Fluid Catalytic Cracking Selectivities of Gas Oil Boiling Point and Hydrocarbon Fractions Robert H. Harding,* ,† Xinjin Zhao, ‡ Kuangnan Qian, † Kuppuswamy Rajagopalan, ‡ and Wu-Cheng Cheng ‡ W. R. Grace & Co.sConn. and Grace Davison, Washington Research Center, 7500 Grace Drive, Columbia, Maryland 21044 The product selectivities of the fluid catalytic cracking (FCC) process are strongly dependent on the properties of the petroleum gas oil reactant. In order to elucidate the complex relationship between gas oil chemical composition and product selectivity, a new technique has been developed which experimentally determines the product distribution of specific gas oil fractions in a realistic chemical environment. This “incremental yield analysis” approach is examined with a representative industrial gas oil. The gas oil is characterized and then divided into boiling point fractions by distillation and into hydrocarbon-type fractions with a chromatographic method. The FCC selectivity of each gas oil fraction is then determined by microactivity testing of blends of each fraction with the original gas oil. Results show that hydrocarbon type is a more significant determinant of the product spectrum than boiling point. Introduction Fluid catalytic cracking (FCC) is an important indus- trial process which is used to upgrade heavy petroleum gas oils into gasoline, diesel fuel, light olefins, and other valuable products (Venuto and Habib, 1979). Naturally- occurring gas oils are a complex blend of hydrocarbons, including paraffins, isoparaffins, naphthenes, aromatics, and asphaltenes. A typical gas oil also contains signifi- cant quantities of multiringed molecules containing heteroatoms, such as nitrogen and sulfur. Worldwide gas oil reserves are evolving toward higher molecular weight crudes and a greater concentration of hetero- atoms with continued use. Therefore, a great deal of current research is oriented toward the development of catalysts and processes for cracking the heavy end of the gas oil (Mitchell et al., 1993). In this paper, we explore the FCC cracking kinetics and selectivities of a gas oil as a function of boiling point range and feed composition. Past research in this area has tended toward correlating the properties of a wide range of gas oils with the products produced by a single catalyst (Fisher, 1990; Pachovsky and Wojciechowski, 1975). Historically, it has been difficult to deconvolute the effects of molecular weight and hydrocarbon type, since these properties tend to correlate in natural gas oils. The kinetic rates and selectivities of gas oil during cracking are difficult to determine due to the convolution of the kinetics with deactivation, diffusion, competitive adsorption, and bimolecular interactions such as hy- drogen transfer (Pine et al., 1984). In order to gain a better understanding of the determinants of FCC selectivities, researchers have studied a hierarchy of hydrocarbon mixtures. The simplest and most complete experiments have focused on the catalytic cracking of single model compounds (Townsend and Abbot, 1993; Corma et al., 1992). These experiments have determined the primary kinetic path- ways for many of the individual components in the gas oil. These model compound experiments have eluci- dated the basic carbenium ion pathways of FCC (Venuto and Habib, 1979) and have explored the role of second- ary pathways, which depend on carbonium (Haag and Dessau, 1984) and cyclopropenium (Sie, 1993) interme- diates. Although a significant quantity of kinetic in- formation has been derived at this level, studies of single model compounds do not completely address the selec- tivity of that model compound in the gas oil environment or the interaction of gas oil molecules (Martin et al., 1989). Model compound tests also tend to be reported at low coke levels. The variation of coke levels makes rate comparisons difficult (Lin et al., 1992). In addition, since model compounds representing high molecular weight gas oil species are difficult to obtain, these experiments do not typically address the role of gas oil boiling point or issues such as concarbon coke produc- tion. A second tier of experiments have used mixtures of model compounds to explore FCC cracking kinetics and selectivities. Mixtures of typically 2-5 individual model compounds have given much insight into the role of competitive adsorption (Abbot and Wojciechowski, 1987; Santilli and Zones, 1990), initiation (Abbot, 1990), and bimolecular reactions (Krannila et al., 1992) in FCC. Mixtures of model compounds have also been used to deconvolute effective first-order rate constants from competitive adsorption and deactivation (Harding et al., 1993). Although they provide significant insight into the nonlinear kinetic interactions of hydrocarbons, these mixture experiments have the same limitations as the individual model compounds: the smaller size ranges and limited molecular variety are only an idealization of gas oil cracking. Experiments which mix a single model compound into a complete gas oil form a third tier of molecular detail. These experiments are designed either to determine how a specific molecular type affects the global product yields (such as coke production) or to determine the cracking selectivities of specific model compounds in a realistic chemical environment (Cook and Colgrove, 1994). Since the cracking of these model compounds is measured in a large pool of gas oil species and their products, only the most significant product species can be directly measured. Chemical labeling of the model compound before cracking can increase the number of product species that can be experimentally assigned. Although these model compound experiments are per- † W. R. Grace & Co.sConn. ‡ Grace Davison. 2561 Ind. Eng. Chem. Res. 1996, 35, 2561-2569 S0888 5885(95)00449 0 CCC $12 00 © 1996 A i Ch i lS it + +