Infinite dilution activity coefficients (y") in binary liquid mixtures are data of wide practical and theoretical applicability, but their extensive use has been hindered by the experimental limitations of classical vapor-liquid equilibrium techniques. Two new experimental methods for accurate y" determinations are presented. These techniques, both faster and more accurate than classical methods, are complementary, and their applicability to different types of binary systems is discussed. SCOPEThe modern design and modeling of fluid-phase separation equipment require an accurate mathematical representation of chemical activities in a multicomponent liquid mixture. Solution models now use binary data only to represent a multicomponent mixture, and have proven central to the rational design of distillation, absorption, extraction, stripping, or even crystallization units. Although many sets of binary data exist in the literature, the limiting factor in most multicomponent process designs is in the acquisition of valid binary data needed to determine parameters in the mathematical models. For most miscible organic systems, the existing data are in the form of classical VLE (vapor-liquid equilibria) measurements, which are intrinsically slow and difficult experiments. Further, the resulting data, even if accurate, do not contain the most useful information either for a basic understanding of solution behavior or for practical application.It is the purpose of the current study to demonstrate a different approach to gathering the data needed for studying the solution thermodynamics of binary liquid mixtures as well as for designing multicomponent separations. The advantages of using infinite dilution activity coefficients were first proposed over twenty years ago, and their use in design and scale-up has often been demonstrated. However, the actual measurement of such quantities by classical means-either from VLE or LLE data-is very imprecise, and until recently newer experimental methods have been limited in applicability. This work presents two complementary experimental methods for measuring limiting activity coefficients in a wide variety of organic systems. Since these methods are both easier to use and intrinsically more accurate than the classical techniques, the use of limiting activity coefficients presents useful, new experimental and analytical approaches. CONCLUSIONS AND SIGNIFICANCELimiting activity coefficient (7%) data are most effective both in the description of binary mixtures and in the scale-up to multi-component systems. The limiting activity coefficient characterizes the behavior of a single solute molecule completely surrounded by solvent. As such, it generally indicates a maximum nonideality and offers incisive information to the theorist, since the order-disorder effect disappears; it also offers economy of effort to the experimentalist, since this datum has wider applicability than a measurement at any other concentration.The major reason that yx methods have been little used lies in the dif...

d/(g cm"3) = 0.91694 -(8.149 X 10'4Xf/°C) -(5.1 X 10"7Xf/°C)2 = 3 X 10"5 g cm"3(1) where d is the density at temperature t. Values calculated from eq 1 are listed in column 3 of Table I. The good agreement with the experimental values demonstrates the concordance of both data sets. Column 5 contains literature data. The single value at 27 °C reported by Scatchard and Wilson is in excellent agreement with our value interpolated by using eq 1; these authors used essentially the same purification procedure as we used. The four values of Onken ( 7) are in poor agreement with our values. This discrepancy in d and the previously mentioned discrepancy in n2^a re consistent with the presence of a small amount of water contamination in his samples. The effect of this would be to lower the refractive index and raise the density in the way observed. Glossary d density, g cm"3 t temperature, °C standard deviation Literature Cited (1) Guzman, F.; McLure, I. A. Int. Data Ser. B 1980, 109.(2) Weast, R. C" Ed. "Handbook of Chemistry and Physics", 55th ed.:

Infinite-dilution activity coefficients were measured for 147 systems using an Improved differential ebulllometrlc technique. The results compare well with the limited literature data available. The observed temperature dependence of the data was also found to be reasonable.

We examined the fouling and corrosion that took place when 316 stainless steel and pure iron wires were electrically heated to 540À680°C in a liquid bath of the atmospheric bottoms fraction of a crude oil. The foulant was determined to be heterogeneous, with a thick macroscale outer layer of pitch, covering a microscale sheath of coke, which was in turn both covering and interspersed with a microscale layer of iron sulfide. This foulant was observed to delaminate from the wire surface, presumably as a result of both the generation of growth stresses and the action of gas bubbles that were evolved during the fouling process. Unexpectedly but conclusively, we observed that the underlying wire surface was heavily corroded. In the case of the stainless steel, we observed a microscale chromium oxide layer that separated the foulant from the underlying metal. This layer presumably reduced the rate of metal dissolution. The degree of corrosion was much higher in the pure iron samples, where such a layer did not exist. Our hypothesis is that there is a synergy between the measured macroscopic fouling and the underlying microscopic corrosion, where the iron from the wire reacts with the sulfur in the oil to build up the thick sulfide.

No abstract