 |
| Benoist's X-ray transparency curves. |
The early workers on indium regarded that element as a diad, and analogous to - zinc. This view was based upon the facts that indium occurs naturally in association with zinc, the two metals are very similar in their analytical reactions, and no sodium- or potassium-indium sulphates could be prepared which crystallised in octahedra and could be formulated as alums on the assumption that indium was tervalent. The bivalency of indium was, however, not acceptable to Mendeleeff, who could only place indium in his periodic table on the assumption that it was a triad, of atomic weight c. 114. In support of the tervalency of indium, both Mendeleeff and Lothar Meyer advanced various chemical reasons; and Mendeleeff and Bunsen each showed, by determinations of the specific heat of the metal, that the tervalency of indium was a necessary assumption from the point of view of Dulong and Petit's Law. Subsequent work has fully substantiated this assumption. In 1873 Roessler succeeded in preparing indium ammonium alum, and in 1885 Soret prepared the rubidium and caesium alums. Most convincing of all, in 1888 Nilson and Pettersson prepared and determined the vapour densities of three chlorides of indium, and showed that the results were in agreement with the molecular formulae
InCl,
InCl2, and
InCl3, with
In = 114 approximately. This value has been confirmed by another method, which may be outlined here. The specific opacity of an element for X-rays of a definite quality is independent of its state of aggregation and of the temperature; it is also independent of whether the element is free or in combination, so that the specific opacity of a compound may be calculated from the opacities of its constituent elements. It has been found by Benoist that for X-rays of one definite quality, the specific opacity of an element increases in a regular manner with the atomic weight. This is perhaps best indicated by plotting the equivalent transparencies of the elements (which are proportional to the reciprocals of the opacities) against the atomic weights. The points lie in a smooth curve, such as fig. curve I., and the curve is somewhat similar to a hyperbola (fig., curve III.), which is the graphical representation of the connection between specific heat and atomic weight (Dulong and Petit's Law). There is, however, a separate curve of transparencies for each quality of X-rays used. Thus in fig., curve I. refers to rays of medium hardness, and curve II. to soft rays. Therefore, in seeking the atomic weight of an element by the X-rays method, each possible multiple of the combining weight is assumed, the elements which would immediately precede and follow it on the curve noted in each case, and the transparencies of each of these elements, and of the element under investigation, determined for a particular quality of X-rays, or better, for two decidedly different qualities. The results have the great advantages over those based upon specific heats that they are not influenced by temperature or physical state and are applicable to gaseous elements.
The preceding method has been applied to the case of indium, both the free element and its acetylacetonate being used. It was found that both for the medium rays (fig., curve I.) and the soft rays (curve II.) the equivalent transparency of indium is very nearly 1.10. The results point clearly to the value c. 114 for the atomic weight.
The atomic weight of indium is therefore three times its combining weight in its highest halogen compounds and its basic oxide. Early determinations of the atomic weight of indium, made by Reich and Richter, Winkler, and Bunsen, were made almost entirely by a faulty method - the synthesis of the sesqui-oxide - and yielded low results. Closely concordant results have since been obtained by Thiel and Mathers, each of whom analysed the trichloride and tribromide of indium:
| InCl3: 3AgCl | 51.473:100.000 (Thiel)
51.442:100.000 (Mathers)
| In=114.968
In=114.834 |
| InCl3:3AgBr | 62.923:100 000 (Thiel)
62.932:100.000 (Mathers) | In=114.753
In=114.803 |
The most probable result, according to Clarke, is 114.86, and the international value (1916) is
In = 114.8.
Indium readily dissolves in molten tin, and lowers the freezing-point. The "atomic fall" is 1.86°, the theoretical value for a monatomic molecule being 3.0°. Hence in dilute solution in tin, indium is mainly diatomic. From the results obtained by Richards and Wilson in a study of the electromotive forces of indium amalgam concentration cells, it follows that indium is essentially monatomic in solution in mercury; or more probably exists as a compound of molecular formula
InHgx, most likely
InHg4. Indium is essentially monatomic in dilute solution in sodium, the " atomic fall" being 3.6° and the theoretical value for a monatomic element 4.4°.
