Triphase catalysis

Chemical Engineering Science 55 (2000) 5013}5033

Triphase catalysis: a rigorous mechanistic model for nucleophilic substitution reactions based on a modi"ed Langmuir}Hinshelwood/Eley}Rideal approach Justinus A. B. Satrio, Holger J. Glatzer, L. K. Doraiswamy* Department of Chemical Engineering, Iowa State University of Science and Technology, 2114 Sweeney Hall, Ames, IA 50011-2230, USA Received 29 December 1999; accepted 2 May 2000

Abstract In the present work, a general kinetic model based on the traditional kinetic mechanisms of the Langmuir}Hinshelwood and Eley}Rideal types has been developed for nucleophilic substitution reaction systems involving a triphase catalyst in which the intrinsic reaction rate at the catalyst active site is the rate-limiting step. The present mechanistic model overcomes the limitation of the commonly used pseudo-"rst-order model by incorporating the e!ect of the leaving anion on the ion-exchange step, and can be used to determine whether a triphase catalytic system is limited by the organic reaction step, ion-exchange step, or a combination of both the steps. Reactions to synthesize octyl acetate by using di!erent octyl halides as the organic reactant have been used to show how the parameters values obtained can be used to classify systems. These values showed physical relevance since the data were in high conformity with trends that could be expected from the physical and chemical properties of the halide anions and the corresponding alkyl halides. 2000 Published by Elsevier Science Ltd. All rights reserved. Keywords: Phase transfer catalysis; Triphase catalysis; Kinetics; Immobilized phase transfer catalysts; Mathematical modeling; Multiphase reactions

1. Introduction Phase transfer catalysis (PTC), a technique to bring the reactants in two immiscible phases together by adding a phase transfer (PT) catalyst, has been widely used in industry, particularly in organic synthesis, in the last few decades. Compared to the other available techniques, such as dipolar aprotic solvents which are expensive and di$cult to recover, PTC o!ers advantages in terms of reaction conditions and economy. The conventional soluble PTC, however, has one disadvantage, i.e. although various classical chemical separation techniques, such as distillation or extraction, can be used, the separation of catalyst from the product can be complex and may significantly a!ect the cost and purity of the product. A method to overcome it is by immobilizing the PT catalyst on a solid support such as a polymeric resin or an inorganic solid. This method, known as triphase catalysis (TPC) (see Regen, 1975), has huge operational ad-

* Corresponding author. Tel.: 1-515-294-4117; fax: 1-515-294-2689. E-mail address: [email protected] (L. K. Doraiswamy).

vantages over conventional PTC: the catalyst can be separated by a simple "ltration step, thus it can be recycled for further use until it is mechanically degraded. Furthermore, TPC opens up the possibility of conducting PT catalyzed reactions in a continuous system, e.g. a packed bed. Extensive reviews by many researchers in the general "eld of PTC have been published both in the form of monographs and papers in journals, for example Dehmlow and Dehmlow (1993), Starks, Liotta and Halpern (1994); Sasson and Neumann (1997), and Halpern (1997). A more recent review by Naik and Doraiswamy (1998) discusses di!erent aspects of PTC and TPC from both the chemistry and engineering viewpoints.

2. Objective of present study For a triphase catalytic reaction at high agitation rates, it is usually assumed that the intrinsic reaction rate at the active sites is the rate-limiting step since mass transfer and di!usional resistances are much lower. The overall rate of reaction between an organic substrate RX and an inorganic nucleophile Y\ to form organic

0009-2509/00/$ - see front matter 2000 Published by Elsevier Science Ltd. All rights reserved. PII: S 0 0 0 9 - 2 5 0 9 ( 0 0 ) 0 0 1 3 8 - X

5014

J. A. B. Satrio et al. / Chemical Engineering Science 55 (2000) 5013}5033

product RY in the presence of TP catalyst QX then can be expressed as a function of the concentrations of RX and Q>Y\, i.e. Rate"k [RX] [Q>Y\] . (1) Q In many studies it was commonly assumed that the reaction rate of TPC systems is simply a function of the concentration of the organic reactant. For instance, Desikan and Doraiswamy (1995) and Glatzer, Desikan and Doraiswamy (1998) reported that the triphase catalyzed reaction between benzyl chloride and sodium acetate, where the salt concentration is in excess, shows pseudo-"rst-order behavior. This behavior is commonly observed in reaction systems where the inorganic nucleophile is in excess and where the leaving anion from the organic phase does not interfere with the overall reaction. Hence, the concentration of the inorganic nucleophile transferred to the organic phase is constant throughout the reaction. In our recent laboratory experiments, we studied the octyl bromide-potassium acetate reaction system and found that the reaction rate cannot be explained by pseudo-"rst-order kinetics even with a large excess of potassium acetate in the aqueous phase. This is shown by the nonlinear trend in the plots of !log(1!X) versus time in Fig. 1. It is believed that this was caused by the

fact that the concentration of the inorganic anion transferred by the TP catalyst cation was not constant during the course of the reaction. It was concluded that, instead of assuming steady-state equilibrium, there must be a dynamic equilibrium of attachment and detachment of the inorganic nucleophile (acetate anion) to the catalyst cation which is a!ected by the increasing concentration of the leaving anion (bromide) produced by the organic step and transferred to the aqueous phase. The presence of bromide anion in the aqueous phase a!ects the concentration of activated catalytic sites. It is believed that a mechanism used in traditional heterogeneous catalysis, such as the Langmuir}Hinshelwood adsorption mechanism, particularly of the Eley}Rideal type, can explain this phenomenon. The application of a Langmuir}Hinshelwood type mechanism to a triphase catalytic system was attempted by Yadav and Mistry (1995) for the oxidation of benzyl chloride to benzaldehyde with hydrogen peroxide using a system with a capsule membrane PT catalyst. However, the model was reduced to a simple single-parameter (pseudo-"rst-order) expression when used to "t the experimental data; thus the original multiparameter model was not validated. In the present work, a rigorous model based on the modi"cation of the Langmuir}Hinshelwood/Eley}Rideal

Fig. 1. Plots of the octyl bromide-potassium acetate system conversion data "tted to "rst-order kinetics. Reaction conditions: organic/aqueous phase volumes: 0.070/0.070 l; catalyst (polymer supported TBMAC): 0.015 mol/l ; temperature: 953C; agitation speed: 600 rpm. (Legend: octyl bromide _ (mol/l )/ potassium acetate (mol/l ).)


5015

mechanism will be developed, and experimental data from the reaction between octyl bromide in the organic phase and potassium acetate in the aqueous phase to yield octyl acetate will be used to verify it. This reaction system belongs to a general class of esteri"cation reactions where the nucleophile is extracted from the aqueous phase using a phase transfer catalyst. The reaction is mediated by polymer-supported tributylmethylammonium chloride and conducted in the batch slurry mode.

zyl chloride with aqueous sodium acetate in the presence of tributylmethylammonium chloride and reported faster reaction rates with the polymer-supported form than with its soluble analog. More recently, Glatzer and Doraiswamy (2000a) made a comparative assessment of heterogeneous and homogeneous PT catalysts with different categories of PTC systems. It was possible to identify conditions under which the supported catalyst performed distinctly better than its soluble counterpart. A methodology for economic choice of PT catalyst has been suggested by these authors.

3. Literature review

3.2. Kinetic mechanism of TPC systems

3.1. Activity of triphase catalyst

As shown in Fig. 2, a liquid}liquid}solid triphase reaction system consists of an organic liquid phase containing a substrate (typically the dispersed phase), an aqueous liquid phase containing a reagent (typically the continuous phase), and a solid-supported catalyst. Similar to Starks' extraction mechanism for homogenous PTC systems, the mechanism of the phase transfer cycle in a TPC system consists of an ion-exchange step in the aqueous phase followed by the organic-phase reaction step. A major di!erence exists, however. In Starks' mechanism it is assumed that the PT catalyst moves freely

For simple liquid}liquid PTC displacement reactions, such as the reaction between an organic substrate RX and an inorganic reagent (M>YU) in the presence of a PT catalyst: (R.1) RX #(M>Y\) PRY #(M>X\) , Starks et al. (1994) reported that several factors a!ect their reaction rates: (1) structure of R groups, (2) activity of the leaving group X\, (3) nucleophilicity of the displacing group Y\, (4) relative ease of transfer of X\ and Y\ between phases, (5) solvent types, (6) reagent concentrations, (7) agitation intensity, (8) temperature, and (9) catalyst structure. These factors have also been known to a!ect reactions that involve TP catalysts. When all other factors are "xed, theoretically the catalyst's structure becomes the only factor that di!erentiates reactions that use a triphase catalyst from those that use its soluble analog. In triphase catalysis, since the active sites of the catalyst are immobilized on the solid support, the catalyst distribution in the reaction system is more restricted. Reactants from both organic and aqueous phases must migrate from their respective bulk phases to the catalyst surface to contact the catalytic sites. Furthermore, the reactants also must di!use within the solid support in order to contact the sites under the surface. These external and intraparticle mass transfer requirements can signi"cantly a!ect the reaction rate; thus it is commonly believed that triphase catalysts have a lower reactivity compared to their soluble analogs in two-phase reaction systems. However, there are cases reported in the literature where triphase catalysts show higher reactivity. Tundo and Badiali (1989) reported that catalysts made by immobilization of onium salts on inorganic supports (silica and alumina) allow high nucleophilic activity in bromide displacement on octyl methanesulfonate which result in higher reaction rates than for the same reaction in a homogeneous phase. It was believed that the rate increase results from both anion activation and adsorption of the substrate by the inorganic support. Desikan and Doraiswamy (2000) studied the esteri"cation of ben-

Fig. 2. A typical liquid}liquid}solid triphase reaction system.

5016


between the organic and aqueous phases whereas in a TPC system the catalyst movement is restricted and the organic and aqueous reagents must be brought to the catalyst cation in sequence. Fig. 3 shows a schematic representation of the mechanism of a typical TPC system. The immobilization of a PT catalyst to a solid support also introduces characteristics in the reaction system that are typical of heterogeneous catalysis. For instance, instead of considering a planar phase boundary through which the catalyst transports the anions as assumed in classical two-phase systems, one will need to consider a volume element which contains the active catalytic sites as well as the continuous and dispersed bulk phases. For example, as shown in Fig. 2, a polymer-supported TP catalyst consists of an organic phase that contains the hydrophobic polymer backbone solvated by the organic solvent and an aqueous phase that contains water and the inorganic nucleophile (Ohtani, Inoue, Mukudai & Yamashita, 1997). It is likely that the TP catalyst sites involved in the reactions are the ones that are present on the interface between the phases. Given the dynamic nature of the interface in this system, the catalyst sites can alternate between being available and unavailable for the reactions. From studies on ion-exchange resins, it is commonly suggested that the ion-exchange reaction is very fast. This enables one to assume that the ion-exchange reaction is always in equilibrium. Therefore, the overall reaction rate in a L}L}S triphase reaction is determined by the following kinetic steps: (1) mass transfer of reactants, (2) intraparticle di!usion of reactants, and (3) intrinsic organic reaction rate at the active sites. These kinetic steps show similarities of triphase catalysis to traditional heterogeneous catalysis, but with more signi"cant complexity. While traditional heterogeneous catalysis involves di!usion of reactants through a single gaseous or liquid phase into the solid support, triphase catalysis requires

transport of reactants from both liquid phases to the solid surface and di!usion to the active sites. Consequently, the di!usion-reaction scenario is much more complicated. Interaction between the two immiscible phases within the solid support and other in#uencing factors were studied by many researchers. Various mechanisms have been proposed, e.g. Tomoi and Ford (1981), Telford, Schlunt and Chau (1986), Schlundt and Chau (1986), Hradil, Svec, Konak and Jurek (1988), Svec (1988), and Ruckenstein and Hong (1992). Summaries of these studies were reported by Desikan and Doraiswamy (1995) and Naik and Doraiswamy (1998). Another point that adds to the complexity of triphase catalytic systems is the determination of phase continuity within the solid support. The continuities may be determined by the volume ratio of the organic phase to the aqueous phase adsorbed by the solid support whose lipophilicity/hydrophilicity also plays a signi"cant role. Naik and Doraiswamy (1998) suggested that a triphase catalyst supported on a polymer causes an inverse continuity. The lipophilic polymer support imbibes the organic solvent, thus making the organic phase the continuous phase with the dispersed aqueous phase droplets being transported through the organic phase to come in contact with the immobilized catalytic sites. Other factors, such as the type of solvent, reactant concentration, and the type of the inorganic anion, organic leaving group and catalyst cation, have been reported to a!ect the organic and aqueous phase distribution (Ohtani et al., 1997). The complex scenario behind the phase distribution within the catalyst support seems to be the main di$culty in developing a general kinetic model for triphase catalytic systems. Estimation of the distribution of organic and aqueous phases within the solid support will require a knowledge of the contributions of many other factors, such as solvent type, reactant concentration, and type of support. Modeling the phase distribution, by itself, is a very challenging task if not an impossible one. To date, no kinetic models on triphase catalysis which consider the complicated dynamics of phase distribution within the catalyst particle have been reported in the literature. 3.3. Kinetic modeling of TPC systems

Fig. 3. A schematic diagram of reaction mechanism for TPC system.

So far, only a handful of papers have reported detailed kinetic and modeling studies on TPC. One of the "rst models was developed by Marconi and Ford (1983). This model, which was based on standard equations developed for porous catalysts, uses the e!ectiveness factor to describe the e!ects of mass transfer resistances outside and within the supported catalyst particles. It assumes a pseudo-"rst-order reaction by considering explicitly the organic phase and completely neglecting the ionexchange kinetics and all transfer resistances to the


5017

aqueous phase; thus, it will not hold for chemical systems where the ion-exchange reaction cannot be ignored. This happens when the inorganic anion is more di$cult to transfer than the leaving anion from the organic phase. A model which considers mass transfer of reactants in the bulk aqueous and organic phases, di!usion of reactants within the catalyst particle, and intrinsic reactivities of the ion-exchange and organic reactions, has been proposed by Wang and Yang (1991). This model assumes a steady-state condition for the mass balance equations. A dynamic model was developed by the same authors (Wang & Yang, 1992) which was further modi"ed by Desikan and Doraiswamy (1995, 1999) to account for the reversibility of the ion-exchange reaction and for nonisothermality. All these models assume that the reactions are elementary and independent in each phase.

fer involves the di!usion of OAc\ in the aqueous phase in the bulk and within the catalyst particle. Step 2: Ion-exchange reaction between the nucleophile and the leaving anion (bromide anion, Br\) attached to the catalyst cation Q> to form active sites Q>OAc\ Step 3: Transfer of the organic substrate (octyl bromide, RBr) from the bulk phase into the catalyst particle. The transfer involves the di!usion of substrate in the organic phase in the bulk and within the catalyst particle. Step 4: Reaction between the substrate and the nucleophile at the active sites of the catalyst located at the interface to form the organic product (octyl acetate, ROAc) and to reform the attached leaving anion on the catalyst cation. Step 5: Transfer of the leaving anion and the organic product from the catalyst particle to the bulk aqueous phase and organic phase, respectively.

4. A plausible theoretical model

At reaction conditions where the rates of mass transfer are much higher than the rate of reaction, the mass transfer steps can be ignored. Thus, we assume that the reaction mechanism consists of an ion-exchange reaction step between OAc\ and Q>X\ (for the "rst ion-exchange cycle X\"Cl\ and for all subsequent cycles X\"Br\) to form an active site, Q>OAc\, followed by reaction of RBr at this site to form the "nal product, ROAc, and an inactive site, Q>Br\. These steps may be described as follows: Ion-exchange step:

4.1. Derivation of the model The present model is developed to explain the kinetics of nucleophilic substitution reaction systems involving a triphase catalyst in which the intrinsic reaction rate at the catalyst active site is the rate-limiting step. The following assumptions are made: 1. Reactions take place at the catalyst cations located at the interface in the catalyst particle. 2. Phase distribution within the catalyst is constant and not a!ected by phase composition changes. 3. Di!usion coe$cients of reactants are constant. 4. There is no change in phase volumes during reaction. 5. Organic reactant and product are insoluble in the aqueous phase. 6. Organic and aqueous phase bulks are well mixed. 7. Isothermal conditions prevail throughout the course of reaction. 8. Extraction at the interface is in equilibrium. Let us assume a triphase catalytic reaction where octyl bromide (RBr) dissolved in toluene reacts with potassium acetate (K>OAc\) dissolved in the aqueous phase to yield octyl acetate (ROAc) and potassium bromide (K>Br\) in their respective phases. The TP catalyst is denoted as (Q>Cl\). The overall reaction can be expressed as (R.2) RBr

#(K>OAc\) PROAc #(K>Br\) . The steps that are thought to be involved in the reactions include the following mass transfer and the surface reaction steps: Step 1: Transfer of the nucleophile (acetate ion, OAc\) from the bulk phase into TP catalyst particle. The trans-

(R.3) (Q>Br\) #(OAc\) (Q>OAc\) #(Br\) . Q Q Organic phase reaction step: (R.4) (Q>OAc\) #(RBr) (Q>Br\) #(ROAc) . Q Q The overall reaction can be seen as analogous to an Eley}Rideal reaction mechanism that involves reaction between an adsorbed reactant with an ùn-adsorbeda reactant from the bulk phase. We treat the ion-exchange step as the adsorption of the "rst reactant on the inactive catalyst sites to form active sites, and the organic phase reaction step as the reaction of the second reactant with the adsorbed reactant at the catalyst sites. Note that the de"nition of ìnactivea and àctivea sites for a TP catalytic system is di!erent from that for traditional heterogeneous catalysis. In the latter an active site is a catalyst site at which the adsorption and reaction steps take place. Reaction does not take place at an inactive site. An active site can be either vacant or occupied depending on whether it contains an adsorbed atom/ molecule/complex or not. In TPC, a site is meant to be the catalyst cation, i.e. Q>. It is considered inactive when it is attached to the catalyst's original anion or to the by-product anion (in our case Cl\ or Br\). On the other hand, a site is active if it is attached to the inorganic nucleophile, i.e. OAc\ anion. The organic phase reaction occurs at this activated site.

5018


The reversible ion-exchange reaction step may be compared to the Langmuir}Hinshelwood adsorption/desorption mechanism. Using the traditional notation of heterogeneous catalysis, we can express the reversible ion-exchange reaction (R.3) as (R.5) OAc\#S> ) Br\ Br\#S> ) OAc\, where S> is the triphase catalyst's cation. Assuming the formation of a transitional site Br\ ) S> ) OAc\ between the forward and reverse reaction steps, i.e.

Br\#S> ) OAc\, we can split the steps as two separate equilibrium attachment/detachment steps. The forward reaction step can be seen as the àttachment/detachmenta of OAc\ anion on the inactive site, S>Br\, i.e. (R.7) OAc\#S> ) Br\Br\ ) S> ) OAc\. Similarly, the reverse reaction step can be seen as the àttachment/detachmenta of Br\ anion on an active site S> ) OAc\, i.e. (R.8) Br\#S> ) OAc\ Br\ ) S> ) OAc\. Assuming the rates of attachment and detachment to be in equilibrium, we obtain Eqs. (2) and (3) for reactions (R.7) and (R.8), respectively:

h

-

"K [OAc\] (1!h !h ), - - -

(2)

-

"K [Br\] (1!h !h ), -

(3)

where K and K are the equilibrium attachment/detachment constants for OAc\ and Br\ anions, respectively; [OAc\] and [Br\] are the concentra tions in the aqueous phase for OAc\ and Br\ anion, respectively; and h , h , and h are the fractions - - of the total numbers of TPC cations attached to OAc\, Br\ and both OAc\ and Br\ anions, respectively. It is postulated that transition sites Br\ ) S> ) OAc\, once formed, are transformed instantaneously to either active sites S> ) OAc\ or inactive sites S> ) Br\. Thus, Eqs. (2) and (3) can be written as h "K [OAc\](1!h !h ), - - -

(4)

h "K [Br\](1!h !h ). -

(5)

Combining the expressions for h and h , we obtain - a hyperbolic equation for the fraction of active TPC sites as K [OAc\] - h " - 1#K [OAc\] #K [Br\] -

[S> ) OAc\] K [OAc\] - "[S>] , (7) 1#K [OAc\] #K [Br\] - where [S>] and [S> ) OAc\] are the total concentra tion of catalyst and the concentration of catalyst attached to OAc\ anions, respectively. Finally, combining Eqs. (1) and (7), we obtain the following expression for the rate of the organic reaction: "k [RBr] [S>] K [OAc\] - ; . (8) 1#K [OAc\] #K [Br\] - The model assumes that the conversion rate is a linear function of the total concentration of catalyst. To account for a possible non-linearity between catalyst concentration and the conversion rate, the model can be modi"ed as !r

(R.6) OAc\#S> ) Br\ Br\ ) S> ) OAc

h

or in term of catalyst concentration:

(6)

"k [RBr] [S>]? K [OAc\] - , (9) ; 1#K [OAc\] #K [Br\] - where a is the power-law exponent on the concentration of catalyst. !r

4.2. Physical interpretation of the model The rate expression derived above shows that the organic phase reaction rate is a function of the concentration of octyl bromide in the organic phase, the concentrations of acetate and bromide anions in the aqueous phase, and the amount of catalyst. While octyl bromide and the acetate anion have positive e!ects on reaction rates, the bromide anion's e!ect is negative. Once the reaction starts, the leaving bromide anions from the organic phase reaction compete with the acetate anions in the aqueous phase for association with the catalyst cations. The values of the equilibrium attachment/detachment constants of acetate and bromide anions, K and K respectively, indicate the degrees of the - anion's a$nity to the catalyst. The ratio of K to - K indicates the `degree of competitivenessa between the acetate and bromide anions for coordination with the catalyst cations. It is interesting to note that this ratio can be seen as the equilibrium selectivity coe$cient for the ion-exchange step de"ned as (Hel!erich, 1962) K [Br\]\[S> ) OAc\]\ K " - " , (10) 1 K [OAc\]\[S> ) Br\]\ where K is the equilibrium constant for the forward - reaction step and K the equilibrium constant for the


reverse reaction step. More favorable reaction conditions are obtained with higher value of K . This can be 1 achieved with higher values of the forward reaction step constant and lower values of the reverse reaction step constant. Higher K than K would indicate that - acetate anions are more likely to be coordinated with the catalyst cations than bromide anions, thus more acetate is available at the catalytic sites to undergo the organic phase reaction with octyl bromide. From this model, in general, we can see that a fast triphase reaction system will have a fast organic phase reaction step, indicated by a high value of k and a large anion transfer rate indicated by a large ratio of K to K . 7 6 In the case of reaction between octyl bromide and potassium acetate, this is represented by the ratio of K to - K . These parameter values may be used as a guide in classifying TPC reaction rates. Based on these values, a triphase reaction regime diagram similar to the PTC reaction regime diagram developed by Starks et al. (1994) can be prepared. Such a diagram is shown in Fig. 4. According to this diagram, the overall conversion rate of a triphase catalytic system is limited by the ionexchange step when the organic reaction rate is fast and its leaving anion is strongly coordinated with the catalyst cation compared to the inorganic nucleophile, as indicated by a low value of the ratio of K /K . As the 7 6 leaving anions start to form and are transferred to the aqueous phase they will increasingly occupy the catalytic sites due to their high degree of coordination until an equilibrium condition is reached. As a consequence, the value of the hyperbolic term in the model quickly reduces to a low value which will not further change as the reaction progresses. Once this condition is reached the overall reaction will follow pseudo-"rst-order kinetics. In the opposite extreme, a triphase catalytic system is limited by the organic reaction step. Hence, the organic reaction rate is slow and K /K is relatively large which 7 6 indicates that the inorganic nucleophile can easily re-

5019

place the leaving anion that is in coordination with the catalyst cation. The slow organic reaction rate produces a low amount of leaving anions. Coupled with its relatively low a$nity to the catalyst cation, the leaving anion will only slightly interfere with the coordination of the inorganic nucleophile with the catalyst and consequently the value of the hyperbolic term in the model will only change marginally. When the salt concentration is in excess the value of this term can be approximated to unity. Thus, generally, the overall reaction rate of triphase catalytic systems that fall in this category can be explained by pseudo-"rst-order reaction kinetics. In the diagram, several triphase catalytic systems fall in the category called the transitional region. In this category, the importance of both organic reaction and ionexchange steps are relatively equal toward the overall conversion rates. The formation of the leaving anions from the organic reaction is moderately fast. These leaving anions will interfere with the ion-exchange step although the interference would not be severe due to the relatively weak a$nity of the leaving anion to the catalyst cation. In the model, the interference of the leaving anions will cause the value of the hyperbolic term to decrease throughout the course of the reaction. Triphase catalytic systems that fall in the transitional region can be detected by the fact that they do not show pseudo-"rstorder behavior even when the systems have large excess salt concentrations. As a concluding remark, we can say that the behavior of the hyperbolic term in the model indicates the category to which a triphase catalytic system belongs. However, it should be noted that the initial reaction rate constant can still be obtained by assuming a pseudo-"rst-order model regardless of the category of the system. At initial reaction time, since the leaving anion has yet to form, the hyperbolic term can be assumed to be constant. The value can be approximated to unity if the salt concentration is in excess.

5. Experimental procedures

Fig. 4. Reaction regime diagram for triphase catalysis.

To check the validity of the model, octyl acetate concentration vs. time data for the reaction between octyl bromide and potassium acetate in the presence of a triphase catalyst were obtained. Unless stated otherwise, all reagents were of analytical grade and not further puri"ed. Octyl bromide was dissolved in toluene and potassium acetate in deionized water. The triphase catalyst used in this study was tributylmethylammonium chloride (CH N[(CH ) CH ] Cl) bound on polysty rene. For catalytic performance study, the triphase catalyst was compared with benzyltributylammonium chloride (C H CH N[(CH ) CH ] Cl). This particular salt was selected as the appropriate soluble analog since the immobilization of PT catalyst on polystyrene adds

5020


a benzene ring to the functional group which makes the functional group of the polystyrene bound catalyst to be best represented by benzyltributylammonium chloride (Desikan & Doraiswamy, 2000). Both catalysts were obtained from Fluka Chemical Corp. One gram of the heterogeneous catalyst is equivalent to 0.85 mmol (corresponds to 0.277 g of homogeneous catalyst). All the kinetic experiments were carried out in a 300 ml agitated vessel from Parr Instruments which was lined with a Te#on insert and equipped with a two-blade paddle. The temperature could be controlled to within 13C. For all reactions, the total volume of the mixture was 140 ml with equal volume fractions for the liquid phases. In a typical experiment, potassium acetate was added to deionized water to make 70 ml potassium acetate solution having a speci"ed molar concentration. The solution was transferred into the reactor vessel. Predetermined quantities of toluene and phase transfer catalyst were then fed into the reactor. The solution was brought to the required temperature at slow agitation. As soon as the target temperature was reached, octyl bromide was added (reaction time"zero) and the mixture subsequently agitated at a rate of 600 rpm. This agitation rate was high enough to remove any e!ect of external mass transfer resistance. Samples of about 1.5 ml were drawn from the reactor periodically. The sampled organic phase was "ltered using 0.15 lm Te#on syringe "lters. The concentrations of octyl bromide and octyl acetate were measured by using a Perkin-Elmer gas chromatograph (Model 3000 Autosystem with FID) with

external standards. A packed column (Carbopack, 10% SP-2250 from Supelco Inc.) with a length of 2.0 m and a diameter of 1/8 inch was used for the analysis. The analysis was run for 9 minutes at 1503C. Reaction rate data were obtained by measuring the slopes of octyl acetate concentration vs. time plots using POLYMATH威 software.

6. Results and discussion 6.1. Ewect of reaction variables 6.1.1. Salt concentration In this study, the catalytic performance of polymer-supported tributylmethyl-ammonium chloride (TBMAC) was compared to that of its soluble analog, benzyltributylammonium chloride (BTBAC). The concentration of octyl bromide in the organic phase was held constant at 1.0 mol/l , while that of potassium acetate in the aqueous phase was varied from 1.0 to 10.0 mol/l . The molar amounts of triphase catalyst and soluble catalyst used were 2.1 mmol which corresponds to a catalyst concentration of 0.015 mol/l . Fig. 5 shows plots of octyl bromide conversions for reactions in the presence of soluble catalyst with salt concentrations ranging from 1 to 10 mol/l at 953C. The conversions obtained without catalyst at a salt concentration of 10 mol/l are also shown. In the absence of catalyst, there was practically no conversion even after

Fig. 5. Actual conversion plots for the octyl bromide-potassium acetate system in the presence of soluble BTBAC catalyst at various combinations of octyl bromide and potassium acetate concentrations. Reaction conditions: organic/aqueous phase volumes: 0.070/0.070 l; catalyst: 0.015 mol/l ; temperature: 953C; agitation speed: 600 rpm. (Legend: octyl bromide (mol/l )/potassium acetate (mol/l ).)


11 h. For the catalytic reactions, the reaction rate was strongly a!ected by the salt concentration. It was high when a large excess of potassium acetate is used. Thus almost 80% conversion was reached within 2 h of reaction time with a salt concentration of 10 mol/l . On the other hand, it took more than 11 hours to reach 70% conversion when the salt concentration was lowered to 5 mol/l . For 2 mol/l this value reduced to below 10% and for 1 mol/l to below 5%. As a concluding remark, we can state that the e!ect of salt concentration in the aqueous bulk phase is much greater than a possible "rst-order dependence would indicate. Fig. 6 shows analogous plots for the supported catalyst (TBMAC). We can see from these plots that the salt concentration has a less signi"cant e!ect on conversion when a heterogeneous catalyst is used to promote the reaction. In the low salt concentration range (1}3 mol/l ), conversions signi"cantly increase with in crease in concentration. However, as the concentration is further increased, the incremental e!ect becomes progressively small until it becomes almost negligible. Thus, with the polymer-supported catalyst there seems to be a limit to the e!ect of increasing salt concentration on conversion. From Figs. 5 and 6 we can also conclude that reactions with soluble BTBAC catalyst are signi"cantly faster than those with polymer supported TBMAC catalyst at high salt concentrations. However, at low salt concentrations (1 and 2 mol/l ), the supported catalyst performs better

5021

than the soluble catalyst. This "nding con"rms that at certain reaction conditions the polymer-supported catalyst can be a better choice than its soluble analog. The e!ect of salt concentration on the catalytic activity of supported PT catalyst and its soluble analog may be explained by variations in the rate of the ion-exchange reaction and in the distribution of PTC cations between the organic and aqueous phases. The "rst e!ect is well known, i.e. generally increasing the concentration of salt (i.e. inorganic nucleophile) will increase the rate of ion exchange, resulting in a higher conversion. This e!ect should be valid for both the systems, soluble and supported PT catalysts. The e!ect of salt concentration on the PT catalyst cation distribution is di!erent for soluble and supported PT catalysts. For systems with soluble PT catalyst, the ratio of the catalyst concentration in the organic phase to that in the aqueous phase is de"ned as the stoichiometric extraction constant E (Dehmlow & Dehmlow, 1993) /7 [Q>Y\] . E " /7 [Q>] [Y\]

(11)

From Eq. (11), we see that an increasing salt concentration will cause larger amounts of (Q>Y\) to be extracted to the organic phase. With larger amounts of (Q>Y\) available in the organic phase, the rate of the organic phase reaction should increase also. Thus, increasing the salt concentration does not only increase the availability

Fig. 6. Actual conversion plots for the octyl bromide-potassium acetate system in the presence of polymer-supported TBMAC catalyst at various combinations of octyl bromide and potassium acetate concentrations. Reaction conditions: organic/aqueous phase volumes: 0.070/0.070 l; catalyst: 0.015 mol/l ; temperature: 953C; agitation speed: 600 rpm. (Legend: octyl bromide (mol/l )/potassium acetate (mol/l ).)

5022


of the inorganic nucleophile but also physically enhances the ability of the catalyst to reside in the organic phase. This e!ect should be more pronounced in reactions using a hydrophilic catalyst which mostly resides in the aqueous phase, less pronounced in reactions that use a lipophilic catalyst, and absent in reactions that use a catalyst which is completely soluble in the organic phase. Benzyltributylammonium chloride is relatively hydrophilic, which explains the strong e!ect of salt concentration on conversion for the current reaction system. In the case of a polymer-supported catalyst, the situation is di!erent. The catalytic sites are immobilized and a$nities to either phase strongly depend on the type of catalyst support. Therefore, a change of salt concentration in the aqueous phase should not signi"cantly a!ect the distribution of catalytic sites between the phases. It is expected that the salt concentration should in#uence only the number of catalyst cations that become active sites. At low salt concentrations the number of active sites will be proportional to the bulk concentration. Thus, increasing the initial salt concentration will increase the number of catalyst active sites resulting in higher reaction rates. At higher salt concentrations, however, an increase will no longer a!ect the number of available active sites, since its maximum number is determined by the equilibrium selectivity coe$cient K . 1 6.1.2. Concentrations of organic and aqueous phase reactants To study the e!ects of the concentrations of organic and aqueous phase reactants, rate data were obtained by varying the concentrations of octyl bromide and potassium acetate. The concentration of octyl bromide in the organic phase ranged from 0.25 to 4.0 mol/l and of potassium acetate in the aqueous phase from 0.25 to 10 mol/l . These e!ects are shown in Figs. 7a and b. Fig. 7a shows the reaction-rate plots at a potassium acetate concentration of 2.0 mol/l and an octyl bromide concentration which varied from 0.25 to 4.0 mol/l . It can be seen that increasing the initial octyl bromide concentration causes a large increase in the initial reaction rate which drops quickly during the initial stages of the reactions. The large decrease in the rate, especially at high octyl bromide concentrations, can be explained by the formation of bromide anions which are transferred back into the aqueous phase. The high initial rates cause a fast accumulation of bromide anions in the aqueous phase which compete with the acetate anions for coordination with the active sites. Since bromide has a higher a$nity to the catalyst cation than acetate, the number of active sites (i.e. catalyst cations attached to acetate anions) decreases in the presence of bromide anions and consequently the reaction rates are lowered. The response of the system is di!erent when the initial concentration of potassium acetate varies as shown in

Fig. 7b. It can be seen that in the low salt concentration range (0.5}4.0 mol/l ) an increasing initial salt concen tration has a positive e!ect on the reaction rate, but the decrease in rate over time is much lower than that was observed in reactions in which the initial concentration of octyl bromide was varied. At higher salt concentration ranges, no signi"cant di!erences in the rates were observed. Under these conditions the number of active sites is at the optimum and thus further increase in salt concentration cannot be utilized to enhance the rate. 6.2. Estimation of model parameters To obtain the reaction rate parameters, k , K and - K , data from a modi"ed 3;3 factorial experimental design for three di!erent temperatures, i.e. 65, 80, and 953C, were obtained. The two variables were the concentration of octyl bromide in the organic phase, at levels 0.25, 0.5, and 1.0 mol/l , and that of potassium acetate in the aqueous phase, at levels 0.25, 0.5, and 1.0 mol/l . The maximum excess of organic reactant over inorganic reactant was allowed to be two-fold. Therefore, the combination 1.0/2.0 was substituted for 1.0/0.25 (octyl bromide/potassium acetate). The concentration of catalyst was held constant at 0.015 mol/l . Ten samples were drawn from the reactor during the course of the reaction (10 h). For each data point, the concentrations of octyl bromide, acetate, and bromide, as well as the rates of reaction were obtained. The data were "tted to the hyperbolic model (Eq. (8)) by using a non-linear least-squares method (Marquardt method). SAS statistical software was used for this purpose. The calculation of model parameters was done in two steps. In the "rst step, the data were "tted to estimate all three parameters. To check for consistency, the parameters were estimated using all 3;3 data levels and 1;3 data for each level of the concentration of octyl bromide. The results for all three temperatures are shown in Table 1. We can see from the data that the equilibrium constant for acetate, K , is almost independent of temperature - with a value close to 0.001. Based on this a reduced parameter model was "tted to the data holding K constant at 0.001. The results, shown in Table 1, - give more consistent values for k and K . The "nal values of the parameters are those obtained from the reduced model. These values were used to calculate the conversion rates. The predicted and experimental values are shown in Figs. 8}10. For all three temperatures, the model is capable of predicting the actual rates well. It is expected that the organic reaction rate constant, k , and the ion-exchange equilibrium constant, K , follow Arrhenius relations. From the k values of three temperature levels, the activation energy of the organic reaction was found to be 76.8 kJ/mol. For the


5023

Fig. 7. Actual reaction rate plots for the octyl bromide-potassium acetate system in the presence of polymer-supported TBMAC catalyst at various combinations of octyl bromide and potassium acetate concentrations. (a) Variation of octyl bromide concentration, (b) Variation of potassium acetate concentration. Reaction conditions: organic/ aqueous phase volumes: 0.070/0.070 l; catalyst: 0.015 mol/l ; temperature: 953C; agitation speed: 600 rpm. (Legend: octyl bromide (mol/l )/potassium acetate (mol/l ).)

same reaction system, Glatzer and Doraiswamy (2000b) reported a value of 74.4 kJ/mol which was calculated from initial reaction rates.

Assuming an ideal condition at which the thermodynamic complexity is greatly simpli"ed, we treat the equilibrium selectivity coe$cient of the ion exchange, K , to be 1

5024


Table 1 Parameter values of the model for the octyl bromide-potassium acetate system Temp (3C)

95

Temp (3C)

3-Parameter model Data based on OBr conc. levels

k ;10 (l /min mol cat.)

K ;10 (mol/l )\ -

K ;10 (mol/l )\

All 1.0 mol/l 0.5 mol/l 0.25 mol/l

2.36$0.21 2.65$0.45 3.49$1.41 2.65$0.34

1.02$0.27 0.85$0.27 0.74$0.51 1.26$0.40

1.82$0.35 1.49$0.48 1.87$0.68 1.93$0.31


2.39$0.09 2.33$0.14 2.61$0.21 3.02$1.32

1.00 1.00 1.00 1.00

1.79$0.20 1.75$0.30 2.20$0.55 1.73$0.31

Final values

2.39$0.09

1.00

1.79$0.20


K ;10 (mol/l )\ -

K ;10 (mol/l )\

1.23$0.20 1.37$0.46 0.69$0.13 0.72$0.12

0.74$0.18 0.53$0.30 1.68$1.11 1.82$0.66

2.12$0.37 1.96$0.67 2.75$0.90 3.29$1.15


0.98$0.04 1.00$0.08 0.90$0.05 0.98$0.06

1.00 1.00 1.00 1.00

2.65$0.36 2.75$0.66 2.12$0.58 2.49$0.78

Final values

0.98$0.04

1.00

2.65$0.36


K ;10 (mol/l )\ -

K ;10 (mol/l )\

0.25$0.02 0.23$0.02 0.20$0.03 0.27$0.04

1.09$0.18 1.24!0.29 1.19$0.27 1.03$0.23

2.52$0.36 3.00$0.62 1.76$0.56 2.95$0.84


0.26$0.01 0.26$0.01 0.22$0.01 0.28$0.01

1.00 1.00 1.00 1.00

2.41$0.29 2.42$0.42 1.66$0.51 2.91$0.77

Final values

0.26$0.01

1.00

2.41$0.29

2-parameter model

3-parameter model Data based on OBr conc. levels All 1.0 mol/l 0.5 mol/l 0.25 mol/l

80

Temp (3C)

2-parameter model

3-parameter model Data based on OBr conc. levels All 1.0 mol/l 0.5 mol/l 0.25 mol/l

65

2-parameter model

Note: limits of parameter values are based on 95% con"dence interval.

equal to the thermodynamic equilibrium constant K . Thus, the values of K /K which are the same as - K give the values of K of the ion exchange. From the 1 K values, the enthalpy change of the ion-exchange is approximated to be 10 kJ/mol. The value is quite close to the standard enthalpy changes accompanying ion exchange which is usually less than 8.4 kJ/mol (Hel!erich, 1962).

To check the relation between PT catalyst concentrations with the conversion rates, the value of a was estimated by "tting the hyperbolic model (Eq. (9)) with the calculated values of k , K , and K to an - experimental data set from reactions by varying the amount of catalyst at octyl bromide/potassium acetate concentrations of 0.5/1.0 mol/l at 953C. Three levels of


5025

Fig. 8. Predicted and actual reaction rate plots for the octyl bromide-potassium acetate system in the presence of polymer-supported TBMAC catalyst at 953C. (a) 0.25 and 0.5 mol/l octyl bromide concentration levels, (b) 1.0 mol/l octyl bromide concentration level. Reaction conditions: organic/ aqueous phase volumes: 0.07/0.07 l; catalyst: 0.015 mol/l ; agitation speed: 600 rpm. (Legend: octyl bromide (mol/l )/potassium acetate (mol/l ).)

catalyst concentration were used: 0.00375, 0.0075, and 0.015 mol/l . The data were "tted to the model to obtain the value of a. The value of a was found to be 0.9601 which is very close to unity. Percentage of the data

explained by the model is 98.7%. Hence, the conversion rate is a linear function of the catalyst amount. Plots of experimental rate data and predicted rates calculated by using an a value of 1 are shown in Fig. 11. The

5026



plots show that the model is able to "t the experimental data well. 6.3. Limitations of the model To further test the model, experimental data from reactions with octyl bromide concentration of

1.0 mol/l and potassium acetate concentration ranging from 3.0 to 10.0 moles/l with a catalyst concentration of 0.015 mol/l at 953C were "tted to the model. The values of k , K , K that were obtained from previous - simulations were tested for the new conditions. The plots of the predicted and experimental reaction rates are shown in Fig. 12. It can be seen that the "t of the model


5027


becomes worse as the initial salt concentration increases. At very high salt concentrations (8}10 mol/l ), the model signi"cantly overpredicts the reaction rates. One possible explanation for this observation is the dissociation of potassium acetate in the aqueous phase. At low salt concentrations, potassium acetate will mostly be in its dissociated form. It can be assumed that the concentration of acetate anions in the aqueous phase is

the same as the concentration of the salt. However, at high salt concentrations, this is no longer the case. The concentration of acetate anions will be much lower than that of the salt. The acetate concentration data were obtained based on the potassium acetate concentration. An improvement of the data is possible by incorporating the dissociation constant of potassium acetate in the model and thus obtaining a more realistic

5028


Fig. 11. Predicted and actual reaction rate plots for the octyl bromide-potassium acetate system at various amounts of polymer-supported TBMAC catalyst. Reaction conditions: octyl bromide/potassium acetate concentrations: 0.5/1.0 mol/l; organic/aqueous phase volumes: 0.070/0.070 l; temperature: 953C; agitation speed: 600 rpm.

Fig. 12. Predicted and actual reaction rate plots for the octyl bromide-potassium acetate system with high salt concentrations in the presence of polymer- supported TBMAC catalyst. Reaction conditions: organic/ aqueous phase volumes: 0.070/0.070 l; catalyst: 0.015 mol/l ; temperature: 953C; agitation speed: 600 rpm. (Legend: octyl bromide (mol/l )/potassium acetate (mol/l ).)


estimate of the concentration of acetate anions in the aqueous phase. 6.4. Model conxrmation and reaction system classixcation: octyl acetate synthesis using other octyl halides as reactants To con"rm the validity of the model for other chemical systems of the same category, two other octyl halides, namely chloride and iodide, were used to synthesize octyl acetate by reacting them with potassium acetate. These two additional systems were chosen to show how the di!erent halides a!ect the values of the model parameters. Among the three halides, iodide is the most lipophilic. It will form a strong coordination with the catalyst anion which is di$cult to be replaced by other anions, thus signi"cantly hindering the ion-exchange step. For this reason iodide is considered to be a catalyst poison in typical PTC applications. On the other hand, alkyl iodides are the most reactive which will result in high organic reaction rates. The other extreme is the chloride anion. Chloride is the hardest anion and the least lipophilic which makes its coordination with the catalyst anion easy to be replaced by another anion. For this reason chloride is the best leaving anion which is favorable for the ion-exchange step. On the other hand, alkyl chlorides are less reactive than the other alkyl halides, thus the rate of the organic reaction with alkyl chlorides is expected to be much lower. We see that the e!ects are opposite and as we move from iodide to bromide to chloride it is likely that the conversions obtained with bromide will be the highest. Based on this knowledge, we can expect the following orders for k and K /K : 7 6 k : octyl iodide'octyl bromide'octyl chloride, K /K : octyl iodide(octyl bromide(octyl chloride. 7 6 From this qualitative comparison, it is expected that octyl acetate synthesis from octyl iodide and octyl chloride will be limited by the ion-exchange step and the organic phase reaction step, respectively. The synthesis with octyl bromide represents a combination of the two steps, thus lying in the transition region indicated in Fig. 4. To con"rm the above analysis, experimental data were obtained for each reaction system at 953C. The octyl halide conversion and the reaction-rate plots are shown in Figs. 13a and b, respectively, for all three reaction systems at the concentration levels of 1.0/2.0 mol/l (octyl halide/potassium acetate). In Fig. 13a it can be seen that the conversion of octyl iodide is initially the highest among the three halides, but it levels o! in less than 30 min after the reaction is started. This is shown even more clearly by the reaction-rate plots in Fig. 13b. Data analysis of the system showed that once the reaction rates level o! (i.e. after 30 min), the reaction shows pseudo-

5029

"rst-order behavior which indicates a constant number of catalytic active sites. The reaction rates of octyl chloride are the lowest. Data analysis of the system showed pseudo-"rst-order behavior that again suggests a constant number of catalytic active sites during the entire reaction. With octyl bromide the reaction rates are always higher than with octyl chloride. Compared to those with octyl iodide, the reaction rates achieved with the bromide are initially slower, but as time progresses the values for octyl bromide become signi"cantly higher. The "nding that octyl bromide conversion does not follow pseudo-"rst-order behavior suggests that both reaction steps (organic and ion-exchange) control the overall conversion. The above observations were con"rmed by the model. Data for octyl iodide and octyl chloride systems from two experimental runs for each system were "tted to the model. It was assumed that the attachment equilibrium constant for acetate anion, K , is independent of the - type of leaving anion, thus the value of K "0.001 - obtained in the previous simulations with the octyl bromide system was used for these two systems. The resulting values of the parameters were used to calculate the reaction rates. The predicted and experimental rates for the octyl iodide and octyl chloride are shown in Figs. 14 and 15, respectively. Despite the limitations of data, we can see that the model is able to "t the data well for both the systems. Table 2 shows the values of the parameters for all the three reaction systems at 953C. The organic reaction rate constant for octyl iodide system is the highest which explains its highest reactivity at the beginning of the reaction. The equilibrium attachment constant for the leaving anion, iodide, is also the highest which indicates that, among the inorganic anions studied, iodide has the strongest coordination with the catalyst cations. The iodide anions formed from the organic reaction step quickly occupy almost all the catalytic sites. Once the iodide anions are attached to catalytic sites they are very di$cult to replace by acetate anions, and thus render the sites inactive. This `poisoninga e!ect of iodide on the catalytic sites explains the steep decrease of octyl iodide conversion rates shortly after the reaction has started. Contrary to the results of the octyl iodide system, the octyl chloride system has the lowest organic reaction rate constant and the lowest equilibrium attachment constant for its leaving anion, chloride. The low degree of coordination of chloride to the catalyst cation favors the ionexchange with acetate. Furthermore, since the organic reaction rate is slow, the amount of chloride anion formed and transferred to the aqueous phase is low which reduces the degree of competition between chloride and acetate in coordination with the catalyst cation. Because of these conditions, the ion-exchange step will not a!ect the overall reaction rate. In the model, the

5030


Fig. 13. Actual octyl halide conversion and reaction rate plots for octyl acetate synthesis using di!erent octyl halides in the presence of polymer-supported TBMAC catalyst. (a) Octyl halide conversion plots, (b) Reaction rate plots. Reaction conditions: organic/ aqueous phase volumes: 0.070/0.070 l; catalyst: 0.015 mol/l ; temperature: 953C; agitation speed: 600 rpm.

value of the hyperbolic term is approximately equal to unity, thus reducing it to pseudo-"rst-order kinetics. The octyl bromide system places second among the three halide systems in both the organic reaction rate constant and the equilibrium attachment constant of its leaving anion. Bromide anions transferred to the aqueous phase interfere with the coordination of acetate with the catalyst cation, and consequently the ion-exchange step cannot be ignored.

7. Conclusion In the present work, a general kinetic model for triphase catalytic systems has been developed. This model is based on the traditional kinetic mechanisms of the Langmuir}Hinshelwood and Eley}Rideal types modi"ed to suit the special case of catalysis by solidsupported PT catalyst (i.e. TPC). The synthesis of octyl acetate from reaction between octyl bromide and


5031

Fig. 14. Predicted and actual reaction rate plots for the octyl iodide-potassium acetate system in the presence of polymer-supported TBMAC catalyst. Reaction conditions: organic/ aqueous phase volumes: 0.070/0.070 l; catalyst: 0.015 mol/l ; temperature: 953C; agitation speed: 600 rpm. (Legend: octyl iodide (mol/l )/potassium acetate (mol/l ).)

Fig. 15. Predicted and actual reaction rate plots for the octyl chloride-potassium acetate system in the presence of polymer-supported TBMAC catalyst. Reaction conditions: organic/ aqueous phase volumes: 0.070/0.070 l; catalyst: 0.015 mol/l ; temperature 953C; agitation speed: 600 rpm. (Legend: octyl chloride (mol/l )/ potassium acetate (mol/l ).)

5032


Table 2 Parameter values of the model for the di!erent octyl halide-potassium acetate systems studied Organic reactant


K ;10 (mol/l )\ -

K ;10 (mol/l )\ 6

K "K /K ;10 - 6

Octyl iodide Octyl bromide Octyl chloride

5.24$1.01 2.39$0.09 0.07$0.04

1.00 1.00 1.00

44.8$25.3 1.79$0.20 0.85$0.65

0.33$0.19 5.65$0.63 28.3$21.7

potassium acetate has been used to test the validity of this model. This model can be used to determine whether a triphase catalytic system is limited by the organic reaction step, ion-exchange step, or a combination of both the steps. Reactions to synthesize octyl acetate by using di!erent octyl halides as the organic reactant have been used to show how the parameters values obtained can be used to classify the systems. These values showed physical relevance since the data were in high conformity with trends that could be expected from the physical and chemical properties of the halide anions and the corresponding alkyl halides.

[S> ) OAc\],[S> ) Br\]

Notation

h

E /7

h

extraction constant, dimensionless k organic reaction rate con stant, l /min mol cat K ion-exchange equilibrium constant, dimensionless K ,K ,K attachment/detachment 6 -A equilibrium constant, l /mol K equilibrium selectivity coef1 "cient, dimensionless (M>X\) , (M>Y\) aqueous-phase species [OAc\] ,[Br\] ,[X\] concentration of inorganic anion in aqueous phase, mol/l (Q>X\) , (Q>Y\) triphase catalytic sites Q Q [Q>Y\] concentration of triphase catalyst attached to Y\ anion per total system volume, mol/l RX , RY organic phase species [RX] concentration of organic substrate in the organic phase, mol/l S> triphase catalyst's cation [S>] total concentration of triphase catalyst per total system volume, mol/l

concentration of OAc\ and Br\ attached to triphase catalyst cation per total system volume, mol/l

Greek letters a h - -

power law parameter for the amount of catalyst, dimensionless fraction of total number of TPC cations attached to OAc\ fraction of total number of TPC cations attached to Br\ fraction of total number of TPC cations attached to both OAc\ and Br\

References Dehmlow, E. V., & Dehmlow, S. S. (1993). Phase transfer catalysis (3rd ed.). Weinheim, Germany: Verlag Chemie. Desikan, S., & Doraiswamy, L. K. (1995). The di!usion-reaction problem in triphase catalysis. Industrial and Engineering Chemistry Research, 34, 3524}3537. Desikan, S., & Doraiswamy, L. K. (1999). A dynamic nonisothermal model for triphase catalysis in organic synthesis. Industrial and Engineering Chemistry Research, 38, 2634}2640. Desikan, S., & Doraiswamy, L. K. (2000). Enhanced activity of polymer supported quaternary ammonium salts. Chemical Engineering Science, in press. Glatzer, H. J., Desikan, S., & Doraiswamy, L. K. (1998). Triphase catalysis: A new rotating disk contactor for measuring mass transfer coe$cients. Chemical Engineering Science, 53(13), 2431}2449. Glatzer, H. J. & Doraiswamy, L. K. (2000a). Rate enhancements due to autocatalysis and heterogenization in phase transfer catalysis: A comparative study. Chemical Engineering Science, 55(21), 5149}5160. Glatzer, H. J. & Doraiswamy, L. K. (2000b). Triphase catalysis: A correlation for Sherwood number using the rotating disk contactor (RDC) developed earlier. Chemical Engineering Science, submitted. Halpern, M. E. (Ed.). (1997). ACS symposium series 659: Phase-transfer catalysis, mechanisms and syntheses. Washington D.C., U.S.A: American Chemical Society. Hel!erich, F. (1962). Ion exchange. New York, USA: McGraw-Hill Book Company, Inc.

J. A. B. Satrio et al. / Chemical Engineering Science 55 (2000) 5013}5033 Hradil, J., Svec, F., Konak, C., & Jurek, K. (1988). Localization of reaction sites in supported catalysts. Reactive Polymers, 9, 81}89. Marconi, P. F., & Ford, W. T. (1983). Catalytic e!ectiveness due to mass transfer limitations in triphase catalysis by polymer supported quaternary onium salts. Journal of Catalysis, 83, 160}167. Naik, S. D., & Doraiswamy, L. K. (1998). Phase transfer catalysis: Chemistry and engineering. A.I.Ch.E. Journal, 44(3), 612}646. Ohtani, N., Inoue, Y., Mukudai, J., & Yamashita, T. (1997). Polystyrene-supported onium salts as phase-transfer catalysts. Their state and reactivity in phase-transfer reaction conditions. In M. E. Halpern, ACS symposium series 659: Phase-transfer catalysis, mechanisms and synthesis. Washington D.C., USA: American Chemical Society (Chapter 19). Regen, S. (1975). Triphase catalysis. Journal of American Chemical Society, 97(20), 5956}5957. Ruckenstein, E., & Hong, L. (1992). Hydrophilic recognition by polymer-supported phase transfer catalyst and its e!ect on reaction activity and selectivity. Journal of Catalysis, 136, 378}391. Sasson, Y., & Neumann, R. (Eds.) (1997) Handbook of phase transfer catalysis. London, U.K: Blackie Academic & Professional, an imprint of Chapman and Hall. Schlundt, P., & Chau, P. C. (1986). The reaction of a polymer-bound phase transfer catalyst in a novel cyclic slurry reactor. Journal of Catalysis, 102, 348.

5033

Starks, C. M., Liotta, C. L., & Halpern, M. (1994). Phase transfer catalysis. New York, U.S.A: Chapman & Hall. Svec, F. (1988). What is the real mechanism of polymer-supported phase transfer catalysis? Pure and Applied Chemistry, 60(3), 377}386. Telford, S., Schlunt, P., & Chau, P. C. (1986). Mechanism of polymersupported phase-transfer catalysis. E!ect of phase ratios on low percent ring substitution microporous polystyrene resin. Macromolecules, 19, 2435}2439. Tomoi, M., & Ford, W. T. (1981). Mechanisms of polymer-supported catalysis. 1. Reaction of 1-bromooctane with aqueous sodium cyanide catalyzed by polystyrene-bound benzyltri-n-butyl-phosphonium ion. Journal of American Chemical Society, 103(13), 3821}3832. Tundo, P., & Badiali, M. (1989). High activity in displacement reactions catalyzed by quaternary onium salts immobilized on inorganic matrices. Reactive Polymers, 10, 55. Wang, M. L., & Yang, H. M. (1991). A pseudo-steady-state approach for triphase catalysis in a batch reactor. Industrial and Engineering Chemistry Research, 30, 2384}2390. Wang, M. L., & Yang, H. M. (1992). Dynamic model of the triphase catalytic reaction. Industrial and Engineering Chemistry Research, 31, 1868}1875. Yadav, G. D., & Mistry, C. K. (1995). A new model of capsule membrane phase transfer catalysis for oxidation of benzyl chloride to benzaldehyde with hydrogen peroxide. Journal Molecular Catalysis A: Chemistry, 102, 67}72.