Membrane-Assisted Reactor Engineering

Like zeolites that combine shape selectivity with catalysis, membranes combine separation with catalysis to enhance reaction rates. The dual functionality of zeolites derives from the nature of the catalytic material, whereas that of membranes derives from the nature of the reactor material. The catalyst in the membrane reactor can be a part of the membrane itself or be external to it (i.e., placed inside the membrane tube). The chief property of a membrane is its ability for selective permeation or permselectivity with respect to certain compounds. Organic membrane reactions are best carried out in reactors made of inorganic membranes, such as from palladium, alumina, or ceramics. Good descriptions of these reactions and the membranes used are available in many reviews, for example, Gryaznov (1986, 1992), Stoukides (1988), Armor (1989), Govind and Ilias (1989), Bhave (1991), Zaspalis and Burggraaf (1991), Hsieh (1989, 1991), Shu et al. (1991), Shieh (1991), Gellings and Bouwmeister (1992), Tsotsis et al. (1993b), Harold et al. (1994), Saracco and Specchia (1994), Sanchez and Tsotsis (1996). A recent trend has been to develop polymeric-inorganic composite type membranes formed by the deposition of a thin dense polymeric film on an inorganic support (Kita et al., 1987; Rezac and Koros, 1994, 1995; Zhu et al., 1996). Another class of membranes under development for organic synthesis is the liquid membrane (Marr and Kopp, 1982; Eyal and Bressler, 1993). The permselective barrier in this type of membrane is a liquid phase, often containing a dissolved “carrier” or “transporter” that selectively reacts with a specific permeate to enhance its transport rate through the membrane. Our main concern in this chapter will be with inorganic membrane reactors. We commence our treatment with an introduction to the exploitable features of membrane reactors (with no attempt to describe membrane synthesis). Then we describe the main variations in design and operating mode of these reactors, develop performance equations for the more important designs, and compare the performances of some important designs with those of the traditional mixed- and plug-flow reactors. Finally, we present a summary of the applications of membrane reactors in enhancing the rates of organic reactions.

Microphase-Assisted Reaction Engineering

A relatively recent concept in organic reaction engineering is the use of submicron particles to enhance the rate of a reaction. These are usually microparticles of solids, but can also be microdroplets of liquids, or even microbubbles of gases. They can be external agents, participating reactants, or precipitating solids. In this chapter, we cover the role of small particles as a whole, which may be regarded as constituting an additional colloid-like phase normally referred to as the microphase. We begin by classifying the microphase in terms of its mode of action and then proceed to an analysis of the following categories of importance in organic technology: microslurry of (1) catalyst or adsorbing particles in a reactive mixture; (2) solid reactant particles in a continuous phase of the second reactant; and (3) solid particles precipitating from reaction between two dissolved reactants, one of which can be a solid dissolving and reacting simultaneously with the other reactant. The microphase in the first case is externally added, whereas that in the last two cases is a reactant or a product. The field is still developing (with many unproven theories), and hence we restrict the treatment to a simple analysis of selected situations based on reasonable assumptions (thus avoiding often unjustified complexity). A microphase can be described as an assemblage of very small dispersed phase particles with average size (dp) much less than the diffusional length scale of the solute. Usually dp < l0μm, compared to the diffusional length scale which is of the order of 50-60 μm. Although the microphase is a distinct phase, the phase in which it is present is commonly regarded as pseudohomogeneous. In a stricter sense, however, it should be regarded as a microheterogeneous phase. Indeed, several studies have been reported on modeling heterogeneous microphase systems (Holstvoogd et al., 1986, 1988; Yagi and Hikita, 1987). In view of the ability of the particles of such a system, pseudohomogeneous or pseudoheterogeneous, to get inside the fluid film, they can enhance the transport rate of the solute through the film. Experimental observations in typical gas-liquid and slurry systems have clearly demonstrated (see Ramachandran and Sharma, 1969; Uchida et al., 1975; Sada et al., 1977a,b, 1980; Alper et al., 1980; Pal et al., 1982; Bruining et al., 1986; Bhaskarwar et al., 1986; Bhagwat et al., 1987; Mehra et al., 1988; Mehra and Sharma, 1988a; Hagenson et al., 1994) the enhancing role of a microphase made up of fine particles. The case of a second liquid phase acting as a microphase or of a solid product performing a similar function has also been studied and found to enhance the reaction rate (Janakiraman and Sharma, 1985; Mehra and Sharma, 1985, 1988b; Anderson et al., 1998). Mehra et al. (1988) and Mehra (1990a,b, 1996) presented a detailed account of the role of different types of microphases in rate enhancement. In all these cases, either a microphase is separately introduced or one of the reactants or products acts as a microphase.

Multiphase Reactions and Reactors

The first three chapters of this part dealt with two-phase reactions. Although catalysts are not generally present in these systems, they can be used in dissolved form in the liquid phase. This, however, does not increase the number of phases. On the other hand, there are innumerable instances of gas-liquid reactions in which the catalyst is present in solid form. A popular example of this is the slurry reactor so extensively employed in reactions such as hydrogenation and oxidation. There are also situations where the solid is a reactant or where a phasetransfer catalyst is immobilized on a solid support that gives rise to a third phase. A broad classification of three-phase reactions and reactors is presented in Table 17.1 (not all of which are considered here). This is not a complete classification, but it includes most of the important (and potentially important) types of reactions and reactors. The thrust of this chapter is on reactions and reactors involving a gas phase, a liquid phase, and a solid phase which can be either a catalyst (but not a phasetransfer catalyst) or a reactant, with greater emphasis on the former. The book by Ramachandran and Chaudhari (1983) on three-phase catalytic reactions is particularly valuable. Other books and reviews include those of Shah (1979), Chaudhari and Ramachandran (1980), Villermaux (1981), Shah et al. (1982), Hofmann (1983), Crine and L’Homme (1983), Doraiswamy and Sharma (1984), Tarmy et al. (1984), Shah and Deckwer (1985), Chaudhari and Shah (1986), Kohler (1986), Chaudhari et al. (1986), Hanika and Stanek (1986), Joshi et al. (1988), Concordia (1990), Mills et al. (1992), Beenackers and Van Swaaij (1993), and Mills and Chaudhari (1997). Doraiswamy and Sharma (1984) also present a discussion of gas-liquid-solid noncatalytic reactions in which the solid is a reactant. In Chapter 7 we saw how Langmuir-Hinshelwood-Hougen-Watson (LHHW) models are normally used to describe the kinetics of gas-solid (catalytic) or liquid-solid (catalytic) reactions, and in Chapters 14 to 16 we saw how mass transfer between gas and liquid phases can significantly alter the rates and regimes of these two-phase reactions.

Reactor Design for Complex Reactions

Procedures were formulated in Chapter 5 for treating complex reactions. We now turn to the design of reactors for such reactions. Continuing with the ethylation reaction, we consider the following reactor types for which design procedures were formulated earlier in Chapter 4 for simple reactions: batch reactors, continuous stirred reactors (or mixed-flow reactors), and plug-flow reactors. However, we use the following less formal nomenclature: A = aniline, B = ethanol, C = monoethyaniline, D = water, E = diethylaniline, F = diethyl ether, and G = ethylene. The four independent reactions then become Using this set of equations as the basis, we now formulate design equations for various reactor types. Detailed expositions of the theory are presented in a number of books, in particular Aris (1965, 1969) and Nauman (1987). Consider a reaction network consisting of N components and M reactions. A set of N ordinary differential equations, one for each component, would be necessary to mathematically describe this system. They may be concisely expressed in the form of Equation 5.5 (Chapter 5), or . . . d(cV)/dt = vrV (11.1) . . . The use of this equation in developing batch reactor equations for a typical complex reaction is illustrated in Example 11.1.

Complex Reactions

When a reactant or a set of reactants undergoes several reactions (at least two) simultaneously, the reaction is said to be a complex reaction. The total conversion of the key reactant, which is used as a measure of reaction in simple reactions, has little meaning in complex reactions, and what is of primary interest is the fraction of reactant converted to the desired product. Thus the more pertinent quantity is product distribution from which the conversion to the desired product can be calculated. This is usually expressed in terms of the yield or selectivity of the reaction with respect to the desired product. From the design point of view, an equally important consideration is the analysis and quantitative treatment of complex reactions, a common example of which is the dehydration of alcohol represented by We call such a set of simultaneous reactions a complex multiple reaction. It is also important to note that many organic syntheses involve a number of steps, each carried out under different conditions (and sometimes in different reactors), leading to what we designate as multistep reactions (normally called a synthetic scheme by organic chemists). This could, for example, be a sequence of reactions like dehydration, oxidation, Diels-Alder, and hydrogenation. This chapter outlines simple procedures for the treatment of complex multiple and multistep reactions and explains the concepts of selectivity and yield. For a more detailed treatment of multiple reactions, the following books may be consulted: Aris (1969) and Nauman (1987). We conclude the chapter by considering a reaction with both catalytic and noncatalytic steps, which also constitutes a kind of complex reaction. Because both chemists and chemical engineers are involved in formulating a practical strategy for accomplishing an organic synthesis, it is important to appreciate the roles of each.

Electroorganic Synthesis Engineering

Historically, electrochemical processes have been limited to the production of inorganic compounds, and commercial processes based on electroorganic synthesis have found only limited application. It appeared to be an “odious truth” (Fry, 1972) that electrochemical techniques were ignored in organic synthesis. But the past 25 years have witnessed the introduction of a fairly large number of new electroorganic processes with attendant advances in electrochemical process analysis. The most remarkable has been Monsanto’s highly successful electrochemical route for the production of adiponitrile. A particularly notable advance is the electrosynthesis of fine chemicals and natural products. Combinations of electrosynthesis with other strategies of rate or selectivity enhancement such as catalysis by PTC and by enzymes (Chapters 19 and 20) are also adding exciting possibilities to organic synthesis. Simultaneously, fundamental understanding of the principles of organic electrochemistry, electrode kinetics, and transport processes in electrochemical systems has grown rapidly in the last decade. A number of books and reviews have appeared on electroorganic chemistry during this period, for example, Eberson and Schafer (1971), Fry (1972), Beck (1974), Perry and Chilton (1976), Rifi and Covitz (1975, 1980), Weinberg (1974, 1990), Swann and Alkire (1980), Kyriacou (1981), Fletcher (1982), Baizer and Lund (1983), Baizer (1973, 1984), Shono (1984), Fletcher and Walsh (1990), Little and Weinberg (1991), Bowden (1997), Bockris (1998), Hamann (1998). This period also saw the emergence of electrochemical reaction engineering as a distinct discipline of chemical reaction engineering, as evidenced by a number of books and reviews on the subject, for example, Picket (1979), Udupa (1979), Danly (1980, 1984), Alkire and Beck (1981), Weinberg et al. (1982), Alkire and Chin (1983), Fahidy (1985), Mine (1985), Goodridge et al. (1986), Rousar et al. (1986), Heitz and Krysa (1986), Ismail (1989), Scott (1991), Prentice (1991), Goodridge and Scott (1995). Electroorganic synthesis offers opportunities for performing many of the conventional organic reactions at controlled rates and greater product selectivities without the addition of any catalyst. The processes almost always employ milder conditions and are characterized by greatly reduced air and water pollution. Further, there are a number of syntheses that can only be carried out electrochemically, such as the Kolbe synthesis and electrochemical perfluorination.

Phase-Transfer Reaction Engineering

There are many situations in organic synthesis where it is desirable to bring about reaction between reactants present in two (or more) immiscible phases. Agents known as phase-transfer catalysts are used for this purpose. Their role in initiating or accelerating such reactions has been proven extensively since the early seventies, and the principles of their operation are being increasingly understood [see Weber and Gokel, 1977; Reuben and Sjoberg, 1981; Frechet, 1984; Freedman, 1986; Goldberg, 1992 (English translation); Dehmlow and Dehmlow, 1993; Starks et al., 1994; Yufit, 1995; Sasson and Neumann, 1997; Naik and Doraiswamy, 1998]. To date, an estimated 500 different commercial chemical processes (mostly for small volume chemicals) using about 5-25 million pounds per annum of phase-transfer catalysts have been reported (Starks et al., 1994), and well over 6,500 compounds have been synthesized in the laboratory using PTC (Keller, 1979, 1986). A large number of industrial applications of phase-transfer catalysis are found in the pharmaceutical, agrochemical, and fine chemicals industries. Additionally, it is now being increasingly used in processes related to the environment, in process modifications for eliminating the use of solvents, and in reactions related to the treatment of poisonous effluents. Not surprisingly, then, there has been a constant stream of publications and patents every year. Phase-transfer catalysis (PTC) is an area that has largely been the province of the preparatory organic chemist (defined broadly to include organometallic and polymer chemists). It is only since the early eighties that the engineering aspects of phase-transfer catalysis are being explored, including such traditional features as mass and heat transfer and reactor design. Our main objective is to present a brief but coherent engineering analysis of PTC, following an introduction to its basic principles. When two reactants are present in two different, immiscible liquid phases (usually one aqueous and the other organic), they can often be brought together by addition of a solvent that is both water-like and organic-like (e.g., ethanol, which derives its hydrophilic nature from its hydroxyl group and its lipophilicity from the ethyl group). However, the rate enhancement tends to be limited due to excessive solvation of the nucleophile.

Gas-Liquid and Liquid-Liquid Reactor Design

The design equation for any reactor expresses its output as a function of the input. For a gas-liquid reactor, the output may be expressed as [Output] = f [input] In Chapter 14, we formulated a number of regimes with corresponding conditions and governing rate equations. In the present chapter, we recast the rate equations in a general form that indicates the relative roles of reaction, liquid film diffusion, and gas film diffusion. Then we briefly discuss the design principles of the more common classes of fluid-fluid reactors. Detailed treatments of design may be found in the books of Astarita (1967), Danckwerts (1970), Shah (1979), Levenspiel (1972, 1993), Doraiswamy and Sharma (1984b), Bisio et al. (1985), and Kastanek et al. (1993).

Reactor Design for Solid-Catalyzed Fluid-Phase Reactions

Catalytic reactions are carried out in reactors with a fixed, fluidized, or moving bed of catalyst. Although the chemical kinetics of the reaction obviously remains the same for all these reactors, the hydrodynamic features vary considerably. Because no complete description of these features is possible, it is convenient to postulate different situations and develop mathematical models to represent these situations for each type of reactor. It is also important to note that wherever solid catalysts are used, the question of catalyst deactivation cannot be ignored. Several books and reviews covering a variety of situations have been written, including those marked with an asterisk in the list of references. They are recommended for general reading. Our intention in this chapter is limited, however: formulate approaches to the design of two main classes of catalytic reactors, fixed and fluidized bed; briefly describe selected procedures along with a few numerical (or methodological) examples to illustrate their use; and outline a procedure for incorporating the effects of catalyst deactivation in reactor design and operation. There are basically two types of fixed-bed reactors: (1) multitubular, in which tubes of approximately 1.5 to 4.0 cm in diameter are placed as a bundle within a shell through which a heat exchange fluid is circulated to control the temperature profile within the reactor; and (2) adiabatic, in which the catalyst is placed directly inside a reactor (with no a priori limitation to the diameter), and heat removal is accomplished by multistaging the bed and removing the heat of reaction by heat exchange between stages. Four major models have been proposed for describing the behavior of a packed tubular reactor (see Doraiswamy and Sharma, 1984). Of these, the most extensively used is the quasi-continuum model in which the fluid-solid system is assumed to act as a single pseudohomogeneous phase with effective properties of its own (as for any true single phase). Thus the procedures developed in Chapters 4 and 10 for the homogeneous model can be used to determine the axial profiles of concentration and temperature. One can also allow for radial transport gradients within each tube [two-dimensional (2-D) models], as opposed to the simpler models in which these gradients are neglected—the one-dimensional (1-D) models.

Estimation of Properties of Organic Compounds

A primary requirement of any reactor design or process development computation is knowledge of the major properties of the compounds involved. Although most of these can be obtained from the literature, there is still a need to estimate them from correlations. The main difficulty is the large number of correlations proposed for a given property and the need to select the best from among them. No single correlation works with equally high precision under all conditions. On the other hand, correlations that can be used with acceptable levels of precision over a wide range of conditions are also available for a number of properties. The slight sacrifice of accuracy is often more than compensated for by the ease and generality of application of these methods. Our emphasis here will be on such correlations. For a detailed treatment, reference should be made to books devoted exclusively to properties estimation. The book by Reid, Prausnitz, and Poling (1987), along with its earlier versions by Reid and Sherwood (1958, 1966) and Reid, Prausnitz, and Sherwood (1977), and the works of Janz (1958), Hansch and Leo (1979), and Lyman, Reehl, and Rosenblatt (1982) are noteworthy. The following methods selected for a few properties are based in part on the recommendations contained in these treatises. The two most important bases for formulating correlations for estimating the properties of organic compounds (indeed of any compound) are the law of corresponding states (LCS), and the method of group contributions (GC). LCS is based on the concept that all substances exhibit identical properties under conditions equally removed from their critical states. The “equally removed” state for any property is usually expressed as the ratio of its value at that state to the value at the critical state and is referred to as the reduced property. Thus Tr — T/TC, Pr = P/PC, Vr = V/Vc and ηr = η/ ηc are the reduced temperature, pressure, volume, and viscosity, respectively. If the simple ideal gas law PV/RgT — Z, where Z is the compressibility, can be recast in terms of reduced properties as PrVt/RgTr = Z/ZC, then the PVT behavior of all fluids can be represented as Pr versus Tr plots for different values of Zr.