Experimental gas-solid adsorption measurements were made on the binary and ternary mixtures of H,S, CO,, and C,H, on H-mordenite molecular sieve zeolite at 30°C. The C, H, containing binaries and the ternary mixture exhibited nonideal behavior, as evidenced by the azeotropelike crossovers in the composition domains and by the peaks in the total amount adsorbed surface. The ternary data were successfully predicted with the activity coefficients calculated by the spreading-pressure-dependent equation. Models that neglect the spreading-pressure dependence can only qualitatively predict the nonidealities.

The ideal adsorbed solution (IAS) theory of adsorption of gas mixtures is extended to the case of energetic heterogeneity. A heterogeneous ideal adsorbed solution (HIAS) behaves ideally on a particular site, but energetic heterogeneity causes a segregation in the composition of the adsorbed phase. Equilibrium properties are obtained by integrating over a multivariate energy distribution based upon perfect positive correlation of site energies. The fact that predictions from HlAS are always an improvement over IAS indicates that heterogeneity is a factor that must be considered in theories of mixed-gas adsorption. Errors in HlAS predictions may be caused by steric exclusion of the larger molecules from micropores accessible to smaller molecules. IntroductionLangmuir (191 8) was the first to recognize the importance of the spatial variation of the potential energy of adsorption, and Ross and Olivier (1964) established the methodology for studying heterogeneity of solid adsorbents in terms of an adsorption integral equation. Several recent reviews (House, 1983; Jaroniec and Brauer, 1986) were devoted to surface heterogeneity and numerical techniques for solving the integral equation for the energy distribution. Hoory and Prausnitz (1967) were the first to extend the adsorption integral equation to mixtures. They studied the bivariant Gaussian distribution, and Jaroniec and Borbwko (1 977) used the bivariant log normal distribution. In both cases the appropriate energy distribution is derived from experimental data for single-gas adsorption, and a correlation parameter p is used to fit data for binary gas adsorption. For adsorbates with similar chemical characteristics, such as CH4 and C2Hs, a strong correlation ( p = 1) between the energy distributions was obtained. A value of p = 0 means no correlation between adsorptive energies and randoin matching of sites. In the random case the bivariant energy distribution is the product of the energy distributions of the individual gases (Jaroniec, 1975), but this is a "gross approximation" (House, 1983) and unrealistic from a physical point of view. This paper is a continuation of the work cited above. The HlAS method developed in this paper is thermodynamic in nature, in the sense that it is not associated with a specific energy distribution or a particular equation for the pure-gas adsorption isotherm. Adsorption Integral EquationThe specific amount N (mol/kg) of a pure gas adsorbed on a heterogeneous adsorbent at pressure P and temperature T is given by the integral ( Ross and Olivier, 1964):is the probability density function for the distribution of adsorptive energies, and n is the specific adsorption if the surface were homogenous and characterized by energy t.The individual amount Ni of ith component adsorbed on a heterogeneous surface from a gas mixture containing r components is (Jaroniec and Rudzinski, 1975): (C1, €2,. . . , e,) dtl dt,. . . dt, (2) g(tl, t2, . . . , t,) is the joint probability density function for the distribution of adsorption energies and y ...

Generalized depletion curves for desorption (and corresponding breakthrough curves for adsorption) were calculated for systems characterized by the Langmuir-type multicomponent equilibrium equation and controlled by the film type rate model. In contrast with adsorption where the nonkey (or less strongly adsorbed) component curves display overshoots above feed concentration, in desorption the key component depletion curves exhibit the instabilities in the form of inflections and curvatures. As in the one component case, the differences in the depletion and breakthrough curves may be related to the rate phenomena. The undulations in the key component depletion curves may be characterized by derivatives of the rate data. The major significance of these instabilities is to elongate the depletion curves, which in turn requires the expenditure of added effort during regeneration. Process modifications are indicated, which could suppress the instabilities. The predicted trends were c o n k e d by experimental depletion curves. SCOPEMost of the studies dealing with the design of sorption step in the operation of the cyclic processes. Recently it based separation processes have focused on the description was shown that in systems characterized by favorable isoof the adsorption step; and even these have been confined therms (d2W/dC2 < 0) the depletion curves are considermostly to single component systems. From experience, ably elongated in comparison to the corresponding breakhowever, the desorption is known to be the controlling through curves. Mechanistically the elongation may be

No abstract

No abstract