Dark Matter

Galaxy Lensing:

We also probe the dark matter halos of galaxies through gravitational lensing: for the small scale, central mass density constraints we analyze the strong lensing effect. In some lensing systems the lensed sources have HST-resolvable surface brightness substructure and are so extended that their images cover a fair fraction of the Einstein circle. Their images can then be used to study their relative surface brightness mapping and to obtain upper limits on the dark matter substructure, eg., on the subhalo fraction above a mass of 10^7 solar masses. We have studied such a case in Bauer et al., submitted , and showed (at a 2 sigma level) that the lens contains less than 4 percent of mass in substructure within the Einstein-cylinder (for a CDM subhalo mass function slope of about 1.9).

At the same time one can use these extended source systems and also systems with several distinct sources at the same redshift to measure the steepness of the mass density profile around one effective radius. In Grillo et al.2010 we identified and analyzed such a "golden lens" system.

The galaxies' outer mass profile (scales of 20kpc to a few 100 kpc) and, more general, the galaxy-mass correlation function (scales up to several couple of 100kpc) can be obtained from the galaxy-galaxy weak lensing effect. We study galaxies in the CFTHLS-Wide survey for that, with redshifts between 0 and 1 and apparent I-band magnitudes brighter than about 24 mags. We show results of this ongoing analysis here.

Most of the even brightest stars in nearby galaxies can not be resolved even with HST because they are heavily crowded and the flux of them is small compared to the sum of the fluxes of all stars in the same resolution element (PSF). However, if stars vary in time and if their variation is strong enough to exceed the photon noise in the same resolution element, their variation can be identified with the difference imaging method. This difference imaging method can be used to find intrinsically variable stars and stars where the variability is induced by the microlensing effect. Microlensing events are achromatic and have a unique light curve which should make discrimination from intrinsically variable stars possible, if sampled well enough. The microlensing effect itself can be caused by stars in the foreground of stars ("self-lensing") or by compact dark matter "machos" in the haloes of galaxies. Thus a monitoring of nearby galaxies can provide upper limits to the compact dark matter fraction in haloes. We have performed such a monitoring for about 10 years towards M31 and show events found. We also demonstrate that the analysis of individual events like WECAPP GL1 can already point to a halo lensing event, provided that the stellar properties (size, colors, brightness, spatial distribution and velocities) of stars in M31 are well understood. This understanding can be improved by investigtions of M31 as detailed in the dynamical/upper part of this webpage (see also Montalto et al. 2009 for our recent determination of the dust properties in M31). 

Dark Matter: Lensing by Clusters of galaxies and Large Scale Structure

Deep multi-color imaging of the cluster A1689 () with ACS on board of HST has tremendously improved our understanding of the central mass distribution of this cluster: this regards the concentration, mass density slope, stripping of galaxy haloes in dense enviromennts and dark matter substructure (See, e.g., Broadhurst et al. 2005, Halkola, Seitz, Pannella, 2006 & 2007 ).

In a joint effort lead by M. Postman about 20 scientists have applied in a Multi-Cycle-Treasury proposal to map about 25 clusters deeply in 14 filters spanning the near-UV to near-IR. The time was rewarded, observations will start with cycle 18 in September 2010. Our major science goals are to map the cluster mass distribution with strong lensing and weak lensing (supported from ground) and to identify the most distant galaxies at z>7.       

QSOs are magnified by the gravitional lens effect of matter in their foreground. This leads to the well known QSO-galaxy associations for a flux-limited, bright QSO sample.

At the same time most of the QSOs vary with time. Large area surveys with high cadence like Palomar Quest (see Bauer et al. 2009a ) allow to investigate the variability as a function of observation time lag, QSO redshift, observation wavelength, and the bolometric luminosity and other parameters of the QSO-host galaxy (black hole mass estimate as obtained from spectral lines). One can normalize out the dependence on redshift, mass and time lag and is left with a linear relation between the logarithms of the normalized variability and luminosity. This relation is shown in the left figure. It was derived from about 300000 thousand flux measurement pairs for 5000 QSOs with SDSS spectroscopy. Our idea now is (Bauer, Seitz et al. submitted) to interprete this relation as a flux-standard candle: A variability estimate of quasars provides a (noisy) luminosity estimate, which after comparison with the true luminosity yields an estimate for the gravitational magnification of the QSO. We indeed have shown that QSOs that are predicted to be magnified by foreground galaxy clusters and groups are also estimated to be magnified by a similar amount according to our new method. This could develop into be an alternative method to analyze the lensing effect of large scale structure in e.g. the LSST era.

The analysis of the gravitational distortion of galaxy shapes by foreground dark matter is one of the most promising tools to constrain the properties of dark matter and dark energy. The feasibility depends on having sufficiently deep and large area data, such that shapes of many galaxies can be analyzed. The most crucial point, however, is that the shape measurement biases present in all shape measurement methods (see Bridle et al., submitted, 2009arXiv0908.0945 , or Fig.5 of Gruen et al. , submitted, astro-ph 1002.0838 ) have mostly to be eliminated. Currently shear biases are decreased by introducing multiplicative shear-calibration-factors derived from simulated galaxy "data". Since these simulations have to be done anyhow to investigate and reduce biases, we instead propose in Gruen et al., submitted, astro-ph 1002.0838, to use these simulated data directly for training neural networks for gravitational shear estimates. These neural networks can then indeed account for biases that depend on the source properties (light profile, signal to noise, size & PSF-properties) in a more complex way.

Recent papers