Page images




ART. XXVIII.—On the Color Relations of Atoms, Ions and Molecules ; by M. CAREY LEA. Part I. (With Plate IV.)

The atoms of which elements are composed differ remarkably in color from the elements themselves; their colors are inore important and more characteristic than those of the elements, and if we divide the entire series of elements into two classes, those whose atoms are always colorless whatever may be their valency, and those whose atoms are either sometimes or always colored, we shall find that this division harmonizes in a striking way with their chemical properties. The colors of the elements have no such significance or importance.

This present paper accepts the Arrhenius theory of dissociation, the author believing that the evidence in its favor is too strong to be resisted. But the facts to be stated, and the conclusions to be drawn therefrom are independent of any theory.

It is somewhat remarkable that it is never possible to deduce the color of an atom from that of the element which it forms by combining with another similar atom. Between the two, the atom and the element, there seems to be no color relation whatever. It is from the combinations of an atom with one or more dissimilar atoms, kathions with anions, that we can with facility and certainty deduce the color of the atoms themselves.

So far I have spoken of atoms only, but the matter in hand is simplified by considering dissociated ions also. And as a first step it becomes necessary to establish the following proposition : namely, that in any colored inorganic compound in solution, the color belongs essentially to the metallic atom, whether it exists in a free state as an ion or combined with a dissimilar atom or atoms to form a molecule. That is to say it does not belong to the ion with exclusion of the molecule as some have held, nor to the molecule with exclusion of the ion as has been held by others. Color when it appears, is the essential property of the atom, possessed by it in the free state and carried by it into any electrolyte which it forms.

This fact can be conclusively proved from the researches of Glan and of Ewan. Each of these chemists studied the action of copper salts and more particularly of the sulphate. Both observers used virtually the same method. A ray of light was passed through a small stratum of strong solution and then through a larger quantity of distilled water and the amount of absorption was noted. In another parallel tube the two were mixed and the amount of absorption was compared, in each case for particular wave lengths. Ewan used an eight-fold dilution. Glan usually a seven-fold but sometimes a three or five. In all cases the difference found was extremely small, scarcely if at all exceeding the magnitude of experimental error.* Had the color been a property of the ion only, or of the molecule only, the differences found would have been very great, in opposite directions.

Ewan calculates from the numbers found by Kohlrausch that in a solution of cupric sulphate containing 2:38 gram equivalents to the liter, the dissociation amounts to 15-3 per cent. A dilution reducing the proportion to 0.2856 equivalents to the liter increases the dissociation to 3107 per cent. Therefore, if the color depended upon the ions only, the total absorption would be more than doubled. On the other hand if the color depended upon the molecules only, the absorption would be materially diminished. Now the work of Ewan, of Glan, and of previous observers shows that neither of these large changes takes place but that the absorption varies between narrow limits. Therefore it is certain that the color belongs to the atom, whether it exists as an ion or whether by union with a dissimilar ion, it forms part of an electrolyte.

These proofs would seem to be sufficient but others can be had in abundance. In dilute solutions of cobaltous salts the ions exhibit the color characteristic of cobalt. But cobalt cyanide also exhibits that color ; it is quite anhydrous, no ions can be present and therefore the color must be due to the atoms. Cobalt carbonate exists in two forms, one freely hydrated, the other anhydrous, both show the characteristic color. In the hydrated salt the ions may possibly be dissoci. ated, in the other they cannot be. Nickel cyanide is also anhydrous and yet shows the characteristic nickel color.

* Both Ewan's and Glan's results will be found tabulated in Dr. Ewan's paper in the Philosophical Magazine, vol. xxxiii at p. 336. (1892.)

Chromic chloride after sublimation still shows the characteristic pale violet color of chromic salts.

These instances might be multiplied but they will sufficiently show that characteristic color belongs to the atom as well as to the ion.

When the atom enters into a molecule such as is produced by the union of an elementary kathion with an elementary anion, of course that molecule will not give exactly the same absorbtion spectrum as the free ion : the vibration of the ion is free whereas that of the atom in the molecule is constrained. What is contended for is that in the great majority of cases the color of the ion and that of the atom are substantially the saine. Exceptional cases will occur, difficult to explain under any theory. Chromous chloride gives a blue solution in water, chromous acetate a red one: can therefore the chromous ion be both red and blue?

This case is quite the reverse of the results obtained by Ostwald in his well-known work done on the permanganates.

The green solution when concentrated of copper chloride is a case often cited to show that green molecules may yield blue ions. But this may be explained in quite a different way. It is known that many salts exist in strong solutions, as complex molecules. Hittorf has shown that cadmium iodide in strong solution in water exists as (Cd 1,), Lenz* gives as a formula for the solution of potassium iodide in alcohol

a(KI), B(KI),, c(KI),, etc. the coefficients a, b, c diminishing rapidly.

It is quite probable that the green color of some copper compounds may be due to complex molecules and among them, copper chloride.

Criteria of Color. In making determinations of the colors of ions, and therefore of atoms, I have used the following criteria.

1. When an electrolyte gives a solution in water which is colorless when dilute, both the kathion and the anion are colorless. Thus for example as lithium bromide gives a colorless solution in water it follows that the ions of lithium and bromine are colorless. It has been already said that no relation whatever exists between the color of an atom and that of its molecule. So two colorless bromine atoms form the intensely colored element bromine. If further proof be required of the colorless nature of the bromine atom it is easily obtained by considering that the compounds formed by that atom with

* Mem. de l'Ac. Imp. de St. Petersburg, 30.

all the alkaline and earthy metals are colorless even when not dissociated. As potassium iodide forms a colorless solution in water it follows that the atoms of both potassium and iodine are colorless. Sodium monosulphide gives a colorless solution in water, therefore the atoms of sodium and of sulphur are colorless. But the polysulphides are yellow in solution : it fol. lows therefore that in them a certain portion of the sulphur may exist in a molecular condition.

2. If an electrolyte gives a colored dilute solution in water it is necessary first to consider the constitution of the anion. If this is a single atom, then the color of the solution belongs entirely to the kathion. For it is a very remarkable fact that all elementary anions are colorless. This is all the more curious that the endless number or organic coloring matters are built up chiefly of these colorless anions and of colorless hydrogen.

This criterion enables us to judge as to a large number of atoms. As oxides, fluorides, chlorides bromides, etc, all have colorless anions, whatever color is shown by their dilute solutions must be due to the kathions.

3. Even if the anion is composite, information can often be gained. Many composite anions are colorless, for example So, So that when sulphates give solutions that are colored when dilute the color must be due to the kathion.

The same is the case with nitrates, phosphates, carbonates, etc. In these cases and in many others, knowing the anion to be colorless we are certain that the kathion must contribute any color that is present.

These criteria will afford the means of deciding upon the color of the entire series of elementary atoms.


In the preceding sections proof has been offered that the colors of the atoms are substantially the same as those of the corresponding ions. Especially the fact that a solution of copper sulphate by a dilution which doubles the number of its dissociated ions scarcely shows the slightest change of total absorption is strong, one might say, final.

But as respects classification which will form the chief subject of the present paper it is absolutely uniinportant whether these views be accepted or not. We may classify the elements according to the color and want of color of their ions or quite

« PreviousContinue »