The thermoanalysis of metal single crystals and a new thermoelectric effect of bismuth crystals grown in magnetic fields

Part I. The Thermoanalysis of Bi Single Crystals: Production of crystals.—Crystals of Bi of any desired orientation were grown by the method of Goetz, one half of each normally, the other half within a transversal magnetic field. The orientations, predetermined by a seed crystal, were not affected by this process. Two methods of growth—the continuous and the discontinuous—were used. The thermoanalysis of a crystal.—A method and its experimental realization—the thermoanalyzer—were developed to measure and to localize any changes of the thermoelectric properties along the lengths of crystals without applying mechanical stresses to them, by progressive local heating of the specimens. Thus very small distortions and imperfections were detectable due to their thermoelectric asymmetry. Types of thermoanalytic diagrams.—The types of diagrams to be expected in the cases of a perfect single crystal, a double-, and a triple-crystal are discussed and a method is developed to analyze simple diagrams by means of the theory of heat-conduction. Part II. Experimental Results Obtained for Normal and Magnetic Crystals with a Discussion of the Same: The thermoanalysis of normal crystals.—The application of the thermoanalysis to normal single crystals leads to the discovery of very small (0.1 mm2) regions within a crystal which are distorted due to very slight variations of the cooling conditions during the growth of the crystal. It is however possible to avoid these faults and to produce comparatively perfect specimens. Hence it was possible to test the results of different methods of crystal production and thus to refine the method used as well as to measure quantitatively the influence of a magnetic field applied to the crystal during the time of its formation. The thermoanalysis of "magnetic" crystals.—It was found that the normal half of a crystal has a thermoelectric e.m.f. against the "magnetic" half. The sign and size of this e.m.f. depend on many circumstances, though mainly on the orientation the growing crystal has with regard to the direction of the field lines. The effect as a function of the orientation.—The effect is a maximum if the principal axis of the crystal grows normal to the lines of force and it is very small (probably zero) if the axis grows parallel. It depends furthermore on the orientation of the crystal with regard to the direction of the thermoelectric current, since the two orientations in which the principal axis is normal to the field show different effects: a small one if the axis is normal to the current, and a very large one if it is parallel.—The thermal e.m.f. obtained for the latter case is 4.3 microvolts/degree which would correspond to a change in orientation of ca. 21° though no actual change could be observed. The effect as a function of the method of growth.—The thermal e.m.f. depends largely upon the method of growth, i.e., whether the crystal is grown by the continuous or the discontinuous method. The effect as a function of impurities.—The thermal e.m.f. for one and the same method of growth depends also on the amount of chemical impurities. Four kinds of purest Bi of different provenience were used in which the total amount of other substances was less than 0.2%. The metals were spectroscopically examined and it appeared that the purest (electrolytic) metal showed the smallest effect. It was ca. 40 times smaller than the effect of another metal. Traces of Ag and Pb seem to affect the "magnetic" sensibility of Bi most whereas Sb is rather ineffectual. The effect as function of the field strength.—The effect is influenced by the strength of the applied field through this problem is not yet settled quantitatively. The influence of a small field (102 Gauss) is comparatively large, whereas fields of more than 13000 Gauss seem to decrease the effect. The maximum seems to depend largely on the kind of the impurity. The effect as a function of temperature.—The Peltier-effect (extrapolated from the E-f(θ) curve) shows a sharp discontinuity between 75° and 90°C, indicating a different relation with regard to the transformation of Bi in this region. No indication could be found that annealing above this point (16 hours) and aging (1 month) destroys the difference between the normal and the "magnetic" half of a crystal. Discussion.—The results obtained are brought into relation with the diamagnetic anisotropy of Bi. They are discussed with regard to the investigations of other authors and it appears possible to describe the effect as due to a change in the secondary (mosaic) lattice of the Bi crystal