Why is the domain wall mobility in bulk yttrium iron garnet so low?

The domain wall mobility observed in bulk samples of yttrium iron garnet is about 3 × 10⁵ms^-1T^-1, whereas simple phenomenological theory predicts, from the ferromagnetic resonance linewidth, 3 × 10⁵ms^-1T^-1. The following explanation has been proposed. The critical velocity (∼ 300 ms^-1for YIG) which a wall cannot exceed would be attained in a drive field of only 10^-6T if the predicted mobility were realised, but in such low fields the coercivity obscures wall velocity measurements. The low mobility quoted was observed in fields above 10^-5T and this might involve motion in an anomalous low-mobility mode. We have tested this proposal by (1) a careful search for high mobility wall motion in low fields (2) looking for high mobility in a sample in which the predicted mobility depends upon temperature and is intermediate between the values quoted above (3) measuring the velocity versus field relationship up to close to the critical velocity. We found no evidence for high mobility motion, or that the normal velocity-field relations are due to an anomalous mode. A non-linear region in the velocity field curve, which is observed only for specimens in which the mobility is lower than qpredicted, may be significant. But we have no explanation to replace that which our experiments refute.