2012ApJ...757..116F

We measure apparent velocities (v_app) of absorption lines for 36 white dwarfs (WDs) with helium-dominated atmospheres–16 DBAs and 20 DBs–using optical spectra taken for the European Southern Observatory SN Ia progenitor survey. We find a difference of 6.9±6.9 km/s in the average apparent velocity of the Hα lines versus that of the He I 5876 Å lines for our DBAs. This is a measure of the blueshift of this He line due to pressure effects. By using this as a correction, we extend the gravitational redshift method employed by Falcon et al. to use the apparent velocity of the He I 5876 Å line and conduct the first gravitational redshift investigation of a group of WDs without visible hydrogen lines. We use biweight estimators to find an average apparent velocity, < v_app > _BI, (and hence average gravitational redshift, < v_g > _BI) for our WDs; from that we derive an average mass, < M > _BI. For the DBAs, we find < v_app > _BI = 40.8±4.7 km/s and derive < M > _BI= 0.71^+0.04_- 0.05_M☉_. Though different from < v_app > of DAs (32.57 km/s) at the 91% confidence level and suggestive of a larger DBA mean mass than that for normal DAs derived using the same method (0.647^+0.013_- 0.014_M☉_; Falcon et al.), we do not claim this as a stringent detection. Rather, we emphasize that the difference between < v_app > _BI of the DBAs and < v_app > of normal DAs is no larger than 9.2 km/s, at the 95% confidence level; this corresponds to roughly 0.10 M_☉. For the DBs, we find < v ^He_app > _BI= 42.9±8.49 km/s after applying the blueshift correction and determine < M > _BI= 0.74^+0.08_- 0.09_ M_☉. The difference between < v ^He_app > _BI of the DBs and < v_app > of DAs is ≤ 11.5 km/s (∼0.12 M_☉), at the 95% confidence level. The gravitational redshift method indicates much larger mean masses than the spectroscopic determinations of the same sample by Voss et al. Given the small sample sizes, it is possible that systematic uncertainties are skewing our results due to the potential of kinematic substructures that may not average out. We estimate this to be unlikely, but a larger sample size is necessary to rule out these systematics.