No abstract

No abstract

The experiments of Meehan having shown that charcoal increases in volume when taking up carbon dioxide, it became an obvious step to correlate this expansion with the quantity of gas adsorbed. In this paper a form of extensometer is described which, while sufficiently sensitive and accurate for the measurement of the percentage linear expansion ( x ) of a rod of wood charcoal, was yet compact enough to permit its being sealed up in an all-glass apparatus, so that simultaneous determinations could be made of the weight ( s ) of gas taken up per unit weight of the absorbent. The graphs obtained by plotting the variables x and s have certain interesting characteristics which it will be convenient to describe before dealing with the detailed results. As fig. 1 shows, the coefficient dx/ds increases steadily with s ; thus far the charcoal behaves similarly to many of the common gels, which contract proportionately less in the later stages of drying, owing to the micellæ coming in contact with each other. The swelling of charcoal, however, appears to follow a more regular course than that of the elastic gels, and in the case of ammonia, carbon dioxide, and sulphur dioxide is given within the limits of accuracy of our experiments by hyperbolic equations of the type x = k s /S— s ... where k and S are constants characteristic of each gas. In fig. 1 the experimental values of x have been plotted (for convenience of representation) against s /S. The smooth curves for ammonia of the above form, using appropriate values of the constants. An additional term (a small additive constant over the experimental range) is necessary to represent the behaviour of water vapour.

The experiments described in Part I of this series led to the conclusion that the expansion which charcoal undergoes in contact with water vapour, carbon dioxide, ammonia, and sulphur dioxide is connected with the existence in the adsorbed phase of a pressure due to the free movement of the molecules in directions parallel to the interface. In a later paper it was pointed out that if the relation between the percentage linear expansion ( x ) and the surface pressure (F) be assumed to take the form x = λ F, (1) where λ is a constant, then the equation found empirically to relate the ad­sorption and expansion in the case of these gases reduces to an expression of the Schofield-Rideal equation of state: F (A – B) = i RT. (2) Here A stands for the area per (gram) molecule, and is equal to M∑/ s if s is the weight in grams of an adsorbate (of molecular weight M) taken up by 1 gram of charcoal the effective surface area of which is ∑; B is the “incom­pressible” cross-sectional area per gram molecule; and i is a constant determined by the mutual cohesion of molecules and is equal to unity when such cohesion is absent.

When a rod of well-baked wood charcoal contained in an extensometer is plunged into methyl alcohol, a noise like the quenching of hot iron is produced, and the moving parts of the extensometer are violently disturbed, as by a sudden impact. Though such effects are to be classed as irreversible and even irreproducible (where the charcoal is visibly fractured), quantitative investigation shows the expansion phenomenon to be no more inherently so than are the effects which accompany the admission of gas to an evacuated vessel. There can be no doubt that the primary factor determining the violence of the effect is the ability of the adsorbed molecules to spread over the surface of the charcoal, not (as has been suggested by analogy with other systems) in a series of infrequent jumps, but with great rapidity. With increase of molecular weight of the alcohol, the effect becomes less violent, especially if only one end of the charcoal is immersed in the liquid; but in none of the systems so far examined has a time-lag been found between adsorption and expansion. Granted the free mobility of a fraction of the molecules forming the adsorbed phase on charcoal, mere economy of hypothesis demands that one should enquire, before making ad hoc assumptions as to the nature of the charcoal surface, whether or not the phenomena are quantitatively similar to the better understood effects which occur at the surfaces of liquids; the present communication records an attempt to provide an answer to this question by appeal to experiments carried out with a series of adsorbates of closely related constitution. On the basis of the results described in this and in earlier papers, it can be asserted without hesitation that the two sets of phenomena do in fact run parallel to a remarkable degree. No doubt remains that the expansion actually provides a measure of the lowering of the surface energy of the charcoal mass; this is comprehensible, so far as when wood is carbonized the tendency of the carbon to reduce its surface energy to a minimum must be a primary cause of its shrinkage, and must determine, among other factors, the extent to which this shrinkage takes place. Further, a direct comparison of our expansion data with those for the lowering of the surface tension of mercury by the vapours of the same alcohols brings to light so many points of resemblance as to indicate a similarity of mechanism in the two cases. No evidence whatever was found in favour of the view (of which one of the present authors was an original advocate) that the swelling is a lattice distortional effect caused by sessile adsorbed molecules; and though the investigation of other systems may reveal the existence of an expansion effect where the molecular motions are confined to isolated areas of low potential (and so make it necessary ultimately to reconcile the two conflicting viewpoints), such a development is not foreshadowed in the data here presented.