66(5 TRANSACTIONS OF SECTION B. 



carbon monoxide with other metals, notably with cobalt, iron, copper, and platinum, 

 have only led to negative results.' In a later paper,' published by one of us in 

 conjunction with Dr. F. Quincke, we announced the formation of iron carbonyls, 

 which were subsequently more fully described by one of us in conjunction with 

 J)r. Carl Langer.'^ 



In continuing the experiments with cobalt, and taking advantage of the method 

 of Sir James Dewar for facilitating the formation of nickel carbonyl, by acting upon 

 the finely divided metals by carbon monoxide under a high pressure and corre- 

 spondingly high temperature, instead of working at ordinary pressure and ordi- 

 nary temperature, as we had bitherto done, we succeeded in obtaining small quan- 

 tities of a cobalt carbonyl when the pressure was raised to 50 atmospheres and the 

 temperature to 160° C, and the quautity formed was considerably increased by 

 raising the pressure to 100 atmospheres and the temperature to 200° C. 



The cobalt oxide was prepared from a pure oxalate, free from nickel and iron, 

 by heating to about 200° C. The cobalt oxide was filled into a steel retort, 

 which was fitted with two steel pipes for the gas circulation. All iron parts were 

 lined with copper, so as to prevent the action of carbon monoxide on the iron. This 

 retort was heated in an oil bath. Behind the outlet valve we fixed a small gun- 

 metal cylinder, which contained a filter of flock asbestos, and behind this we had 

 an additional filter of cotton-wool in a glass tube. From there the gases had to 

 pass a condensing bottle, and the spent gases were burnt at the outlet. The cobalt 

 oxide was placed into the retort and reduced with hygogen at 6-10 atmospheres 

 pressure, and at the lowest possible temperature. A slow current of carbon mon- 

 oxide was then passed through th^ apparatus at 100 atmospheres pressure. The 

 temperature of reaction lies between 150° C. and 200° C. Both filters were kept 

 at 50° 0. by means of warm water. The condensing bottle was cooled in ice. In 

 this bottle we obtained large orange crystals of the cobalt carbonyl. 



The analysis of these crystals was made by decomposing them with bromine 

 water. The carbon monoxide was measured, and the cobalt estimated in the 

 solution. The result was : — 



niM _j .' Co. CO. Co/CO. 



l:.3-96 

 ] :4-0.3 



"We conclude therefore that this compound isCo(CO)^, corresponding to the liquid 

 nickel carbonyl. 



The cobalt carbonyl is gradually decomposed in air, leaving a deep violet sub- 

 stance, which has not yet been investigated. The crystals are best preserved by 

 sealing them in a glass tube under hydrogen or carbon monoxide. The cobalt 

 carbonyl is very slawly attacked by non-oxidising acids, as hydrochloric or sul- 

 phuric. The reaction is accelerated by the presence of an oxidising agent. The 

 carbonyl is quickly, almost instantaneously, decomposed by oxidising acids, such 

 as bromine, chlorine, nitric acid, and the corresponding cobalt salt is formed under 

 elimination of carbon monoxide, according to the following equation: — 



Co(CO), + Brg = CoBr^ + 4C0. 



The cobalt carbonyl is insoluble in water, but more or less soluble in the several 

 organic solvents, such as carbon bisulphide, ether, naphtha, alcohol, and nickel 

 carbonyl. If these solutions are kept standing for some time, or if they are 

 warmed, decomposition sets in. These and other solutions and their products of 

 decomposition have not yet been investigated. 



The specific gravity measured by the suspension method at 18° C. is 1*827, but 

 we do not regard this figure as final, as the substance is so easily decomposed, and 

 if only kept for a few hours the specific gravity increases considerably, due to 

 decomposition. The vapour tension in vacuo could not be measured with certainty, 

 owing to decomposition. The vapour tension at 15° ^C. measured 11 mm., but 



^ Jovrnal of the Chemical Society, 1891, p. 604, 

 « Jbid,, p. X091. 



