Title:

Studies of magnetic solids at low temperatures

The series of compounds which crystallize in the orthorhombic Pbnm distortion of the perovskite structure has proved a lucrative source of experiments in solid state physics. In the general formula ABO_{3}, either or both the metallic ions may be magnetic and since the two site symmetries are also different this provides several different degrees of complication. The main study in this thesis is of the two members GdAlO_{3} and DyAlO_{3} in which the B site is occupied by the diamagnetic aluminium ion. We therefore only have to consider the interactions of the rare earth ions with each other and with their surroundings. Three types of experimental measurements have been carried out. Specific heat measurements in the range 0.5°K to 20°K were performed in a conventionally designed He^{3} cryostat using a germanium resistance thermometer. Measurements of the magnetic moment and most of the magnetic susceptibility measurements were carried out by a ballistic method and the induced signal measured on an integrating digital voltmeter. Applied magnetic fields up to 90kOe were available for use in the moment experiments. The temperature range 0.4°K to 4.2°K could be covered using pumped He^{3} or He^{4}. Susceptibility measurements in the pumped liquid hydrogen range and some measurements in the liquid He^{4} range were carried out using a mutual inductance technique. The most prominent feature of the specific heat measurement on GdAlO_{3}, was a lambda anomaly peaked at 3.870°K which we take to be the ordering temperature. Around 1°K there is another smaller bump which is a Shottky anomaly due to the crystal field splitting. The total entropy under the magnetic part of the specific heat was measured to be R ln8 showing that the Gd^{3+} ion is in its ^{8}S_{7/2} ground state. Measurement of the magnetic susceptibility of GdAlO_{3}, along the three orthorhombic axes showed results similar to that expected for a simple two sublattice antiferromagnet with the baxis as the direction of magnetic alignment. Above the Néel temperature there was very little anisotropy. As expected for an antiferromagnet with small anisotropy, the phenomenon of "spin flop" was observed in the magnetic moment measurement along the baxis. The critical field was found to be 11.58kOe. Magnetic saturation was achieved along the a, b and caxes at 45.6kOe, 39.5kOe and 44.4kOe respectively with an isotropic saturation moment within 1% of 7 Bohr magnetons per atom. Prom a simple molecular field theory analysis of these experiments we obtain the values 21.0kOe for the exchange field and 2.7kOe for the anisotropy field. Closer analysis reveals that the anisotropy is mainly due to the crystal field interaction and the principal axis is canted at ±38° to the baxis in the ab plane for the two different magnetic sites. This means that the ordering mode is really divided into four sublattices with hidden canting in an A_{x}G_{y} configuration following the notation of Bertaut. At low temperatures the G_{y} component, which corresponds to antiparallel nearest neighbours, is dominant showing why many of the properties could be explained on the basis of a simple two sublattice antiferromagnet. The sublattice magnetization was calculated in several different ways and helped show that the internal consistency of molecular field theory is much better than its accuracy in ab initio calculations. Using results from the moment experiments along the baxis the magnetic phase diagram on the HT plane was plotted. The value of the nearest neighbour exchange constant was evaluated from results of several different experiments using the Heisenberg model and a best value of J/k = 0.067°K obtained. Prom the consistency of the results it was concluded that second nearest neighbour exchange is negligible. Specific heat measurements on DyAlC_{3} below 8°K showed only a lambda anomaly peaked at the Néel point of 3.53°K. The entropy under the specific heat was measured to be almost R ln2 showing that the Dy^{3+} ion is in the lowest lying Kramers' doublet of the ^{6}H_{15/2} ground state. Thus the Dy^{3+} is well described by an effective spin of andfrac12; and this gives rise to very anisotropic interactions with its surroundings. Optical measurements have shown that one gvalue is very much larger than the other two so we expect it to behave to a good approximation as an Ising system. The susceptibility measurements below the Néel temperature confirm this belief, with the susceptibility decreasing with decreasing temperature along both the a and baxes. The ratio of the two measurements remains constant at a value corresponding to tan^{2}33° showing that the canting angle is ±33° to the baxis. The susceptibility along the caxis is very small and almost temperature independent. The moment measurements in the ab plane all showed the phenomenon of "spin flip" or reversal of the direction of a sublattice, expected in an anisotropic antiferromagnet. Measurements at an arbitrary angle showed two flips due to the magnetically inequivalent sublattices and the angular dependence could be explained well using dipole interactions alone. Although the magnetic ordering is due mainly to dipole interactions it is shown that there is a significant antiferromagnetic exchange interaction with J/k~1°K between nearest neighbours along the caxis. Susceptibility measurements were carried out on four series of rare earth compounds of the general formulae RNbO_{4}, R_{3}NbO_{7}, R_{2}TiO_{5} and R_{2}Ti_{2}O_{7} as well as several other members of the RAlO_{3} family. Half of the compounds showed ordering transitions but further detailed work could not be carried out because of the difficulty in obtaining suitable specimens.
