Atomic Weight of Osmium

Several considerations lead us to the conclusion that the atomic weight of osmium is of the order of 191. The chief reasons may be summarised as follows:

1. Analyses of and vapour density measurements with osmium tetroxide and osmium octafluoride indicate that the osmium in the molecules of these two compounds has a weight of approximately 190. This sets a superior limit to the atomic weight of osmium, which might, however, be a submultiple of this value.
2. The specific heat of osmium is 0.0311.- Assuming, therefore, a mean atomic heat of 6.4, the atomic weight of the metal, according to Dulong and Petit's Law, is approximately 200.
3. Osmium bears a close resemblance to ruthenium in many of its chemical properties; in fact, in certain respects, such as the formation of tetroxides, these two elements are absolutely unique amongst the metals of the platinum group.
   
   Furthermore, a distinct gradation in properties can be traced as we pass from osmium, through iridium to platinum.
   
   There seems little reason to question, therefore, the propriety of placing osmium as the first member of the third series of triads in Group VIII. This postulates an atomic weight intermediate between 184 (at. wt. of tungsten) and 193.1 (at. wt. of iridium), but nearer to the latter value, inasmuch as an intermediate element between tungsten and osmium is missing from Group VII.
4. Alkali hexachlorosmates are isomorphous with the corresponding derivatives of palladium, iridium, and platinum, and may therefore be assumed, by application of the Law of Mitseherlich, to have the general formula M₂OsCl₆.

Again, potassium osmocyanide is isomorphous with potassium ferrocvanide and potassium ruthenocyanide, and may therefore be judged to have the formula K₄Os(CN)₆.3H₂O.

Analyses of these compounds indicate that the atomic weight of osmium is 190.9.

The early determinations, by Berzelius and Fremy, gave very high values for the atomic weight of osmium. Berzelius analysed potassium chlorosmate by reducing the heated salt in a stream of hydrogen, and separating and weighing the osmium and potassium chloride in the residue, his result being as follows:

2KCl:Os::100.00:133.42 whence Os = 198.9

Fremy's value was Os = 199.7, derived from the analysis of osmium tetroxide.

That these values are much higher than the correct one was foretold by Mendeleeff, and confirmed by Seubert in 1888. The latter chemist showed the substantial accuracy of his results in a second series of experiments, published in 1891.

Seubert analysed two salts, namely, ammonium and potassium chlorosmate. The former was heated in a stream of hydrogen, the residual osmium weighed, the volatile chlorides collected, and the chlorine contained therein determined as silver chloride.

6 expts. (NH₄)₂OsCl₆:Os::100.000:43.459 whence Os = 191.25
4 expts. 6AgCl: (NH₄)₂OsCl₆::100.000:51.283 whence Os = 192.23

The latter was first analysed by reduction in a stream of hydrogen, the hydrogen chloride produced being determined by conversion into silver chloride; the residual osmium plus potassium chloride was weighed, the potassium chloride dissolved out, recovered by evaporation, and weighed separately.

2 expts. K₂OsCl₆: 2KCl::100.000:30.987 whence Os = 190.27
2 expts. 4AgCl:K₂OsCl₆::100.000:84.097 whence Os = 191.23

In a second method of analysis employed with this salt the weighed mixture of osmium and potassium chloride, obtained as before, was ignited in a boat in a platinum tube until the potassium chloride had volatilised, a current of hydrogen passing through the tube during the experiment. The residual osmium was then weighed:

9 expts. K₂OsCl₆:Os::100.000:39.553 whence Os = 190.38
7 expts. K₂OsCl₆:2KCl::100.000:30.931 whence Os = 191.11

In 1912 Seybold published two series of analyses of ammonium chlorosmate, the osmium being determined:

1. 5 expts. (NH₄)₂OsCl₆:Os::100.000:43.212 whence Os = 189.4
2. 3 expts. (NH₄)₂OsCl₆:Os::100.000:43.437 whence Os = 191.1

Seybold considers that the second series of analyses is vitiated owing to absorption of oxygen by the osmium.

The Table given by the International Committee for 1918 gives

Os = 190.9.

It will be clear that neither Seubert's nor Seybold's results are of the high standard reached in many modern atomic weight researches. Further determinations of this atomic weight are urgently needed, the value in current use being stated more precisely than is justified by the evidence at present available.