Liquid phase epitaxy of garnets

A short introduction to the technique of liquid phase epitaxy of single crystal thin films of magnetic garnets is given with the emphasis on the relative merits of vertical and horizontal dipping. For vertical dipping results are given for the thickness profiles from top to bottom of the samples and the inhomogeneities within the film thickness. The thickness profile is explained by fluid motion caused by density segregation. The inhomogeneities are explained by growth rate changes due to transient growth at the beginning and nearly steady state growth when the diffusion boundary layer has been formed. It is shown under special conditions these inhomogeneities in composition in LaGa : YIG films can be an advantage and can increase the domain wall velocity. These effects are compared to films grown by horzontal dipping with rotation. Using spin wave resonance it is shown that this process does not produce as homegeneous films as expected. Explanation of this is given in terms of fluid flow during the growth. Because literature data on the diffusion coefficient D has a wide spread of values, an experiment is described in which D has been measured independent of the model using radioactive tracers. These results, combined with the model that the garnet does not diffuse as a molecule but as species like Y₂O₃ and Fe₂O₃ and at the growing interface react to form garnet, has lead to a new insight into the solvation of these species in the liquid. Magnetic bubble films of the type CaGe substituted iron garnets are used in large numbers for devices. This makes the reproducibility an important factor. To increase the yields of usuable material it has been found necessary to make the magnetic parameter H_collapse independent of the growth temperature because of the thermal constants in LPE furnaces. This has been achieved by combining CaGe melts with Ga. Results are given for a 3 μm bubble material. The LPE technique can be used for non-magnetic garnet materials as well as iron garnets. These materials are usually substituted with rare-earths. The incorporation of Ce into Y₃Al₅O₁₂ is a difficult case. How this can be achieved and difficulties encountered are explained in the last part of the paper.