Two-Phase Ferrites for High-Speed Switching

Abstract
A plot of the inverse switching time vs the applied field yields three regions of different slope for square‐loop ferrites. The low‐field region has the lowest slope and is commonly attributed to domain wall motion. As the field is increased, a break in the curve occurs and the slope increases. This intermediate‐field portion of the curve has been explained by a nonuniform rotational model, and its projection to the field axis is termed the threshold field H 0 for nonuniform rotation. In coincident‐current memory operation, twice the coercive force is the maximum drive which can be employed for switching from one remanent state to the other. Therefore, it is desirable to minimize the extent of the domain wall region so that the coercive force will approximate H 0 as closely as possible. Then 2Hc will correspond to a higher value of the inverse switching time on the nonuniform rotational portion of the curve, and very fast switching can be attained. Prevalent theories describing domain wall motion in the vicinity of nonmagnetic inclusions indicate that domain wall motion can be impeded by the addition of nonmagnetic second phases of proper size and distribution in a magnetic matrix, resulting in an increase in Hc without affecting H 0. A similar approach was used by Nesbitt and Gyorgy on gold‐doped Permalloy. To some extent, the desired effect can be obtained by underfiring the ferrite. However, such firing treatments are difficult to control. A method is described whereby the coercive force is increased while employing a noncritical firing treatment. This has been achieved by the inclusion of palladium as a second phase in a magnesium—manganese ferrite. Particles of the order of 1500 Å were dispersed in the ferrite, and toroidal samples were fired at temperatures between 1250° and 1400°C for firing times up to 12½ h. At 1400°C, within the range of 7½ to 12½ h, variations in Hc were very small. Moreover, in samples containing 10% palladium,Hc increased from 0.9 to 2.7 Oe without changing the H c of 3.5 Oe. In coincident current operation, the 0, 10, and 20% samples switched in 0.70 0.16, and 0.10 μsec, respectively.

This publication has 1 reference indexed in Scilit: