2011MNRAS.412..800S

The caustic technique uses galaxy redshifts alone to measure the escape velocity and mass profiles of galaxy clusters to clustrocentric distances well beyond the virial radius, where dynamical equilibrium does not necessarily hold. We provide a detailed description of this technique and analyse its possible systematic errors. We apply the caustic technique to clusters with mass M₂₀₀≥ 10¹⁴ h^–1 M_☉ extracted from a cosmological hydrodynamic simulation of a ΛCDM universe. With a few tens of redshifts per squared comoving megaparsec within the cluster, the caustic technique, on average, recovers the profile of the escape velocity from the cluster with better than 10 per cent accuracy up to r∼ 4r₂₀₀. The caustic technique also recovers the mass profile with better than 10 per cent accuracy in the range (0.6-4) r₂₀₀, but it overestimates the mass up to 70 per cent at smaller radii. This overestimate is a consequence of neglecting the radial dependence of the filling function {img}. The 1σ uncertainty on individual escape velocity profiles increases from ∼20 to ∼50 per cent when the radius increases from r∼ 0.1r₂₀₀ to ∼4r₂₀₀. Individual mass profiles have 1σ uncertainty between 40 and 80 per cent within the radial range (0.6-4) r₂₀₀. When the correct virial mass is known, the 1σ uncertainty reduces to a constant 50 per cent on the same radial range. We show that the amplitude of these uncertainties is completely due to the assumption of spherical symmetry, which is difficult to drop. Other potential refinements of the technique are not crucial. We conclude that, when applied to individual clusters, the caustic technique generally provides accurate escape velocity and mass profiles, although, in some cases, the deviation from the real profile can be substantial. Alternatively, we can apply the technique to synthetic clusters obtained by stacking individual clusters: in this case, the 1σ uncertainty on the escape velocity profile is smaller than 20 per cent out to 4r₂₀₀. The caustic technique thus provides reliable average profiles which extend to regions difficult or impossible to probe with other techniques.