Domain Wall Mobility in Single-Crystal Yttrium Iron Garnet

Reasonable agreement between the damping constant inferred from the resonance linewidth and that obtained from the domain wall mobility has previously been found for ferrites. However, the recent linewidths reported for yttrium iron garnet (YIG) by Spencer, LeCraw, and Clogston are substantially narrower than those for ferrites. On using this narrow linewidth to calculate the YIG wall mobility, one obtains about ${10}^6\hspace{0.17em}\mathrm{cm}/\sec \hspace{0.17em}\mathrm{oe}$ . We have measured the YIG wall mobility as a function of temperature and find it to be less than ${10}^4\hspace{0.17em}\mathrm{cm}/\sec \hspace{0.17em}\mathrm{oe}$ . This discrepancy can be interpreted in terms of a model for domain wall motion which is based on a type of energy conversion process different from that usually assumed. The customary treatment requires that the domain wall dissipate entirely to the lattice all of the magnetostatic energy stored in a given region during the time that the wall passes through this region. For small damping such as exists in YIG, this requirement appears to be unreasonable. The wall motion model proposed here allows the wall to be followed by a wake in which the original magnetostatic energy is temporarily stored as exchange, anisotropy, and demagnetizing energy. The measured wall velocities in single-crystal YIG are in reasonable agreement with quantitative estimates based on this model.