Cluster Lensing and Supernovae Survey with Hubble (CLASH)

Marc Postman,

The composition of our universe is most intriguing. It is “dark,” with 23% of the mass-energy density made up of weakly interacting (and as-yet-undetected) non-baryonic particles (a.k.a. dark matter, DM), and 73% due to the as-yet-unknown physics that accelerates the expansion of the universe (a.k.a. dark energy). To shed new light on these mysteries, we coupled Hubble’s panchromatic imaging capabilities—the Wide Field Camera 3 (WFC3) and Advanced Camera for Surveys (ACS)—with the gravitational-lensing power of 25 massive galaxy clusters, to test models of the formation of cosmic structure and to probe the high-redshift universe with unprecedented precision. Our 524-orbit Multi-Cycle Treasury program, dubbed CLASH (Postman et al. 2012), successfully completed all its observations in July 2013.

The four primary science objectives of the CLASH program are to (1) map the distribution and characterize the nature of dark matter in galaxy clusters, (2) detect type Ia supernovae (SNe Ιa) out to redshift z ~ 2, to constrain the dark energy equation of state, measure the rate at which they occur, and constrain how their light curves may evolve with time (important for understanding if SNe Ιa are equally good standard candles over a wide range in redshift), (3) detect and characterize the most distant galaxies, focusing especially on the epoch when the universe was less than 800 million years old, and (4) characterize the internal structure and evolution of galaxies in and behind the CLASH clusters.

The mass profiles and central concentrations in CLASH clusters

Gravitational lensing enables precise maps of the distribution of mass in lenses, which are galaxy clusters in the case of CLASH. Lensing also enables high-resolution studies of background galaxies, which are sometimes referred to as arcs due to their curved appearance when they are magnified. The two main limitations on the accuracy of the derived mass distributions are the number of sources identifiable as having multiple images and the number of such sources with reliable redshifts. With deep Hubble imaging in 16 broadband filters, ranging from the ultraviolet (UV; 0.225 microns) to the near infrared (NIR; 1.6 micron), CLASH was designed to increase both numbers.

Prior to CLASH, Hubble observations demonstrated that many strongly lensed galaxies are fainter than 25.5 AB magnitude, which puts them beyond the reach of spectrographs on ground-based telescopes. CLASH’s 16-band photometry provides both deep imaging—the 5σ point-source limit is ~27 AB mag—and accurate photometric redshifts, with errors ~0.03 (1 + z) for many arcs (see Jouvel et al. 2013).

Figure 1 shows a montage of images centered on some of the more prominent arcs seen in CLASH clusters. Figure 2 shows a CLASH cluster with 47 multiply lensed arcs from 12 unique sources, which span the range 1.1 ≤ z ≤ 5.8. Prior to CLASH, only 3 multiply lensed sources were known in this cluster, all from a single background galaxy at z = 1.03.


To map the dark-matter distribution in a cluster, we combine weak-lensing data, primarily from multi-band images taken with the Subaru Prime Focus Camera (Suprime Cam), with strong-lensing data from Hubble. Weak lensing is measured mostly at large radii from the cluster, where the distortions of the shapes of background galaxies are detected statistically. In this regime—Suprime Cam covers a region around a cluster at least 5 Mpc across—the galaxies are not multiply lensed but only slightly distorted. The combined information from Subprime Cam and Hubble allows us generate a complete, two-dimensional picture of each galaxy cluster.

One of the predictions of the cold-dark-matter (CDM) model of structure formation is that there is a relationship between the compactness (or concentration) of the cluster’s central distribution of matter and the virial mass of a cluster-scale dark matter halo. A halo is the overdense region containing the dark matter that, in this case, generates the cluster’s gravitational potential well. Prior to CLASH, several studies found that clusters were significantly over-concentrated, in some cases by factors of 2 or more when compared to the predictions of numerical simulations. These findings caused tension between the observations and the theory. Nevertheless, it was not clear if this tension indicated problems with CDM theory or a selection bias in the data, because the best-studied clusters were often known to be the strongest lenses, and such systems may have systematically higher central-mass concentrations. In CLASH, 20 out of the 25 clusters were selected by their X-ray properties rather than their lensing characteristics. Specifically, the X-ray selected clusters display relatively smooth spherical distributions of hot gas. By choosing potentially relaxed systems at late stages of structural evolution, our sample for testing the CDM paradigm was far less biased.

CLASH-3CLASH-4Figures 3 and 4 show the preliminary relation between mass and concentration for 12 of 20 CLASH clusters selected by X-ray. The relation agrees with the expectations from CDM-based numerical simulations in a spatially flat universe with a sub-critical density of matter. Indeed, after analysis with the full suite of CLASH optical and X-ray data, one of the previously more discrepant clusters, Abell 2261, now agrees with predictions (Coe et al. 2012). The earlier tensions between observed and modeled cluster-mass profiles now seem far less significant (Merten et al. 2014).

CLASH data also allow us, for the first time, to directly test if the dark-matter fluid is pressureless. In General Relativity, a gravitational potential well is shaped by its mass-energy content. However, in that well different types of particles respond to the gravitational potential, Φ(r), in different ways (e.g., Faber & Visser 2006). Dark matter is the dominant component that determines the gravitational potential well of a cluster and if the equation of state of the dark matter has a pressure term, then relativistic particles (photons) and non-relativistic particles (galaxies) flowing in a cluster will yield different mass profiles. In particular, the equation of state (EoS) is w = [pr(r) + 2pt(r)] / [c23r(r)], where pr(r) and pt(r) are the radial and tangential pressure profiles. We have obtained a constraint on the amplitude of the pressure terms in the CLASH cluster MACS J1206.2-0848, which has over 600 spectroscopically confirmed cluster members. That constraint is w = –0.003 ± 0.137 (statistical error) ± 0.170 (systematic error), which is consistent with zero. Because the baryonic content of clusters is only ~15% of the total mass, the above EoS constraint is primarily on the DM fluid (Sartoris et al. 2013).