Abstract
Volatile methyl compounds are formed by the majority of elements standing up to six places before a rare gas, so that as a class they represent all groups and periods of the Periodic System with the exception of Group I. Accordingly, the volatile methyls provide a particularly suitable class of compounds for a comparative study of the covalent bonds which the metals and non-metals concerned form with carbon; and this study in turn throws light on the magnitudes of the internal excitation energies possessed by these elements when engaged in covalent bonding. Former approaches to this problem have been almost entirely theoretical. For convenience, the term" metal-carbon bond" will be used through­ out this paper to denote all bonds of the type stated, whether the element linked to carbon is a metal or not. MEAN DISSOCIATION ENERGIES The heats of formation of the methyl compounds are required for calculating the mean metal-carbon dissociation energies, and, in the case of the highly reactive, spontaneously inflammable liquids or gases under discussion, their determination has provided unusual experimental diffi­ culties. Two distinct lines of approach to this problem have proved successful. One, originally due to Skinner and his co-workers, has made use of bromination or iodination in non-aqueous media in specially designed apparatus. The other, employed by several investigators (in­ cluding the present author), relies on a special technique for opening thin-walled glass ampoules in oxygen under pressure in a bomb calorimeter. Waddington and his collaborators have improved on this by introducing a rotating bomb, and this has been applied to the cases of the methyls of lead and sulphur. With all methods, considerable difficulty has been encountered in obtaining pure or even complete reactions, and a careful analysis of the products is in general necessary to obtain reliable heats of reaction. In some reactions a small element of uncertainty is introduced through the possibility of solid products being produced in an energy-rich form or doubtful crystalline state. Notwithstanding these difficulties, it has usually proved possible by these methods to obtain heats of formation of the methyl compounds with an accuracy that corresponds to limits of error of ± 1 kcal or less in the mean dissociation energies of the metal-carbon bonds. However, in a number of cases uncertainties in auxiliary thermochemical data, especially heats of atomization, introduce larger possible errors.