Chandra follow-up of the Hectospec Cluster Survey: comparison of caustic and hydrostatic masses and constraints on the hydrostatic bias

Context. Clusters of galaxies are powerful probes with which to study cosmology and astrophysics. However, for many applications, an accurate measurement of a cluster's mass is essential. A systematic underestimate of hydrostatic masses from X-ray observations (the so-called hydrostatic bias) may be responsible for tension between the results of different cosmological measurements.Aims. We compare X-ray hydrostatic masses with masses estimated using the caustic method (based on galaxy velocities) in order to explore the systematic uncertainties of both methods and place new constraints on the level of hydrostatic bias.Methods. Hydrostatic and caustic mass profiles were determined independently for a sample of 44 clusters based on Chandra observations of clusters from the Hectospec Cluster Survey. This is the largest systematic comparison of its kind. Masses were compared at a standardised radius (R-500) using a model that includes possible bias and scatter in both mass estimates. The systematics affecting both mass determination methods were explored in detail.Results. The hydrostatic masses were found to be systematically higher than caustic masses on average, and we found evidence that the caustic method increasingly underestimates the mass when fewer galaxies are used to measure the caustics. We limit our analysis to the 14 clusters with the best-sampled caustics where this bias is minimised (>= 210 galaxies), and find that the average ratio of hydrostatic-to-caustic mass at (500) is M-500,M-X/M-500,M-C = 1.12(-0.10)(+0.11).Conclusions. We interpret this result as a constraint on the level of hydrostatic bias, favouring small or zero levels of hydrostatic bias (less than 20% at the 3 sigma - level). However, we find that systematic uncertainties associated with both mass estimation methods remain at the 10-15% level, which would permit significantly larger levels of hydrostatic bias.