The 6dF peculiar velocity survey

Abstract

We present the 6dF peculiar velocity survey (6dFGSv), the largest ever catalogue of galaxy peculiar velocities by a factor of two. This catalogue contains all new spectro-photometric data, including redshifts, galaxy central velocity dispersions ({u03C3}), NIR surface photometry (Re, ({u03BC}e)), and Fundamental Plane (FP) distances and peculiar velocities for 11287 early-type galaxies to a depth of cz = 16500 kms^-1 . All spectral observations were taken as part of the Six Degree Field Galaxy Survey (6dFGS), using the Anglo-Australian Observatory's 6dF multi-fibre spectrograph system at the UK Schmidt telescope, covering the southern sky with [b] > 10{u00B0}. Photometry was based on images and magnitude data in the 2MASS Extended Source Catalogue. Velocity dispersions were calculated from a spectral range of 4000A to 5570A using the Fourier cross-correlation technique. A new technique to correct for the effects of variable instrumental resolution was developed, and shown to improve the accuracy and precision of {u03C3}. The 6dFGSv sample is limited to galaxies with {u03C3} > 100 km s^-1, resulting in median values of S/N= 12. 9{u00C5}^-1, {u03C3} = 163km s^-1, and random errors of 12.7%. A new reduction process was applied to 2MASS image tiles, in order to derive seeing corrected surface photometry for 6dFGSv galaxies in J, H, and Ks bands. The derived parameters were remarkably robust and well behaved, especially considering the eight second observations of 2MASS. The combined FP photometric parameter XFP =log Re -0.33({u03BC}e) has a median random error of 3%, which is negligible compared to the intrinsic scatterof the FP. A sample of 102 known groups and clusters were identified within the 6dFGSv sample, and were used along with high S/N galaxy subsamples to provide global fits of the NIR FP. The fitting was done using a maximum likelihood method, designed to efficiently determine the best inverse fit with a minimum level of bias. Two representative fits of the FP were selected for all subsequent estimations of distances: Case I - log {u03C3} = 0.689XFP+1.581; and Case II - log {u03C3} = 0.643XFP+1.610. Case I was determined from the fit of the best 35 groups, while Case II represented the fit to the largest subsample of almost 8000 galaxies with S/N> 10. The differences in the slope coefficient between the samples was larger than the {u00B1}1{u03C3} errors. It was supposed that this was evidence of the FP being a bent surface, as indicated by some recent studies. Hence, in spite of the homogeneous nature of the data and the high precision of the 6dF NIR FP fits, an increase in slope of the FP with wavelength could not be confirmed. Distances were derived for all groups relative to the Coma Cluster, and after correction for homogeneous Malmquist bias, were rescaled to leave a zero net flow. The rescaling of the zero-points resulted in a peculiar velocity for Coma of ~ 400 kms^- 1, consistent with previous surveys. The rescaled distances from both cases resulted sensibly in essentially zero peculiar velocity forclusters in the Shapley Supercluster Complex (SSC), appropriate for such a massive overdensity. Comparisons of group distances with independent determinations in previous surveys showed impressive agreement, with no sign of bias. Distances to individual galaxies were then calculated using the rescaled zero-point for Coma. Peculiar velocities are calculated in the CMB frame for groups and galaxies, with sufficient correspondence between 6dFGSv group results and the literature to suggest the results are reasonable. The mean peculiar velocity determined from the total sample of galaxies, vpec^- = 254 km s^-1 , is in agreement with the scale of velocities suggested by hierarchical merging scenarios. An initial investigation was made of the source of the Local Group CMB dipole, utilizing group peculiar velocities in the direction of the Great Attractor/SSC. Comparisons with simple one and two attractor flow models, as well as with SN Ia distances and Tully-Fisher distances, confirmed the local streaming motions are best fit with a two attractor model. Preliminary indications of backside infall into the Great Attractor seen in the 6dFGSv data however, suggest the influence of the SSC to be significantly less than that of the Great Attractor at the Local Group.