London: Longmans, Green and Co., 1907
According to the electron theory, varying valency would be explained by assuming that a mono-, di-, or trivalent negative ion consists of a combination of the atom or atomic complex in question with one, two, or three electrons, which are all negative. A positive ion is formed from a given atom or complex by the splitting of of the necessary number of electrons (cf. Chap. VIII.). This conception has hitherto remained only a formal one, and has led to no new results.
A short account, however, may be given of the manner in which J. J. Thomson seeks to reconstruct this scheme by the aid of the known properties of the electrons.
We may, with Thomson, consider the elements of one series, for instance, series 3-
and compare this series with the possible combinations of electrons, containing an outer ring of 20 electrons. These and their nearest neighbours are, according to Thomson, the following:
Of these, the system of 59 electrons contains in its outermost ring the greatest possible number of electrons that may be held together by the inner electrons. Therefore this combination is very unstable, and easily loses one electron, thereby becoming the combination 58 with a unit positive charge. On the other hand, the compound of 58 electrons is so stable in the outer ring that it very easily adds on a negative electron and holds it at the surface of the atom. That is, as soon as the combination of 59 electrons has been transformed into the atom with 58 electrons, it will again bind one electron, so that the whole process consists in the transference of one electron of the 59 to the outside of the atom. This atom will, therefore, possess neither a positive nor a negative charge; it will be without valency, and therefore belong to the group 0.
The outside electron of the atom (58 + 1) may, on the other hand, again go back to the outer ring, so that the atom 59 is reformed. There will therefore be an equilibrium between uncharged 59-atoms and positively charged (58 + 1)-atoms with an outer negative electron. But in both cases the total charge is zero, and we have an element of the group 0, like neon, before us. In an analogous way Thomson tries to demonstate that the 67-atom represents an element of the group 0, like argon.
If we then proceed to the 60-atom, we find that it may lose one electron, and thus be transformed into the 59-atom; but if it loses one electron more (58-atom), it will add it on to its outside again, just as the 59-atom did. The 59-atom derived from the original 60-atom, therefore, represents a monovalent positive element, like sodium.
By extending this reasoning to the 61-atom, we find that it may lose two electrons and behave as a divalent positive atom, like magnesium; the 62-atom will give a trivalent positive atom, like aluminium; the 63-atom a tetravalent positive atom, like silicon (in chloride of silicon, SiCl4); and the 64-atom a pentavalent atom like the phosphorus atom (in PCl5); and so forth.
On the other hand, by adding negative electrons to the outside of each, the 66-, 65-, 64-, and 63-atoms will be converted into negatively charged mono-, di-, tri- and tetra-valent negative atoms, corresponding to chlorine, sulphur, phosphorus, or silicon in their compounds with positive hydrogen and its substituents.
This representation of Mendeleeff's system is very interesting, and it may be hoped that with further work success will meet the efforts to remove the numerous difficulties with which it has now to contend. One of these difficulties is inherent to the scheme itself; the atomic weight ought to increase with the positive valency in every series. This seems not to be the case. There are two very prominent exceptions to this rule: The one is argon, which has a higher atomic weight than potassium (39.8 against 39.1). The figure 38 for argon given by Mendeleeff is only a "theoretical" value, computed in order to retain the simplicity of his scheme; the experiments of Ramsay1 give 39.8 for argon, and potassium has been very accurately determined to be 39.1. The other exception is met with in the case of the two elements tellurium and iodine, with the atomic weights 127.6 and 126.97 respectively.
This difficulty is only a specially striking instance of a more general one; the systems in Thomson's series differ from each other by one electron, so that the difference between two consecutive atomic weights is constant. This does not agree with the much more complicated behaviour of the natural elements. In the series 2 this difference varies between 1.05 and 3.39, i.e. in the proportion 1 : 3.2, in a somewhat irregular manner. In other series this variation is of the same order.
But there are other difficulties not specially characteristic of Mendeleeff's scheme itself, but only of its representation by Thomson. After the second series with 20 electrons in the outer
1 Ramsay and Travers, Proc. Roy. Soc., 1898, 64, 183.
ring, there will come a series with 21 electrons in this ring, then one with 22 electrons, and so on. These series correspond to the series 3, 4, 5, 6, etc., in Mendeleeff's scheme. Now, the number of atoms belonging to one series is very nearly constant, and likewise the difference between the atomic weights of the extreme elements is not very unlike in succeeding series. The series of Thomson behave in quite another way. It is only by chacce that the series with 20 electrons in the outer ring contains 7 elements with valencies corresponding to the second, third, or seventh series of Mendeleeff. In general, the number of elements belonging to a Thomson's series will increase nearly proportionally to the 2/3 power of the atomic weight of its first element; therefore the number of elements in the seventh series ought to be about three times as great as that in the third series and about eight times as great as that in the second series. Thomson's scheme would, therefore, if it were worked out, differ considerably from its prototype.
One might imagine that perhaps this difficulty would disappear if the atoms were represented, not by rings of electrons lying parallel to one another, but by some other arrangement of the electrons, which would correspond better with the three dimensional configuration of the atoms. But, as Thomson remarks, the general outlines of his system would not be substantially altered by such a modification.
Another difficulty, and perhaps the greatest one, is that experiment shows that the mass of the electrons is only about the two-thousandth part of that of the hydrogen atom. According to this idea, it would be natural to suppose that the neon atom consists of 39,800 electrons, the sodium atom of 46,100. It would, therefore, not be one step from neon to sodium, but 6300. Hence it will be necessary to add to Thomson's scheme the supposition that not one electron, but a great number of electrons, together with the corresponding quantity of positive electricity, makes the difference between two consecutive elements. And as the differences between the atomic weights of consecutive elements are not at all constant, it will be necessary to assume that the complex of electrons is very different in different cases. By this amendment the Thomson scheme loses its simplicity, and at the same time much of its theoretical value.
The latest calculations of J. J. Thomson1 may perhaps open out a new perspective. They are based on three different series of measurements, and lead to the result that "the number of electrons (corpuscles) in an atom of an elementary substance is of the same order as the atomic weight of the substance." Thus, for instance, the scattering of Rontgen rays by air gave the result that the number of electrons in a molecule of air is about 25 (according to experiments of Barkla), whereas the mean molecular weight of air is 29. This result reminds one very much of the hypothesis of Prout.
Another objection may be made to Thomson's scheme, because it demands that there should be two elements of the group 0 in every series. Thus, in series 3, neon should correspond both to the 58-atom and to that composed of 59 electrous. We know nothing of such a duplication of the elements of group 0. Further, if, for instance, the 59-atom is unstable and loses one of its electrons, why should it regain its stability by simply adding on the lost electron to its outside, and not just as well by attracting a negative monovalent atom, e.g. a chlorine ion? In this manner the positive potassium ions unite with negative chlorine ions and give salt molecules. One cannot ignore the fact that it seems necessary to find the reason why the 59-atom does not behave in the same way in order to acquire stability; in other words, why the 59-atom cannot as well be a monovalent positve element as one without valency.
It is very interesting to note that at the same time as Thomson worked out his idea of the composition of matter, the Japanese physicist Nagaoka2 was also led to similar assumptions in order to explain the optical, especially the spectral, phenomena of matter. He supposes electrons to move in concentric rings around a positively charged centre; also that the whole positive charge is concentrated at one point, and not equally distributed over a sphere. Nagaoka, as well as Thomson, has worked out an explanation of the breaking down of the radium atom into helium and something else. If this breaking down is finally satisfactorily proved, these theoretical deductions may prove of value.
1 J. J. Thomson, Phil. Mag. (6), 1906, 11, 769.
2 Nagaoka, Nature, 1904, 69, 392.