No abstract

No abstract

Measurements of the ferromagnetic resonance line-width maximum in single-crystal yttrium iron garnet (YIG) were previously reported by Dillon. 1 His data show that the line width increases with decreasing temperature, reaches a maximum below liquid nitrogen temperature, and decreases again to approximately the roomtemperature value at liquid helium temperatures. We have been investigating the nature of this line-width maximum at low temperatures to determine if the effect is intrinsic to ferromagnetic resonance and to determine its origin. The line widths reported in this paper are determined from the absolute absorption of the sample and the saturation magnetization vs T curve measured on single-crystal YIG. The large effects on the line width AH at room temperature due to inhomogeneous broadening by scattering from pits on the surface of a YIG sphere were evaluated previously. 2 Consequently, we measured AH vs T for a series of three spheres prepared using polishing papers having mean grit sizes of 15, 5, and 0.3 microns, respectively. These measurements were made at 9300 Mc/sec. Figure 1 shows that the ratios of maximum line widths to room temperature line widths are 2.5, 4, and 13, respectively, the larger ratios occurring for the better polished spheres. It is also clearly shown that the contribution to the line width due to surface preparation is essentially additive over the temperature range. We conclude therefore that the low-temperature line-width maximum is an intrinsic property of the material and not of the surface. Further, if perfect polishing were attained the line-width maximum would still be approximately 6 oersteds in this sample.Earlier investigations on the same samples 3 have shown a small but repeatable frequency dependence of line width at room temperature. The line width AH (full width between half maximum values) is 0.44 oersted at 3000 Mc/sec and 0.52 oersted at 9300 Mc/sec along the [111] axis. Taking these measurements over the temperature range, there are three essential features. See Fig. 2. The line width maximum at 3000 Mc/sec is reduced by 40 % in magnitude and is shifted from 40°K to 30°K. Also it is seen that the room-temperature frequency dependence of 300 FIG. 1. AH vs_ T for single-crystal spheres of YIG 0.014 to 0.017 inch in diameter. H^c oriented along the [111] axis. The labels on the curves indicate the mean grit size of the final polishing paper used to prepare the spheres. The lowest curve is Fig. 3 repeated here for comparison. 300 FIG. 2. AH vs T measured at 9300 and 3000 Mc/sec. Hfc oriented along the [111] axis. 32

No abstract

A calculation is presented of the generation of microwave elastic waves by the uniform precession resonance in ferromagnets. The excitation of the thickness shear modes in a normally magnetized ferromagnetic disk is treated. For uniform precession the elastic modes are driven only on the surfaces of the disk. The effects of long wavelength spin waves magnetostrictively driven by the elastic waves can be included in the analysis by a change in the elastic modulus. The coupling out of the elastic energy by bonding the disk to a nonmagnetic delay medium is considered for arbitrary values of the elastic impedances of the two media. An equivalent electrical circuit for transducers is presented; the efficiency and bandwidth of disk transducers are calculated for a disk placed in a matched one port resonant cavity. A figure of merit for such transducers is defined, and calculated for yttrium iron garnet. Due to interference between the elastic waves generated on the two disk surfaces, the realizable bandwidth of the transducer is restricted by the thickness of the disk. By bonding two disks together or by using a rod, this restriction may be reduced. 7 W. P. Mason, Physical Acoustics and the Properties of Solids