The mechanisms by which magnetisation changes occur in magnetic materials may be investigated by a variety of hysteresis measurements. During this study both alternating and rotational hysteresis measurements were used to characterise the mechanisms of magnetisation change in a number of permanent magnet materials. Studies of the time dependence of magnetisation, remanent magnetisations and the dependence of the reversible magnetisation on the irreversible magnetisation were undertaken. These studies revealed that in sintered rare-earth iron magnets the magnetisation change is predominately controlled by domain nucleation, with a lesser contribution from domain wall pinning within the boundary regions of the grains. Similar mechanisms control the magnetisation change in the larger grains of melt-quenched rare-earth iron magnets. In the single domain grains of the melt-quenched materials incoherent rotation mechanisms control the changes of magnetisation. Magnetisation change in MnAlC and sintered AlNiCo was found to be controlled by domain wall pinning within the interior of the grains of the materials. Two devices were constructed for the measurement of rotational hysteresis. The first measures the angular acceleration of a sample set spinning in a magnetic field, from which the rotational hysteresis loss may be determined. The second employs rotating search coils to make direct measurements of the component of magnetisation that contributes to rotational hysteresis loss during the rotation of a sample in a field. Both devices were found to produce data consistent with that in the literature and to be useful for the characterisation of rotational hysteresis in permanent magnet materials. A simple model was used to examine the dependence of rotational hysteresis loss on various material parameters. It was found that the value of the rotational hysteresis integral is dependent on interactions and to a lesser extent distributions in anisotropy. This is contrary to assumptions commonly made in the literature but consistent with published experimental data, which has been reinterpreted. Measurements of rotational hysteresis losses in the materials studied were found to be effected by geometric demagnetisation effects. A method by which such data may be corrected for these effects is proposed. Following correction and consideration of the interactions within the materials, the rotational hysteresis data was found to be consistent with the characterisations performed in linearly alternating fields.

Conselho Nacional de Desenvolvimento Científico e Tecnológico / The exchange bias (EB) phenomen occurs due to the coupling between ferromagnetic and antiferromagnetic material and the main characteristics are the rise of unidirectional anisotropy and the rotational hysteresis in torque curves. In this work we have investigated how the unidirectional anisotropy and the rotational hysteresis are influenced by the change of some characteristics in thin films samples. Among these are the stacking of layers, the roughness at the interface between the two materials, the difference in the ferromagnetic layer thickness and the difference between the antiferromagnetic materials. The study was made by using magnetic torque measurements which were interpreted using a granular model for the EB. The parameters obtained from the fitting of the model to the torque curves has also permitted to reproduce data obtained by different magnetic techniques, especially in samples where the unidirectional anisotropy is greater than the uniaxial anisotropy. In NiFe=IrMn samples, the unidirectional anisotropy is favored by stacking layers, by increasing of the interfacial roughness and by the decrease of the ferromagnetic layer thickness, while the rotational hysteresis is substantially increased as the interfacial roughness increases. For the NiFe=FeMn sample the highest values of unidirectional anisotropy and rotational hysteresis were found, as well as a dependence of these parameters with the measuring magnetic field. / O fenômeno de "exchange bias" (EB) ocorre quando um material ferromagnético está acoplado a um material antiferromagnético e apresenta como características principais o surgimento de uma anisotropia unidirecional e a histerese rotacional em curvas de torque. Nesse trabalho foi investigado como a anisotropia unidirecional e a histerese rotacional são influenciadas pela mudança de algumas características em amostras na forma de filme finos. Entre essas características estão o empilhamento de camadas, a rugosidade na interface entre os dois materiais, a diferença na espessura da camada ferromagnética e a diferença entre os materiais antiferromagnéticos. O estudo foi realizado através de medidas de magnetometria de torque nas amostras, as quais foram interpretadas usando um modelo granular para o EB. Os parâmetros resultantes do ajuste do modelo às curvas de torque permitiram também reproduzir dados obtidos por outras técnicas de caracterização magnética, sobretudo em amostras que apresentam a anisotropia unidirecional maior do que a anisotropia uniaxial. Em amostras de NiFe=IrMn, a predominância da anisotropia unidirecional é favorecida pelo empilhamento de camadas, pelo aumento da rugosidade interfacial e pela diminuição da espessura da camada ferromagnética, enquanto a histerese rotacional apresenta aumento significativo somente com o aumento da rugosidade. Em uma amostra de NiFe=FeMn foram encontrados valores maiores tanto para a anisotropia unidirecional quanto para a histerese rotacional, bem como uma dependência desses com o campo de medida.

Fundação de Amparo a Pesquisa no Estado do Rio Grande do Sul / The Exchange Bias (EB) phenomenon consists of an interfacial exchange coupling between a ferromagnetic (FM) material and an antiferromagnetic (AFM) one and has as the main characteristics the rise of a unidirectional anisotropy and a rotational hysteresis (HR) in torque curves. In this work we have investigated the behavior of the magnetic behavior of systems presenting EB through magnetization curves, anisotropic magnetorresistance (AMR)and torque magnetometry. We have grown films, via magnetron sputtering, of NiFe/FeMn (bilayers) with different thickness of the FM layer and film structures of NiFe/IrMn/Ta with different repetitions of the basic s tructure. From the X-ray diffraction and reflectivity we have determined the preferential crystallographic growth direction, the thickness and the grain size of the studied samples. It was implemented at the Magnetism and Magnetic Materials Laboratory (LMMM) an apparatus to measure the AMR in two configurations: resistance as a function of the applied field angle and resistance as function of the field applied in a particular angle direction. To adjust the magnetization and AMR data we have used a phenomenological model for EB considering four energy terms: Zeeman, uniaxial, unidirectional and magnetostatic. The AMR curves have shown an asymmetry around 180º when measured at field smaller than the Bias one. According to the used model used, such asymmetry can be caused either by a misalignment between the anisotropy axes (uniaxial and unidirectional) or by a misalignment between the measurement current direction easy magnetization axis of the sample. It was not observed in the multilayers an increase on the uniaxial anisotropy dispersion with increasing on the trilayers number, as expected. In order to fit average torque curves, we have used an equation with two harmonic terms. The resulting parameters allow us to study the field evolution of the harmonic amplitudes Kθ e K2θ. In the multilayer system as higher total energy is, higher is the energy losses by HR. In bilayers system, the HR energy losses still increases even above the saturation field. Such behavior was associated to the configuration of interfacial spins and suggests the need of an additional contribution to total energy of the magnetic system. / O fenômeno de Exchange Bias (EB) consiste de um acoplamento de troca interfacial entre um material ferromagnético (FM) e um material antiferromagnético (AFM) e tem como características principais o surgimento de uma anisotropia unidirecional e a histerese rotacional (HR) em curvas de torque. Neste trabalho investigamos o comportamento de sistemas magnéticos com EB através de curvas de magnetização, magnetorresistência anisotrópica (AMR) e magnetometria de torque. Foram crescidos filmes via magnetron sputtering, de NiFe/FeMn (bicamadas) variando a espessura da camada de NiFe e filmes de NiFe/IrMn/Ta com diferentes números de repetições dessa estrutura básica (multicamadas). A partir da difração e refletividade de raios-X determinamos a direção preferencial de crescimento cristalográfico, espessura e o tamanho de grão das amostras estudadas. Foi implementado no Laboratório de Magnetismo e Materiais Magnéticos (LMMM) um aparato experimental para medidas de AMR: resistência em função do ângulo do campo aplicado e resistência em função do campo aplicado para um determinado ângulo escolhido. Para o ajuste aos dados das curvas de magnetização e AMR é apresentado um modelo fenomenológico para EB considerando quatro termos de energia: Zeeman, uniaxial, unidirecional e magnetostático. As curvas de AMR apresentam uma assimetria em torno de 180º quando medidas em valores de campo menores que o campo de Bias. De acordo com o modelo utilizado, tal assimetria pode ser causada por um desalinhamento entre os eixos de anisotropia (uniaxial e unidirecional) ou por um desalinhamento entre a direção da corrente de medida e o eixo fácil de magnetização da amostra. Não foi observado um aumento na dispersão da anisotropia uniaxial com o aumento das repetições das tricamadas, conforme esperado. Para o ajuste das curvas de torque médio utilizamos uma equação com dois termos harmônicos. Os parâmetros resultantes do ajuste das curvas de torque permitiram estudar a evolução das componentes harmônicas Kθ e K2θ. Nas multicamadas quanto maior a energia associada ao torque médio maiores são as perdas de energia por HR. Já nas bicamadas, as perdas de energia por HR aumentam mesmo acima do campo de saturação. Tal comportamento foi associado à configuração de spins interfaciais e sugere a necessidade de uma contribuição adicional para a energia total do sistema magnético.