Properties of galaxy dark matter halos from weak lensing

We present the results of a study of weak lensing by galaxies based on 45.5 deg(2) of R-C-band imaging data from the Red-Sequence Cluster Survey (RCS). We define a sample of lenses with 19.5 < R-C < 21 and a sample of background galaxies with 21.5 < R-C < 24. We present the first weak-lensing detection of the flattening of galaxy dark matter halos. We use a simple model in which the ellipticity of the halo is f times the observed ellipticity of the lens. We find a best-fit value of f = 0.77(-0.21)(+0.18) , which suggests that the dark matter halos are somewhat rounder than the light distribution. The fact that we detect a significant flattening implies that the halos are well aligned with the light distribution. Given the average ellipticity of the lenses, this implies a halo ellipticity of (e(halo)) = 0.33(-0.09)(+0.07) , in fair agreement with results from numerical simulations of cold dark matter. We note that this result is formally a lower limit to the flattening, since the measurements imply a larger flattening if the halos are not aligned with the light distribution. Alternative theories of gravity ( without dark matter) predict an isotropic lensing signal, which is excluded with 99.5% confidence. Hence, our results provide strong support for the existence of dark matter. We also study the average mass profile around the lenses, using a maximum likelihood analysis. We consider two models for the halo mass profile: a truncated isothermal sphere (TIS) and a Navarro-Frenk-White (NFW) profile. We adopt observationally motivated scaling relations between the lens luminosity and the velocity dispersion and the extent of the halo. The TIS model yields a best-fit velocity dispersion of sigma = 136 +/- 5 +/- 3 km s(-1) ( all errors are 68% confidence limits; the first error bar indicates the statistical uncertainty, whereas the second error bar indicates the systematic error) and a truncation radius s = 185(-28)(+30) h(-1) kpc for a galaxy with a fiducial luminosity of L-B = 10(10) h(-2) L-B,L-circle dot (under the assumption that the luminosity does not evolve with redshift). Alternatively, the best-fit NFW model yields a mass M-200 = (8.4 +/- 0.7 +/- 0.4) x 10(11) h(-1) M-circle dot and a scale radius r(s) = 16.2(-2.9)(+3.6) h(-1) kpc. This value for the scale radius is in excellent agreement with predictions from numerical simulations for a halo of this mass.