Variation with the concentration, -- - - - however, is not infrequent. Thus, for values for p varying between 24 and 25 are found. This may be due to a change in the character of the magnetic "carriers" with concentration owing to the formation of complex ions, a suggestion supported by the fact that in some cases the p value is markedly influenced by the addition of acids to the solution. It may also be due to the Curie law being no longer obeyed at higher concentrations. In spite of these variations, there is a remarkable agreement as to the order of magnitude of the ionic moment deduced from measurements on different salts; and there is a regular variation of the moment —first increasing and then decreasing—with the number of elec trons in the ion, as shown by the numbers in the following table of the probable rough values of ionic moments, based on the work of Weiss, Cabrera and others. (The positive charge on the ion is indicated by the index; underneath is given the number of electrons, n, in the ion; the moment is expressed as p Weiss mag netons.) Fig. 29.-GRAPH SHOWING THE VARIATION WITH TEMPERATURE OF THE SUSCEPTIBILITY OF SOME PARAMAGNETIC SALTS and number of the diamagnetic ions present as well as of the para magnetic.
In some cases there are abrupt discontinuities in the T curves, X attributable to changes in 0 or C. Chemical changes may occur. For some exceptional salts the paramagnetism remains practically constant. In the following table are given some illustrative results for a few of the normal paramagnetic salts investigated by Theodorides. The positive charge on the ion is given by the index, and the number of electrons follows.
The magneton values found for the ion from measurements on different salts agree fairly well among themselves, and also with those found from solutions. A striking fact is that the magneton values for different ions with the same number of electrons (as and are in close agreement. The magneton number, in fact, as pointed out by Kossel, depends on the number and con figuration of the electrons. It may be supposed that there is a definite value associated with each ion, and that the different values obtained are due partly to experimental uncertainties and partly to real disturbing influences not considered in the theory. Approximate magneton values may, however, be assigned to the ions. These are shown in the following table, and plotted against the electron number in fig. 3o.
The susceptibilities of a number of the rare earth sulphates and oxides have been measured by Cabrera and St. Meyer. It was assumed that the sulphates obeyed Curie's law as does gado linium sulphate. Aqueous solutions of rare earth salts have likewise been examined by H. Decker. The results of the different investigators are in fair agreement; the approximate magneton values calculated for the ions (the trivalent ions from Lai 54 to 68) are plotted in fig. 31.
According to Langevin's theory the intensity of magnetization depends on H/T; the effect of decreasing T is similar to that of increasing H, so that, with the fields which are attainable by the usual means, it is possible to ob tain very high values of the "effective field" if the tempera ture can be sufficiently reduced; the theory as to an approach to saturation, revealed by a devi ation from a straight line relation between the intensity and H/T, can then be tested. Octahydrated gadolinium sulphate (with the active ion 61) has been examined from this point of view by Onnes and his collaborators at Leiden. It is one of the few substances which obey Curie's law down to the lowest temperatures, 1.3°A. The susceptibility is given by X = .0203/T, indicating a value for p of 39.2 magnetons (see above ionic moment curve). This corresponds to an ionic moment = 7.2X giving Below temperatures of 4.25°A the proportionality between magnetization and field broke down. At 1.3°, with H=22,000 gauss, values of a approaching 8 were obtained. Plotting the magnetization against a, the points were found to lie very closely to a Langevin curve (fig. 27). The intensity of magnetization reached corresponded to 84% of the saturation value calculated in the usual way from the initial susceptibility. A striking con firmation is afforded of the general validity of the Langevin theory. The theory, however, is essentially for a gas, in which the molecular magnetic "carriers" can change their orientation. It is remarkable that the results are in complete agreement with what would be expected if the magnetic ions of solid gadolinium sulphate were free to rotate like the molecules in a gas.
The susceptibility of most salts follows a Weiss law, Fig. 31.-IONIC MOMENTS OF RARE EARTH ELEMENTS (ALL THE IONS HAVE A POSITIVE CHARGE OF 3) peratures, a real Curie point has not been found for any of those investigated at low temperatures.
A special investigation has been carried out by L. C. Jackson (Phil. Trans. 1923-26) on the susceptibilities down to 14° A of a series of sulphates of iron, nickel and cobalt, in powder form and as single crystals. The results for the susceptibility along the three magnetic axes of a monoclinic crystal are shown in fig. 32. The curvatures illustrate one type of "cryo magnetic" anomaly. The straight parts of the curves are approx imately parallel, indicating an approximately constant value of and hence magneton value (the values obtained for Co' were 24.9, 24.5, 24.8) and a value for 0 (and so of the molecular field) which varies with the axis. Somewhat similar results were found by Fax for siderose (an impure ferrous carbonate), the magneton value being the same along different axes, but the molecular field varying. Among other substances belonging to the group of "normal paramag netics" are a number of complex coordination compounds. These will be referred to later.