs in clusters. The number densities are for a 20000 square degree survey; the lensing cross-section was approximated following Dalal et al 2004, and the HDF faint galaxy counts (Casertano at al 2000) were used for the source population.

Integrated arc number counts as a function of magnitude limit. LSST will reach magnitude ~24 in each exposure, and ~27 in the final stacked image.

CL0024 as seen by the Hubble Space Telescope and as it would be seen in one year's stacked image from the LSST survey (assuming 0.7 arcsec seeing). The multiple images of the background galaxy are clearly visible.

Galaxy-scale Lenses

In addition to the rare cluster lenses that will be found in the LSST survey, the survey images will produce at least an order of magnitude increase in galaxy-scale lenses. Our simulations predict that the final stacked data set will contain approximately 5000 detectable cases of a background galaxy being multiply-imaged by a foreground system. In addition, we predict that there will be ~150 systems in which the lensed objects are AGN or quasars and the lens system can be identified in a one-epoch image with an integration time of 20 seconds (assuming seeing of 0.7 arcsec). This number increases to ~1500 systems if the seeing is 0.4 arcsec. More generally, with shapes and redshifts of billions of source galaxies LSST will measure the compact dark matter distribution on these scales with precision.

The source population numbers to the survey limiting magnitude may be estimated from the Hubble Deep Field (HDF), suggesting some 3 x 105 galaxies per square degree at z > 1 (Metcalfe et al. 2001, Fernandez-Soto et al. 1999). We may reasonably expect of order 107 multiple image systems to be present in the survey, using a lensing rate of 10-3 as found in the CLASS survey (Browne et al. 2003); however, only a fraction of these lenses will be identified by LSST alone. The lensing cross-section is dominated by massive elliptical galaxies at redshifts 0.3 < z < 1 (e.g. Fukugita & Turner 1991, Blandford et al. 2001); again from the HDF, we may expect approximately 10,000 such "clean lens" galaxies per square degree, providing about 30 square degrees of lensing cross-section in the whole survey. By targeting these ellipticals and searching for achromatic excesses, a substantial fraction of these lenses may be detected; note the great importance of multi-color LSST imaging in this task. The resulting large sample of wide separation lenses will allow high precision statistical tests of the level of small-scale and non-axisymmetric structure in galaxies and groups.

The prospect of discovering a significant number of higher-order catastrophe lenses in the LSST sample is an exciting one: as an example, the "quintuple quasar" lens system has six lensed images, with the lens model predicting two more (Winn et al. 2003, Keeton & Winn 2003). Such multiple image systems will provide much information on lens galaxy structure, while the very high magnifications attainable will provide us with a very powerful "cosmic telescope." The LSST optical data on sources observed in this way will provide very important complementary information to that available in similar scale low cadence surveys across the rest of the electromagnetic spectrum, including those by EXIST in the X-ray band and the Square Kilometer Array in the radio.

Mass in Three Dimension

A glimpse of a universe of mass. Shown here is a map of mass obtained by gravitational lens mass tomography. This 2x2 degree field of mass, obtained in 15 hours of 4-meter telescope exposures in the Deep Lens Survey, would fit easily inside LSST's single snapshot field of view.

LSST will find hundreds of thousands of these massive clusters in a stunning 3-D view of the universe of mass extending over 20,000 square degrees of sky and back to half the age of the universe. Sharp constraints on cosmology and the physics of dark energy will result.
(cf. Wittman et al. 2003 ApJ 579, 218)

The faintest galaxies have a range of colors, each one's color depending on its type and its distance from us. The most distant galaxies have their spectra shifted to longer, redder wavelengths by the Hubble expansion, and their light has taken up to ten billion years to travel to us. Using the colors of the galaxies, it is possible to gauge the distance to the background galaxies. Mirages also rely on distance. This is the clue that unlocks the universe of mass in three dimensions; the more distant the source, the more warped its image. If there is a foreground mass, the mirage effect on the background galaxies is stronger for more distant galaxies. Mass tomography images [PPT]

By measuring both the warp and the distances to the background galaxies, it is possible to reconstruct the mass map and also to place the mass at its correct distance. This enables the exploration of mass in the universe, independent of light, since only the light from the background galaxies is used. By exploring mass in the universe in three dimensions we are also exploring mass at various cosmic ages. This is because mass seen at great distance is mass seen at a much earlier time. So we can chart the evolution of dark matter structure with cosmic time.

Surveying the numbers of cosmic mass clusters in our universe will ultimately lead to precision tests of theories of dark energy. To fully open this novel window of the three-dimensional universe of mass history, we need a new telescope and camera very unlike what we have now. We need LSST. Advances in technology have equipped us to mine the distant galaxies for data — in industrial quantity. LSST's wide-angle gravitational lens survey will generate millions of gigabytes of data and intriguing opportunities for unique understanding of the development of cosmic structure. Our challenge is twofold. These galaxies are faint, and we need to capture images of billions of them. LSST's combination of large light-collecting capability and unprecedented field of view will for the first time open this unique window on the physics of our universe. LSST will provide a wide and deep view of the universe, allowing us to conduct full 3-D mass tomography to chart not only dark matter, but the presence and influence of dark energy.

Do we trust our current view of the universe? Combining these results with other cosmic probes will lead to multiple tests of the foundations of our model for the universe. What will our concept of the universe be when those answers are in? Perhaps the most interesting outcome will be the unexpected; a clash between different precision measurements might prove to be a hint of a grander structure, possibly in higher dimensions. LSST provides that opportunity.

Cluster Mass Tomography: Technique of 3-D weak lens mass tomography.

Tomographic view of the W2001 cluster. Redder areas indicate higher density, with most bluer areas being in the noise. No significant structures are seen in the mass map made from sources at z < 0.3 (left). The cluster appears prominently in the lower left of the mass map made from sources z > 0.3 (right). The field is about 40' on a side. LSST will survey at least three orders of magnitude more area, and enable finer slicing of the redshift distribution, which will provide a detailed picture of the growth of structure with cosmic time.
(Tyson 2000 Physica Scripta T85, 259, Wittman etal ApJ 557 L89 2001, Wittman etal SPIE 4836, 21 (2002)).

The cluster found by 3-d mass tomography was DLS CL 1055-05 at redshift 0.68
(Wittman etal 2003 ApJ 597, 218).

CL 1055-05 optical image showing the galaxies. Combining mass tomography with images of the galaxies yields information on the mechanisms of cosmic structure formation.