Andrzej Stankiewicz

Domain control in magnetic shields using patterned permanent magnet underlayer

The domain state of a magnetic shield in a recording head can be controlled by an adjacent patterned permanent magnet layer. A 1.1-μm-thick electroplated Ni80Fe20 (NF) film with slight uniaxial magnetic anisotropy was patterned into rectangular magnetic shields with various dimensions over patterned thin film made from a 0.1-μm-thick CoCrPt permanent magnet (PM). The shape of the adjacent biasing PM layer should be the shape of a desired final domain in NF. Domain structure in the NF layer and the process of magnetic saturation were imaged using wide-field Kerr microscopy. The NF and PM layers are magnetically coupled and, therefore, a magnetic state with parallel magnetization is preferred. The PM direction of magnetization is set in high magnetic field and the final NF domain state is controlled by the shape of PM features. The simplest stable domain structure in a rectangular thin shield is of an “envelope” type. Using a PM underlayer, either clockwise or counterclockwise domain structure is preferred....

Damping constant estimation in magnetoresistive readers

The damping constant is a key design parameter in magnetic reader design. Its value can be derived from bulk or sheet film ferromagnetic resonance (FMR) line width. However, dynamics of nanodevices is usually defined by presence of non-uniform modes. It triggers new damping mechanisms and produces stronger damping than expected from traditional FMR. This work proposes a device-level technique for damping evaluation, based on time-domain analysis of thermally excited stochastic oscillations. The signal is collected using a high bandwidth oscilloscope, by direct probing of a biased reader. Recorded waveforms may contain different noise signals, but free layer FMR is usually a dominating one. The autocorrelation function is a reflection of the damped oscillation curve, averaging out stochastic contributions. The damped oscillator formula is fitted to autocorrelation data, producing resonance frequency and damping constant values. Restricting lag range allows for mitigation of the impact of other phenomena (e.g., reader instability) on the damping constant. For a micromagnetically modeled reader, the technique proves to be much more accurate than the stochastic FMR line width approach. Application to actual reader waveforms yields a damping constant of ∼0.03.

Magnetic Noise and Spin-Wave Eigenmodes in a Magnetic Tunnel Junction Read Head

With read-head structures becoming very small (<30 nm typical size), thermal fluctuations of the free and reference layers, occurring in the gigahertz range of frequencies, may become relatively large (>20°) and cause the appearance of considerable magnetic noise in the read-back function. For instance, the experimental test of a read-head sensor shows that a considerable low-frequency noise appears in the measured signal in coincidence with the superposition of high-frequency eigenmodes of the stack. To shed light on these experimental data, we perform extensive micromagnetic simulations to analyze the spin-wave spectrum of the read-head sensor and its evolution under the application of an external magnetic field. We show that the nonuniformity of both the bias field and the demagnetizing field induces pronounced spatially dependent dynamics within each layer. Moreover, the derived read-back function is characterized by different peaks that reflect the underlying eigenmodes spectrum, as suggested also by simple macrospin approach. However, for large precession amplitude, nonlinear and mixing effects appear, leading to the presence of extra peaks and of a low-frequency tail in the readback signal, in qualitative agreement with experiments.

Magnetic Reader Instability as a Source of Low-Frequency Noise

Detecting and characterizing magnetic stability is one of the main goals in hard drive reader testing. A simple model of a bi-stable system leads to random telegraph noise, and reveals spectral characteristics of unstable readers, which may be used for instability detection in standard reader testing. In particular, it proves that unstable devices may show noise reduction with temperature increase, which has been confirmed experimentally. Furthermore, characterization of the telegraph noise in magnetic readers is achieved by high bandwidth time-domain measurements, followed by waveform analysis based on a hidden Markov model. An example of actual reader analysis leads to the estimation of activation energy.