Critical study of the determination of ethane by combustion over platinum in the presence excess oxygen

A sample of ethane of known purity was analyzed by the usual slow-combustion method; that is, over a hot platinum spiral in the presence of excess oxygen. The contraction after burning, the carbon dioxide produced, and the oxygen consumed were measured. The average results obtained from a series of 18 determinations indicated the impossibility of attaining the accuracies usually reported in the literature. The average volumetric equation for the reaction was C 2H 6 -|-3.5116 2 «= 1.9940 COii+ 2.5190 contraction. The weighted average eauation was CoH 6+ 3.513 2 «= 1.994 CO2+ 2.520 contraction. The contraction and oxygen consumed were invariably too high, and the carbon dioxide produced was invariably too low. The ethane was calculated from contraction, carbon dioxide, and oxygen separately and in various combinations, both uncorrected and corrected for deviations from theoretical molecular volumes. The average results from different methods of calculation varied from 99.13 percent to 100.76 percent. The best average value found by use of a single formula was 100.11 percent, and resulted from combining the contraction and carbon dioxide as one factor of an equation of which the other factor was oxygen. Such an equation eliminates part of the uncertainty concerning the proper method of correcting for deviation from theoretical molecular volumes. The averages of two pairs of formulas gave results very close to 100 percent. The ethane was also calculated on the assumption that methane, or propane, (known to be absent) was present. (The usual analysis would have been reported as methane and ethane.) Hydrocarbons other than ethane indicated by various methods of calculation from the average of the observed results varied from 0.3 percent of propane to 9.4 percent of methane. A number of sources of error are discussed. These include small measured losses of ethane and carbon dioxide in the distributor, and a small loss of carbon dioxide by solution in the water produced by the combustion and otherwise present. Such errors must be present in every apparatus of this type, and will effect the analysis of all hydrocarbons to some extent. The observed results may be explained by correcting for these errors, and for the known deviation of carbon dioxide from its theoretical molecular volume with a similar correction for ethane, which is somewhat higher than the best estimated value available.