Floyd B. Humphrey

Transient bubble domain configuration in garnet materials observed using high speed photography

The dynamic domain configurations of bubbles in garnet materials have been studied using a sampling optical microscope capable of single exposure photographs with a 10 nsec exposure time. The microscope is an integral part of a sampling system so that the transient shape of the bubble is recorded at various times after a field pulse or, for bubbles in field access devices, during a clock cycle. A triggerable flowing nitrogen gas laser pumping a low Q Rodamine 6G Dye laser is used as an illumination source giving light pulses of ∼1.5 KW for 10 nsec. This light is sufficient to expose Kodak 4 × 16 mm movie film. Standard pulse generators (HP 214A) are used to make free bubble radial velocity measurements. A modified generator allows free bubble collapse measurement to be made. For bubbles propagating at operating frequency within field access devices, a standard bubble exerciser is used, synchronized to the sampling system. In this case, special samples with an internal mirror and epi-mode illumination are used. Illustrative results of bubble domain shapes in a chevron propagating structure and a 90° chevron expander detector are included.

Anisotropy pinning of domain walls in a soft amorphous magnetic material

Ribbons were annealed in the demagnetized state with one wall along the ribbon middle. This wall becomes pinned during the heat treatment. Reentrant reversal occurs when reverse domains are nucleated at the ribbon edge with a threshold field larger than the demagnetizing field; this wall does not annihilate when it meets the pinned wall but leaves a line of reverse domains stabilized by ripple in the anisotropy. These domains permit a regular smooth reversal for the demagnetization process until the ribbon returns to the pinned configuration. The regular loop appears when the ribbon has been completely saturated by a large field. Mobile walls are nucleated on both sides of the pinned wall so that the ribbon does not return to the pinned configuration. Reversal now follows the usual demagnetization curve over the entire cycle. Kerr magnetooptical domain and domain wall observations are used in this investigation. All of the possible wall structures predicted by the model of asymmetric flux closed Bloch walls were identified. >

Mobility and Loss Mechanisms for Domain Wall Motion in Thin Ferromagnetic Films

Domain‐wall mobility for Ni–Fe alloy films has been measured as a function of film thickness from 300 to 1650 A. Between 300 and 800 A, the mobility decreases with increasing film thickness, ranging from 8×103 cm/sec·Oe at 300 A to 3×103 cm/sec·Oe at 800 A. Between 900 and 1000 A, the mobility increases rapidly with increasing film thickness to about 7×103 cm/sec·Oe. Above 1000 A, the mobility increases slowly with film thickness. These mobility data have been compared with predictions based on intrinsic and eddy‐current loss mechanisms utilizing available static domain wall‐shape information. Predictions based on a 180° Bloch or Neel wall model in which the magnetization rotates linearly through the wall are in disagreement with the present data over the entire range of film thickness. For the 300–800 A thickness range, predictions based on Lorentz microscopy wall‐shape measurements are in good agreement with the measured mobility values. The crosstie structure associated with Neel walls in this thicknes...

Magnetic Anisotropy in Flat Ferromagnetic Films: A Review

A review of those characteristics of the anisotropy found in thin Fe–Ni–Co films that are important from a practical viewpoint are presented. Previously published data are supplemented to form a composite ternary diagram of anisotropy field Hk as a function of composition for the alloy system Fe–Ni–Co for 250°C substrate temperature. Anisotropy fields as a function of composition for the three binary alloy systems at other substrate temperatures are included. A composite diagram showing anisotropy field as a function of measuring temperature illustrates the need for temperature stability. Magnetic annealing and aging is discussed. Finally the importance of magnetostriction as a source of the anisotropy is discussed in detail as well as pair ordering. It is concluded that no single source or combination satisfactorily describes the origin of the anisotropy.

Sensitive Automatic Torque Balance for Thin Magnetic Films

An automatic torque balance has been constructed for use in studying the magnetic properties of thin ferro‐magnetic films. The automatic servo force balancing technique using a photoelectric pickoff has been applied to a fused‐silica torsion fiber suspension. A sample up to 600 mg in weight can be handled. The data are automatically recorded by an X‐Y plotter. Sensitivity of 10−6 dyn‐cm and a response time of a fraction of a second have been obtained, providing a significant improvement over similar instruments previously used. An appropriate set of coils provides the magnetic field H. Since the torque on a sample is given by M×H=L, measurement of torque affords a means of studying the behavior of the magnetization M. Provision is made for observing torque either parallel or perpendicular to the film surface. The sensitivity available will permit the study of 1‐cm‐diam films as thin as one‐half an atomic layer under ideal conditions. Examples are given of the measurement of the saturation magnetization, r...

Gradient propagation, overshoot, and creep in magnetic‐bubble garnet material

Gradient propagation of free magnetic bubbles has been investigated using high‐speed transient photography. Overshoot in the translational motion was confirmed and a new mode of motion, creep, was introduced. Creep is induced translational motion caused by bias‐field pulsing after a driven translation and overshoot. It was found that overshoot is bubble‐size (bias‐field) dependent and that creep and overshoot are complementary with the total motion independent of the bias field. A model is presented that provides that the wall structure at the end of the gradient pulse be dominated by these vertical wall twists, i.e., vertical Bloch line pairs with each having a winding direction opposite the other. This structure is packed by the g forces during translation and relaxes after the termination of the pulse to give overshoot. Creep is the result of annihilating the remaining wall twists by bias‐field pulsing as the wall assumes its most stable state.

Relaxation Processes for Ferromagnetic Resonance in Thin Films

Ferromagnetic‐resonance linewidth measurements have been made for two hundred Ni–Fe alloy films (77% Ni) 150–3200 A thick at frequencies from 1–9 Gc/sec with the static field in the film plane. To avoid dispersion effects samples with the smallest linewidth (ΔHmin) were selected for each thickness. For film thickness less than a frequency‐dependent critical thickness Dω, ΔHmin is independent of film thickness. For thicker films where D > Dω, ΔHmin increases linearly with film thickness. The observed Dω values (about 1000 A) are in good agreement with predictions based on magnon scattering involving spin waves degenerate with the uniform mode. Because of the magnetostatic mode modification of the spin‐wave dispersion relation for thin films there are no spin‐wave states degenerate with the uniform mode for D < Dω and magnon scattering is not allowed. The applicability of two different magnon processes, s‐d exchange and two‐magnon scattering, is discussed. Neither mechanism provides a completely satisfactor...

A Nanosecond Kerr Magneto‐Optic Camera

A Kerr magneto‐optic camera has been constructed which takes 10 nsec exposures of the dynamic magnetization configuration in a magnetic thin film during high speed flux reversal. The photographs obtained have 10 μ resolution and sufficient contrast and brightness to record, in a single exposure, the dynamic state during a 50 nsec or longer flux reversal. The time at which the photograph is taken relative to the application of the switching field is variable from ‐100 nsec to +10 μsec with a relative time stability at any setting of ±3 nsec. A Q‐switched ruby laser is used as a light source. The 10 nsec exposure time is obtained with a Kerr electro‐optic shutter. The reversing field has a risetime of 10 nsec, decay time of 20 nsec, variable amplitude (<30 Oe), duration (<10 μsec), and is applied with a stripline. Included in the description of the system are discussions of a blocking oscillator circuit used to drive KU‐27 thyratrons with a jitter of less than 1 nsec, an electron multiplier pulse amplifier ...

Mechanisms of Reversal with Bias Fields, Deduced from Dynamic Magnetization Configuration Photographs of Thin Films

Flux reversal with a transverse field in Ni–Fe thin films has been studied using a Kerr magnetooptic apparatus which permits 10‐nsec exposures to be taken of the dynamic magnetization configuration. Whenever the applied fields just exceed the critical asteroid, noncoherent rotation occurs, and a 50–100 μ wavelength striped configuration appears in the film. Before the stripes form there is a fast relaxation of the ripple as the magnetization rotates coherently to angles greater (5°–20°) than either the critical angle for reversal or the uniform torque minimum. Therefore, previous models of noncoherent rotation based on instabilities in the ripple which were expected to occur before or at the critical asteroid or uniform torque minimum do not explain the observed reversal process. After the stripes form, depending on the magnitude of the applied fields, one of two processes occurs: either (1) the rotation of the magnetization stops until the stripes break up into small nucleated regions or (2) the rotation...

Instantaneous radial wall velocities in magnetic garnet bubble domains

Using an optical sampling microscope with a sample time of 10 nsec, single‐exposure photographs of the transient bubble‐domain configuration resulting from a uniform pulse field applied parallel to the bias field has been recorded at known times with respect to the applied pulse field. In this way the actual dynamic domain configuration of the bubble was observed. Mixed LPE garnet samples with a nominal composition of Y1.57Eu0.78Tm0.65Ga1.05Fe3.95O12 and material parameters, for nonimplanted (implanted), 4πM=184 (205) G, thickness is 6.8 (6.9) μm, Q=16.7 (16.6), α=0.026 (0.029), γ=1.1×107 rad/sec Oe, and A=2×10−7 erg/cm, were investigated. The radius of the bubble plotted as a function of the time delay between the leading edge of the expanding field pulse and the exposure, for delays less than 1 μsec, gave straight lines for all but the low pulse fields (<40 Oe). For these field pulses, a clear velocity saturation of 5.30±0.05 m/sec over the range 40