The adsorption of certain gases and vapors on silver iodide and silver sulfide

Thesis (Ph.D.)--Boston UniversityIn gas-solid systems, physical or van der Waals adsorption is involved if the gas remains attached to the surface of the solid by a weak interaction comparable in strength to the forces responsible for deviations of gases from ideal behavior and for liquefaction. If a gas or vapor is adsorbed by a solid, some of the unsymmetrical forces of the surface are satisfied and the free energy of the system is decreased. Likewise, the entropy is decreased because the gas molecules have become restricted in their motion. As a consequence, the heat content of the system is also decreased. In this dissertation the various theories of adsorption have been reviewed, with a literature survey of the work done on the adsorption of water vapor on various surfaces and on the thermodynamics of its adsorption. Most of the isotherms representing the adsorption of vapors on solids have been obtained with porous solids. A great deal of the work that has been done on non-porous solids has not been carried to the saturation pressure of the vapor. In this research adsorption isotherms have been determined and the data analyzed according to the BET, Harkins-Jura and Huttig theories. Study has been made of the adsorption of nitrogen at -195.8° C. and of water vapor at 25.0° C. and 0.0° C. on silver iodide and silver sulfide. The latter were chosen as adsorbents because they had been reported as effective nucleating agents and it seemed reasonable to assume that adsorption was the first step in the process of nucleation. From a comparison of the surface areas of the silver iodide sample as measured by water vapor and by nitrogen adsorption, a decided disagreement is evident. This discrepancy is apparent in both the BET and Harkins-Jura methods. The determination of the surface area by the "point B" method resulted in good agreement in the nitrogen adsorption method but was not too reliable in the method adsorption of water vapor, because the point on inflection at the completion of a monolayer was less distinct in the adsorption isotherm. According to the Harkins-Jura method, the area for the silver iodide sample, as measured by the adsorption of water vapor at 25.0° C., varies from 0.005 to 0.0556 m.^2/g. as compared to 0.0169 m.^2/g. by the BET method. Corresponding differences are noted at 0.0° C. The maximum area obtained by the Harkins-Jura method is about three and a half times the BET value. Nevertheless, it is still only twenty per cent of the area by the nitrogen BET method, 0.271 m.^2/g. The BET and Harkins-Jura theories give fairly good agreement for the area by nitrogen adsorption. The values from the Harkins-Jura method vary from 0.2030 to 0.3086 m.^2/g. as compared to 0.271 m.^2/g. by the BET method. Analysis of the data according to the Huttig theory gave values for the surface area of 0.0228 m.^2/g. and 0.0176 m.^2/g. at 25.0° C. and 0.0° C., respectively. Although these are larger than those calculated according to the BET theory, there still remains the same disagreement with the nitrogen area, 0.381 m.^2/g. The differences in surface areas cannot be attributed to an alteration of the surface brought about by heating, because a sample was prepared which was dried and degassed at room temperature. Nitrogen and water vapor adsorption experiments carried out on a standard sample of anatase with a reported specific surface area of 13.8 m.^2/g. warranted elimination of the apparatus as a possible source of error. The nitrogen adsorption method gave excellent agreement with the area reported, while the water vapor adsorption method gave a deviation of less than twenty per cent. This could in no way account for the difference found in the case of the silver iodide. Since the two adsorptions in question were carried out at temperatures very far apart, it was decided to determine further the temperature behavior of the apparent surface area. Measurement of the surface area by the adsorption of carbon dioxide at -78.5° C. gave a value of 0.1409 m.^2/g., slightly more than half the surface area as determined by nitrogen adsorption. This seems to indicate that the apparent surface area is temperature-dependent. At room temperature there are definite deviations from the wurtzite structure of silver iodide due to the motion of a silver ion toward three of the four surrounding iodide ions and away from one of them. At a lower tanperature the silver ions being less mobile provide adsorption sites for fairly "localized" adsorption, while at higher temperatures, it would seem probable that the number of available adsorption sites would decrease, particularly for a polar adsorbate. The applicability of the BET theory in this case may be questioned. The method was originally worked out in a narrow and very low temperature range and for the adsorption of gases close to their boiling points, and one of the primary assumptions of the theory is that the adsorbate is non-polar. Although both the silver iodide and the water molecules are predominantly covalent, they are polar, and a certain amount of orientation would be expected in the adsorption of water vapor. This would not be the case in the adsorption of nitrogen. It could be that all the surface physically available to nitrogen molecules would not be effective in providing adsorption sites for water molecules. Also water forms strong hydrogen bonds in the liquid state and is an outstanding example of substances which show marked deviations from typical adsorption isotherms, especially at low pressures. In the case of silver sulfide, there is a difference between the surface areas as determined by nitrogen and water vapor adsorption but it is not nearly so pronounced as that found for silver iodide. The surface area of silver sulfide as measured by water vapor adsorption is about sixty per cent of the area obtained by the method of nitrogen adsorption. The Huttig values for the area are again higher than the BET values. The thermodynamic quantities, ΔH, ΔF, and ΔS, were first calculated from isosteres corresponding to a constant amount of water vapor adsorbed. In the case of both adsorbents, silver iodide and silver sulfide, the heats of adsorption are less than the heat of liquefaction of water at 25.0° C., that of silver aulfide being about 0.5 kcal. less than that of silver iodide. The decrease in free energy is accompanied by a decrease in the entropy of adsorption approaching the entropy of liquefaction of bulk liquid. The values seem to indicate that the state of the adsorbed water is more gas-like than liquid-like. However, most theories ascribe liquid-like properties to the adsorbate, and the BET analysis gave an energy of adsorption exceeding the heat of liquefaction, the reverse of the thermodynamic treatment. When the thermodynamic data were recalculated after determining the equilibrium pressures corresponding to a given fraction of surface covered, the resulting values indicated that the adsorbed layer is more liquid-like in nature. Studies of the dielectric behavior and of the infra-red spectrum of adsorbed water by other investigators support the postulate of liquidlike properties for the adsorbed layer. A preliminary experiment was carried out to determine what would be the effect on the adsorption by silver iodide if a small amount of ammonia were added to the water vapor. Isotherms were constructed which related the volume adsorbed to 1) the total equilibrium pressure, 2) the relative humidity of water vapor, and 3) P/P0, where P was the total equilibrium pressure and P0 was the equilibrium pressure over an ammonia-water solution of the composition which would be in equilibrium with the original gas phase. The first isotherm does not give any selective information and neither of the other two would be expected to represent the actual behavior. First of all, the adsorption of water vapor is definitely enhanced by the presence of the ammonia (the BET value of the surface area is one and a half times that for water alone). However, some ammonia must be adsorbed and there must be a change in the relative composition of the vapor as the adsorption proceeds