Atomic Weight of Cobalt

That the atomic weight of cobalt is approximately 59 and not a multiple or submultiple of this amount is evident from a variety of considerations, namely:

1. The specific heat of cobalt is 0.108. Assuming a mean atomic heat of 6.4, the atomic weight, according to Dulong and Petit's Law, is approximately 59.3.
2. The most appropriate position for cobalt in the Periodic Table is, as explained in the opening chapter of this volume, between iron and nickel. Assuming the atomic weights of iron and nickel to be correct, we should expect the value for cobalt to be greater than 55.84 (at. wt. of iron), but less than 58.68 (at. wt. of nickel).
3. The cobalt alums are isomorphous with those of iron and aluminium, and therefore, by the application of Mitscherlich's Law, they must be assumed to contain two atoms of cobalt, their generic formula being:
   
   M₂SO₄.Co₂(SO₄)₈.24H₂O.

Again, double cobaltous sulphates with alkali sulphates are isomorphous with those of nickel and divalent iron, from which it may similarly be gathered that their formulae are given by the generic scheme:

M₂SO₄.CoSO₄.6H₂O.

Analyses of these compounds indicate that the atomic weight of cobalt is 59.0.

Until 1857 the accepted value for the atomic weight of cobalt was derived from a single experiment made by Rothoff in 1826; 269.2 parts of CoO were found to be equivalent to 1029.9 of AgCl, whence Co = 58.9.

In 1857, Schneider analysed cobalt oxalate, determining the carbon dioxide obtained by its combustion and the metal left after igniting the salt in air and subsequently in hydrogen. Four analyses gave the following result:

Co:2CO₂::100.000:146.665 whence Co = 60.005

The following year Marignac made a number of preliminary experiments on the atomic weight of cobalt. In two experiments anhydrous cobalt sulphate was calcined to oxide:

CoSO₄:CoO::100.000:48.287 whence Co = 58.761

In another four experiments anhydrous cobalt chloride was analysed for chlorine by titration with silver:

2Ag:CoCl₂::100.000:60.118 whence Co = 58.797

In I860 Dumas' analyses of anhydrous cobalt chloride appeared. Five analyses gave the following mean result:

2Ag:CoCl₂::100.000:60.228, whence Co = 59.034

In 1863 Russell reduced pure cobaltous oxide to the metal in a stream of hydrogen, and as the mean of fifteen experiments obtained the following result:

CoO:Co::100.000:78.592, whence Co = 58.738

Three years later, Sommaruga made seven analyses of purpureo-cobaltic chloride by heating it until cobaltous chloride remained, and reducing this salt in a stream of hydrogen:

Co(NH₃)₅Cl₃:Co::100.000:23.827, whence Co = 59.909

In 1867 Winkler made five determinations of the amount of gold precipitated from sodium aurichloride solution by a given weight of cobalt:

2Au:3Co::100.000:45.172, whence Co = 59.386

In 1868, Weselsky analysed ammonium and phenylammonium cobalti-cyanides for cobalt by ignition in a stream of hydrogen, two experiments with the first and four with the second salt, giving the following results:

(NH₄)₃Co(CN)₆:Co::100.000:21.943, whence Co = 59.085

(C₆H₅.NH₃)₃Co(CN)₆:Co::100.000:11.8665, whence Co = 59.017

In 1869, Russell redetermined the atomic weight of cobalt by dissolving the metal in hydrochloric acid and measuring the hydrogen evolved. Computing his results with the modern value for the density of hydrogen, his eleven experiments give the following mean result:

Co:H₂::100.000:3.4112 whence Co = 59.077

In 1871, Lee determined the percentage of cobalt in brucine cobalti-cyanide, strychnine cobalti-cyanide, and purpureo-cobaltic chloride:

(C₂₃H₂₆N₂O₄)₃H₃Co(CN)₆ .10H₂O:Co::100.000:3.7437 whence Co = 59.199

(C₂₁H₂₂N₂O₂)₃H₃Co(CN)₆.4H₂O:Co::100.000:4.5705 whence Co = 59.096

Co(NH₃)₅Cl₃:Co::100.000:23.5795 whence Co = 59.095

In 1886, Zimmermann made ten extremely concordant analyses of cobaltous oxide by reducing it to the metal in a stream of hydrogen:

CoO:Co::100.000:78.635 whence Co = 58.889

Another series of twenty-four experiments on the reduction of cobaltous oxide was published shortly afterwards by Remmler:

CoO:Co::100.000:78.613 whence Co = 58.812 In 1893-1894, Winkler published the results of his experiments on the atomic weight of cobalt. First, pure cobalt was converted into the neutral chloride, and in six experiments the chlorine determined gravimetrically as silver chloride:

2AgCl:Co::100.000:20.864 whence Co = 59.812

In a second series of six experiments the chlorine in the neutral chloride was determined volumetrically by Volhard's method after removing the cobalt with potassium carbonate. The ratio of cobalt to silver was thus established and checked by two experiments in which cobalt was allowed to displace silver from silver sulphate solution:

2Ag:Co::100.000:27.705 whence Co = 59.776

In a third series of eight experiments a weighed amount of cobalt was dissolved in excess of a standard solution of iodine in potassium iodide, the excess of iodine being then titrated with sodium thiosulphate:

I₂:Co::100.000:23.462 whence Co = 59.556

In 1895, Hempel and Thiele's determinations appeared. Three experiments on the reduction of cobaltous oxide gave the following result: CoO:Co::100.000:78.666 whence Co = 58.998

In another series of seven experiments a weighed amount of cobalt was converted into the anhydrous chloride, and the increase in weight noted:

Cl₂:Co::100.000:82.873 whence Co = 58.769

Finally, in four of the preceding experiments the chlorine in the anhydrous chloride was estimated as silver chloride:

2AgCl:Co::100.000:20.556 whence Co = 58.929

The foregoing results are of little or no value in comparison with those that have been subsequently obtained. They vary very considerably, but undoubtedly point to a value round about 59 for the atomic weight of cobalt. The results obtained by Schneider, Sommaruga, and Winkler are much higher than all the others, and it will be noticed that in the hands of Lee, Sommaruga's method gave a much lower value.

In 1897-1899, the work of Richards and Baxter on the atomic weight of cobalt was published. Anhydrous cobaltous bromide was prepared by direct union of the pure elements and the product sublimed in a stream of hydrogen bromide and nitrogen. The bromine in the salt was determined by the two usual methods, the amount of silver required to precipitate it being first measured and then the precipitated silver bromide being collected and weighed. The final results were as follow:

8 expts. 2Ag:CoBr₂::100.000:101.407 whence Co = 58.964

9 expts. AgBr:CoBr₂::100.000:58.255 whence Co = 58.969

These results were confirmed by a series of experiments in which the anhydrous bromide was reduced to metallic cobalt by heating it in hydrogen. Twelve experiments gave the following mean result:

CoBr₂:Co::100.000:26.952 whence Co = 58.972

It is seen that the atomic weight thus obtained is practically the same as that derived from the percentage of bromine in the salt, an indication of the high degree of accuracy of the experimental work.

Richards and Baxter also analysed the chloride and oxide of cobalt. It was not found possible to prepare the anhydrous chloride wholly free from alkali chloride and silica, so that their two reductions of the chloride to the metal are slightly in error:

Cl₂:Co::100.000:83.266 whence Co = 59.047

The analyses of cobaltous oxide were also regarded with suspicion, owing to the difficulty of ascertaining whether the oxide is free from excess of oxygen. Five experiments were made:

CoO:Co::100.000:78.659 whence Co = 58.973

With the subsequent invention of fused quartz vessels it became possible to prepare pure anhydrous cobalt chloride. Accordingly, in 1906, Baxter and Coffin prepared and analysed this salt in the customary Harvard manner, the results being as under:

8 expts. 2Ag:CoCl₂::100.000:60.1975 whence Co = 58.968

7 expts. 2AgCl:CoCl₂::100.000:45.3070 whence Co = 58.969

The excellent agreement between Richards and Baxter's analyses of cobalt bromide and Baxter and Coffin's analyses of cobalt chloride leaves no doubt as to the accuracy of the value

Co = 58.97

for the atomic weight of cobalt, and this is the value given in the International Committee's table for 1918.