A comprehensive theoretical error analysis of the factors involved in the ckromatographic determination of binary diffusion c0efEcient.s shows that it is hardly possible to obtain a precision of better than 0.5 % if conventional peak measurement techniques are employed. XNTRGD'JCI'IGN In a recent comprehensive survey of gaseous diffusion coetficients (D), Marrero and Mason' considered gas chromatography to give results of average reliability. At a temperature of 300 "K, comparisons with D obtained by other methods showed deviations of up to 4 % with an average deviation of about 2 %_ The two-bulb method is considered to be the most reliable, yielding values within 2% of the actual value and, with exceptionally careful contro12, within 1%. In view of the convenience of the equipment, it is natural to consider whether it would be worth while trying to improve upon the reliability of the ckromatorzmphic method. A thorough theoretical analysis -of the relevant design parameters is indicated. Tkis study is seen as a contribution to tkis end, although the treatment is not exhaustive. Several variants of the method exist which dilTer mainly in tke method by which end-effects are eliminated. A single variant., known in tke literature McLaren3 were used for the measurement of the tortuosity factor in packed columns. As far as we know, this method has not been applied to tke systematic determination of diffusion coefficients per se. The fact that it is relatively unexplored was not, however, the sole motivation for its selcction. It also appears to be more convenient and flexible and avoids tke uncertainties related to tke mass-franfer terms in the plate-height expression which are inherent in the other methods. Tkese advantages have not been assessed quantitatively relative to tke otker methods; it is possible tkat with tke necessary care, these methods 2 C. E. CLO-T. W. SMrJTS, EC_ DE CLERK may be refined to the level of the PA method. The latter is expected, however, to be at least representative of the chromatographic methods.
. The specific aim of this paper is to provide design data for the construction of an apparatus for measuring binary gaseous diffusion coefficients to a prescribed