The dependence of free induction decay and echo amplitude of NMR signals in Fe on rf field strength pulse length , and frequency has been determined by an extensive series of measurements The observed variation of free‐induction decay amplitude with and cannot be explained by a rigid oscillation of domain walls, as usually supposed. Such a model fails, even though proper account is taken of enhancement‐factor distribution caused by orientation, and by spatial arrangement of spins perpendicular to the central plane of each wall.
We propose a different model where each domain wall is considered immobile along its periphery. Wall oscillation then resembles that of a drumhead. This modification alters significantly the enhancement‐factor distribution. Furthermore, the average enhancement factor of each wall is proportional to its area. Accordingly, a distribution in wall areas is also postulated. The predictions of this model are in excellent agreement with free‐induction decay spectra.
Enhancement factors of impure samples are smaller than those of pure ones by as much as a factor of three, indicating that the average wall area in such specimens is smaller by a similar factor. This supports the view that impurities augment domain nucleation.
