Chemical Engineering Science 57 (2002) 1355–1377

www.elsevier.com/locate/ces

Phase-transfer catalysis: a new rigorous mechanistic model

for liquid–liquid systems

Justinus A. B. Satrio, L. K. Doraiswamy

∗

Department of Chemical Engineering, Iowa State University of Science and Technology, 2114 Sweeney Hall,

Ames, IA 50011-2230, USA

Received 7 March 2001; accepted 30 January 2002

Abstract

A general kinetic model for phase-transfer catalyzed reactions involving two liquid phases and a homogeneous catalyst has been

developed. The proposed new model introduces the separation ofthe contributions ofthe phase-transfer catalysis (PTC)-enhanced reaction

and the non-PTC reaction toward the overall conversion. The common approach in the past was that the non-PTC reaction was either

ignored or incorporated in the PTC-enhanced reaction. Even more signi5cantly, the model also incorporates terms for explaining the

variability ofcatalyst phase distribution with changes in electrolyte composition in the aqueous phase, which subsequently a7ects the

amount ofinorganic nucleophile that reacts in the organic phase. The industrially important reaction, synthesis ofbenzaldehyde from

benzyl chloride, has been used to validate the model. Three reaction steps, i.e. esteri5cation, hydrolysis, and oxidation, as well as

combinations thereof, were considered. It was found that the model was able to 5t the experimental data well. Model veri5cation was done,

not by parameter estimation by regression, but by determining them from separate sets of experiments. This lends greater credibility to

multiparameter models (like that in the present case). Further analysis showed that the parameter values ofthe individual reactions can be

used to classify the reactions, based on an approach proposed earlier by Starks et al. (Phase Transfer Catalysis, Chapman and Hall, New

Keywords: Phase transfer catalysis; Homogeneous PT catalyst; Kinetics; Mathematical modeling; Multiphase reactions; Benzaldehyde synthesis

1. Introduction

Phase-transfer catalysis (PTC) is one of the most useful

methods ofsynthesizing organic chemicals from two im-

miscible reactants that normally will not contact (and hence

react with) each other. Often it is an attractive alternative

to conventional processes which can be ineAcient due to

high pressure and temperature requirements or due to low

conversions or product selectivities.

It is generally agreed that liquid–liquid (L–L) PTC

reactions in the presence ofa homogeneous phase-transfer

catalyst follow the Starks mechanism. According to this

mechanism the PT catalyst cation travels between the or-

ganic and aqueous phases to transport the inorganic nucle-

ophile and the leaving anion. Other mechanisms proposed

are the interfacial reaction mechanism and the reverse

reaction mechanism (Starks, Liotta, & Halpern, 1994).

∗

Corresponding author. Tel.: +1-515-294-4117;

fax: +1-515-294-2689.

E-mail address: dorai@iastate.edu (L. K. Doraiswamy).

In the intrinsic kinetic modeling ofL–L PTC systems, it

is assumed that, under conditions ofintense agitation, the

phases are in equilibrium with each other, i.e. high and (if

present) constant mass transfer between phases, so that the

rate ofthe overall reaction is controlled by the reaction in the

organic phase. Under these conditions the overall reaction

rate can be assumed to follow pseudo-5rst-order kinetics.

This simple kinetic model is valid provided that the concen-

tration ofthe active PT catalyst in the organic phase is es-

sentially constant. This is possible when one or both ofthe

following two conditions are satis5ed. First, the ratio of the

concentration ofthe leaving anion (i.e. product inorganic an-

ion) to that ofthe nucleophile (i.e. reactant inorganic anion)

in the aqueous phase is essentially unchanged. This condi-

tion usually can be met by ensuring that the reactant an-

ion in the aqueous phase is present in large excess. Second,

the leaving anion (i.e. product inorganic anion) does not in-

terfere with the attachment=coordination ofthe nucleophile

with the PT catalyst cation. This condition is met only when

the aAnity ofthe inorganic nucleophile to the PT catalyst

cation is not much weaker than that ofthe leaving anion.

PII: S0009-2509(02)00061-1

1356

J. A. B. Satrio, L. K. Doraiswamy / Chemical Engineering Science 57 (2002) 1355–1377

There are cases where the pseudo-5rst-order assumption

does not hold, even when the inorganic nucleophile is used

in large excess. In these cases the leaving anions greatly

a7ect the concentration ofthe active PT catalyst in the or-

ganic phase due to their stronger aAnity toward the PT cat-

alyst cation than that ofthe inorganic nucleophiles. For such

cases, a more general kinetic model that incorporates the

e7ect ofthe electrolyte composition in the aqueous phase

is needed. Some studies on non-5rst order kinetics on L–L

PTC systems have been reported, for example: Starks et al.

(1994), and Sasson and Neumann (1997), but no rigorous

modeling has been attempted.

2. Objective of present study

The main objective ofthe present study is to develop a

general kinetic model for a PT catalyzed system involving

two liquid phases and a homogeneous PT catalyst that incor-

porates the e7ect ofchanges ofthe electrolyte composition

in the aqueous phase. Both the PTC-enhanced reaction and

the base reaction (i.e. reaction without the PT catalyst) are

a7ected by the electrolyte composition. In the past, the com-

mon approach was that the non-PTC reaction could either

be ignored or incorporated in the PTC-enhanced reaction. In

some reaction systems the e7ects ofthe ionic concentration

in the aqueous phase on the base and PTC-enhanced reac-

tions can be di7erent. Thus it is considered necessary to sep-

arate the contributions ofthe PTC-enhanced and non-PTC

reactions to the overall conversion.

As a case study for the development and validation of the

model, the industrially important reaction, synthesis ofben-

zaldehyde from benzyl chloride in the presence of a homoge-

nous PT catalyst, will be used. Literature study revealed that

in order to obtain high conversions ofbenzyl chloride with

a high selectivity to benzaldehyde, the reaction might need

to take place in three steps: esteri5cation, hydrolysis, and

oxidation. This multiple-step scheme, besides serving as an

ideal system for studying the e7ects of electrolyte compo-

sition on di7erent PTC systems, has been used to evaluate

the accuracy ofusing values ofmodel parameters obtained

independently to test the validity ofthe model. In this ap-

proach, experimental data for the individual reaction steps

are obtained, from which the parameters of the model are

determined directly. This constitutes an important departure

from parameter estimation by regression.

3. Synthesis of benzaldehyde using PTCmethods

3.1. Previous studies

Many studies have been reported in the literature on the

synthesis ofbenzaldehyde under PTC conditions using var-

ious starting materials. Theoretically, benzaldehyde can be

synthesized by reacting benzyl chloride with an oxidizing

agent. Several studies have reported that, under PTC condi-

tions, benzaldehyde can be synthesized through direct oxida-

tion ofbenzyl chloride by using chromium compounds such

as potassium dichromate (Yadav & Mistry, 1995; Kumpf&

Martinetz, 1984; Fisher & Dowd, 1979; Landini & Rolla,

1979; Cardillo, Orena, & Sandri, 1976). The reaction mecha-

nism ofbenzyl chloride oxidation has also been investigated

(Yadav & Haldavanekar, 1997; Gopalan & Subbarayan,

1979).

Studies show that benzyl alcohol can be readily converted

to benzaldehyde by using an oxidizing agent in the pres-

ence ofa PT catalyst. Various conventional oxidizing agents,

such as hydrogen peroxide (Ma, Ye, Wang, Zhang, & Wu,

1992), dichromate (Pletcher & Tait, 1978) and hypochlo-

rite (Asai, Nakamura, & Sumita, 1994; Do & Chou, 1990;

Abramovici, Neumann, & Sasson, 1985; Lee & Freedman,

1976) have been used. It was reported that high conver-

sion ofbenzyl alcohol with high selectivity to benzaldehyde

could be obtained at very mild reaction conditions.

Since benzyl alcohol can be easily converted to benzalde-

hyde under PTC conditions, an alternative route for synthe-

sizing benzaldehyde from benzyl chloride is to 5rst convert

benzyl chloride to benzyl alcohol. The benzyl alcohol pro-

duced is then oxidized to benzaldehyde. Both steps can be

carried out under PTC conditions. Theoretically, the con-

version ofbenzyl chloride to benzyl alcohol can be accom-

plished by using a single hydroxide anion displacement step,

i.e. by reacting benzyl chloride with OH

−

(hydroxide) an-

ion. However, simple PTC displacements with OH

−

to yield

alcohol are diAcult to accomplish since dialkyl ether is the

preferred 5nal product (Wang, Huang, & Yeh, 1990a; Her-

riott & Picker, 1972). It has also been reported that the selec-

tivity ofbenzyl chloride conversion to benzyl alcohol could

be improved by immobilizing the PT catalyst on a capsule

membrane (Yadav & Mehta, 1993).

To improve the conversion ofalkyl halides to their cor-

responding alcohols, Zahalka and Sasson (1983) suggested

that a two-step procedure involving acetate displacement

followed by hydrolysis of the acetate esters is the best pro-

cedure. Both steps could be accomplished using PTC reac-

tions, and the acetate-containing phase from the second step

could be recycled to the 5rst for more complete utilization

ofthe acetate.

Wang, Huang, and Yeh (1990b) reported that a high yield

ofbenzyl alcohol could be achieved through weak alkaline

hydrolysis ofbenzyl chloride under PTC conditions using a

carboxylate salt such as sodium acetate as cocatalyst. The

idea ofadding carboxylate salts to the system as cocatalysts

seems to be in line with the two-step process suggested

by Zahalka and Sasson (1983). Under PTC conditions, the

presence ofsodium acetate causes benzyl chloride to be

converted 5rst to benzyl acetate since, compared to the OH

−

anion, OAc

−

(acetate) anion is more readily attached to the

PT catalyst cation. In turn, OAc

−

is a better leaving anion

compared to the Cl

−

(chloride) anion. Thus, compared to

J. A. B. Satrio, L. K. Doraiswamy / Chemical Engineering Science 57 (2002) 1355–1377

1357

benzyl chloride, benzyl acetate is more easily hydrolyzed

by the OH

−

anion.

3.2. Proposed possible routes for synthesizing

benzaldehyde from benzyl chloride

Based on information from the literature and data from

laboratory experiments, benzaldehyde may be synthesized

under PTC conditions from benzyl chloride as follows:

(1) Contacting benzyl chloride with an oxidizing agent, e.g.

NaOCl (sodium hypochlorite). The overall reaction is

PhCH

org

+ (OCl

−

)

→ PhCHO

org

+ (HCl)

+ (Cl

−

)

(R.1)

It is thought that the reaction mechanism must involve

the formation of benzyl alcohol intermediate from the

reaction ofbenzyl chloride with the OH

−

anion released

from the interaction of OCl

−

(hypochlorite) anion with

water.

(2) Contacting benzyl chloride with a hydrolyzing agent,

e.g. NaOH (sodium hydroxide), and NaOCl. The over-

all reactions are

Step 1: Hydrolysis ofthe benzyl chloride intermedi-

ate by the OH

−

anion to form benzyl alcohol:

PhCH

org

+ (OH

−

)

→ PhCH

org

+ (Cl

−

)

(R.2)

Step 2. Oxidation ofbenzyl alcohol by the OCl

−

anion to form the 5nal product benzaldehyde:

PhCH

org

+ (OCl

−

)

→ PhCHO

org

+ (Cl

−

)

+ H

(R.3)

The PTC reaction mechanism ofthe reaction is similar

to that ofthe benzyl chloride=NaOCl system. The only

di7erence is that an external source ofNaOH is added.

(3) Contacting benzyl chloride with an esteri5cation agent,

e.g. NaOAc (sodium acetate), NaOH, and NaOCl, lead-

ing to the following steps:

Step 1. Esteri5cation ofbenzyl chloride with the

OAc

−

anion to form benzyl acetate intermediate:

PhCH

org

+ (OAc

−

)

→ PhCH

OAc

org

+ (Cl

−

)

(R.4)

Step 2. Hydrolysis ofthe benzyl acetate intermediate

by the OH

−

anion to form benzyl alcohol:

PhCH

OAc

org

+ (OH

−

)

→ PhCH

org

+ (OAc

−

)

(R.5)

Step 3. Oxidation ofbenzyl alcohol by the OCl

−

anion to form the 5nal product benzaldehyde:

PhCH

org

+ (OCl

−

)

→ PhCHO

org

+ (Cl

−

)

+ H

(R.3)

As shown in Fig. 1, theoretically, from the one-step,

two-step, and three-step reactions, various routes can be

proposed by varying the manner in which these steps are

carried out. Recently, extensive kinetic and feasibility stud-

ies on these routes (culminating in a novel PTC-based

benzaldehyde process) have been reported by Satrio and

Doraiswamy (2001).

4. Mechanism of the three-step reaction

The synthesis ofbenzaldehyde from benzyl chloride pro-

vides an ideal case study for the development of a kinetic

model for multi-step L–L PTC systems that involve multiple

inorganic nucleophiles and organic reagents. The three-step

reaction mechanism is selected for the development of the

model since it is the most comprehensive one. This model

can be readily simpli5ed to the two-step and one-step mech-

anisms.

4.1. Main reactions

In order for the three-step reaction of the conversion of

benzyl chloride to benzaldehyde to take place, theoretically,

the inorganic nucleophiles involved in the steps must exist

together in the aqueous phase. The reaction mechanism then

will be as follows:

Extractions ofthe inorganic anions (OAc

−

, OH

−

, OCl

−

and Cl

−

) from the aqueous phase to the organic phase or

interface by the PT catalyst cations:

)

+ (OAc

−

)

QOAc

↔ (Q

OAc

−

)

;

(R.6a)

OAc

−

)

QOAc

↔ (Q

OAc

−

)

org

;

(R.6b)

)

+ (OH

−

)

QOH

↔ (Q

−

)

;

(R.7a)

−

)

QOH

↔ (Q

−

)

org

;

(R.7b)

)

+ (OCl

−

)

QOCl

↔ (Q

OCl

−

)

;

(R.8a)

OCl

−

)

QOCl

↔ (Q

OCl

−

)

org

;

(R.8b)

)

+ (Cl

−

)

QCl

↔ (Q

−

)

;

(R.9a)

−

)

QCl

↔ (Q

−

)

org

;

(R.9b)

where K

QOAc

, K

QOH

, K

QOCl

, and K

QCl

are the equi-

librium constants ofthe association ofthe PT catalyst cation

with OAc

−

, OH

−

, OCl

−

, and Cl

−

, respectively, in the

aqueous phase; K

QOAc

, K

QOH

, K

QOCl

, and K

QCl

are

the equilibrium phase distribution constants ofQ

OAc

−

, Q

OCl

−

, and Q

−

, respectively, between the

aqueous and organic phases.

1358

J. A. B. Satrio, L. K. Doraiswamy / Chemical Engineering Science 57 (2002) 1355–1377

Fig. 1. Schematic of routes for benzaldehyde synthesis from benzyl chloride (adapted from Satrio and Doraiswamy, 2001).

Reactions in the organic phase or at the interface:

Reactions between organic substrates and the inorganic

nucleophiles attached to the PT catalyst cations (i.e.

PTC-enhanced organic phase reactions):

PhCH

org

+ (Q

OAc

−

)

org

PTC

→ PhCH

OAc

org

+ (Q

−

)

org

;

(R.10a)

PhCH

org

+ (Q

−

)

org

PTC

→ PhCH

org

+ (Q

−

)

org

;

(R.11a)

PhCH

OAc

org

+ (Q

−

)

org

PTC

→ PhCH

org

+ (Q

OAc

−

)

org

;

(R.12a)

PhCH

org

+ (Q

OCl

−

)

org

PTC

→ PhCHO

org

+ (Q

−

)

org

+ H

(R.13a)

Reactions between organic substrates and free inorganic

nucleophiles (i.e. base reactions):

PhCH

org

+ (OAc

−

)

base

→ PhCH

OAc

org

+ (Cl

−

)

;

(R.10b)

PhCH

org

+ (OH

−

)

base

→ PhCH

org

+ (Cl

−

)

;

(R.11b)

PhCH

OAc

org

+ (OH

−

)

base

→ PhCH

org

+ (OAc

−

)

;

(R.12b)

PhCH

org

+ (OCl

−

)

base

→ PhCHO

org

+ (Cl

−

)

+ H

(R.13b)

where k

PTC

and k

base

are, respectively, the rate constants

for the PTC and base reactions of benzyl acetate formation;

PTC

and k

base

are, respectively, the rate constants for

the PTC and base reactions ofbenzyl chloride conversion

J. A. B. Satrio, L. K. Doraiswamy / Chemical Engineering Science 57 (2002) 1355–1377

1359

) + (OAc

)

OAc

)

OCl

)

) + (OCl

)

) + (OH

)

) + (OH

)

Organic Phase

Interface

Aqueous Phase

(OCl

) + H

(HOCl) + (OH

)

NaOCl

(Na

) + (OCl

)

) + (Cl

)

) +

PhCH

OH + (Q

)

PhCH

OH + (Q

OCl

)

PhCHO + (Q

) + (H

)

) +

)

(Cl

) +

OAc

PhCH

OH + (Q

OAc

)

NaOAc

(Na

) + (OAc

)

(OAc

) + H

(HOAc) + (OH

)

OAc

) + PhCH Cl

PhCH Cl

PhCH

OAc + (Q

)

PhCH

OAc + (OH

)

OH + (OAc

)

PhCH

OH + (OCl

)

PhCHO + (Cl

) + (H

PhCH

Cl + (OAc

)

OAc + (Cl

)

PhCH

Cl + (OH

)

OH + (Cl

)

) + (Cl

)

NaOH

(Na

) + (OH

)

OAc )

(Q ) + (OAc )

PhCH

Fig. 2. Diagram ofthe three-step PTC mechanism ofthe synthesis ofbenzaldehyde from benzyl chloride (adapted from Satrio and Doraiswamy, 2001).

to benzyl alcohol; k

PTC

and k

base

are, respectively, the

rate constants for the PTC and base reactions of benzyl

acetate conversion to benzyl alcohol; and k

PTC

and k

base

are, respectively, the rate constants for the PTC and base

reactions ofbenzyl alcohol conversion to benzaldehyde. The

overall PTC mechanism ofthese steps is shown in Fig. 2.

Note that, although OH

−

can be supplied from an external

source such as NaOH, theoretically, the anion can be formed

from equilibrium reactions between the OCl

−

and OAc

−

anions with water.
(OAc

−

)

+ (H

)

OAc

↔ (H

OAc

−

)

+ H

(R.14a)

(OCl

−

)

+ (H

)

OCl

↔ (H

OCl

−

)

+ H

(R.14b)

↔(H

)

+ OH

−

(R.14c)

4.2. Formation of by-product

When benzyl alcohol and benzyl chloride are present in

the organic phase, undesired benzyl ether may be formed.

Benzyl alcohol is slightly soluble in water. In the presence

ofhigh concentrations ofOH

−

in the aqueous phase, benzyl

alcohol will undergo proton transfer with OH

−

to form a

benzyl oxide (PhCH

−

) anion and water.

PhCH

+ (OH

−

)

PhCH2O

↔ (PhCH

−

)

+ H

(R.15)

Theoretically, the PhCH

−

anion formed then will com-

pete with other anions (i.e. OAc

−

, OCl

−

, and Cl

−

anions)

in the aqueous phase for association with the PT catalyst

cation to form (Q

PhCH

−

), which subsequently trans-

fers to the organic phase and reacts with benzyl chloride in

the aqueous phase to form benzyl ether.

PhCH

org

+ (Q

PhCH

−

)

org

PTC

→ PhCH

OCH

org

+ (Q

−

)

org

;

(R.16a)

PhCH

org

+ (PhCH

−

)

aq=int

base

→ PhCH

OCH

org

+ (Q

−

)

aq=int

(R.16b)

To reduce the formation of benzyl ether when there is a sig-

ni5cant fraction of benzyl chloride still left unreacted, nei-

ther benzyl alcohol nor the OH

−

anion should be present in

signi5cant amounts. This explains the necessity ofconvert-

ing benzyl chloride to benzyl acetate intermediate before

forming benzyl alcohol. The OAc

−

anion in the aqueous

phase functions in two ways. First, it serves as the inorganic

reagent in the formation of benzyl acetate, and second, in

large amount, it serves as an inhibitor for the formation of

−

) which increases the formation of benzyl alcohol

and subsequently benzyl ether. Compared to OH

−

, theoreti-

cally OAc

−

is more easily attached to the PT catalyst cation

(Starks et al., 1994).

1360

J. A. B. Satrio, L. K. Doraiswamy / Chemical Engineering Science 57 (2002) 1355–1377

5. Development of a rigorous kinetic model

5.1. Assumptions

The following assumptions are made:

1. All the ion-exchange reaction steps are much faster

than all the organic-phase reaction steps; thus, the

ion-exchange steps are always in equilibrium.

2. The PTC-enhanced and base reactions take place inde-

pendent ofeach other.

3. Di7usion coeAcients ofcatalyst and reactants are con-

stant, thus can be ignored.

4. There is no change in phase volumes during reaction.

5. Organic reactant and product are insoluble in the aque-

ous 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.

9. There is no catalyst decomposition.

10. Solvent properties do not change during reaction.

5.2. Development

The reactions are assumed to be kinetically controlled.

Hence mass transfer e7ects are not included. Accounting

for these e7ects requires a separate study dedicated to this

aspect ofthe analysis.

As shown in Fig. 2, the conversion ofbenzyl chloride to

benzaldehyde takes place via three organic reaction steps:

(1) formation of benzyl acetate from benzyl chloride via

esteri5cation, (2) formation of benzyl alcohol from hydrol-

ysis ofbenzyl acetate, and (3) oxidation ofbenzyl alcohol

to form benzaldehyde. During the course of these reactions,

two other reactions may take place: (1) formation of benzyl

alcohol via hydrolysis ofbenzyl chloride, and (2) formation

ofbenzyl ether from benzyl chloride with the PhCH

−

an-

ion which is derived from benzyl alcohol. The reaction rate

expressions for the formation of benzyl acetate, benzyl al-

cohol, benzaldehyde, benzyl chloride, and benzyl ether are

d[PhCH

OAc]

org

=[PhCH

Cl]

org=int

PTC

OAc

−

]

org=int

+ k

base

[OAc

−

]

aq=int

)

− [PhCH

OAc]

org=int

PTC

−

]

org=int

+ k

base

[OH

−

]

aq=int

);

(1)

d[PhCH

OH]

org

=[PhCH

Cl]

org=int

PTC

−

]

org=int

+ k

base

[OH

−

]

aq=int

)

+ [PhCH

OAc]

org=int

PTC

−

]

org=int

+ k

base

[OH

−

]

aq=int

)

− [PhCH

OH]

org=int

PTC

OCl

−

]

org=int

+ k

base

[OCl

−

]

aq=int

);

(2)

d[PhCHO]

org

= [PhCH

OH]

org=int

PTC

OCl

−

]

org=int

+ k

base

[OCl

−

]

aq=int

);

(3)

d[PhCH

Cl]

org

= − [PhCH

Cl]

org=int

PTC

OAc

−

]

org=int

+ k

base

[OAc

−

]

aq=int

)

− [PhCH

Cl]

org=int

PTC

−

]

org=int

+ k

base

[OH

−

]

aq=int

);

(4)

d[PhCH

OCH

Ph]

org

=[PhCH

Cl]

org

PTC

PhCH

−

]

org=int

+ k

base

[PhCH

−

]):

(5)

To utilize Eqs. (1)–(5), the unknown concen-

trations ofthe PT catalyst in the organic phase, i.e.

OAc

−

]

org

, [Q

−

]

org

, [Q

−

]

org

, [Q

OCl

−

]

org

(and

PhCH

−

]

org

), need to be expressed using variables

that can be directly measured, i.e. concentrations ofthe in-

organic anions OAc

−

, Cl

−

, OH

−

, OCl

−

(and PhCH

−

)

in the aqueous phase. The derivation is explained below.

First, we obtain expressions for the concentrations of the

PT catalyst attached to the OAc

−

, Cl

−

, OH

−

, OCl

−

, (and

PhCH

−

) anions in the aqueous phase and the organic

phase. They are given in Eq. (6) through Eq. (15).
[Q

OAc

−

]

= K

QOAc

]

[OAc

−

]

;

(6)

−

]

= K

QOH

]

[OH

−

]

;

(7)

OCl

−

]

= K

QOCl

]

[OCl

−

]

;

(8)

−

]

= K

QCl

]

[Cl

−

]

;

(9)

PhCH

−

]

= K

QPhCH

]

[PhCH

−

]

; (10)

OAc

−

]

org

= K

QOAc

OAc

−

]

;

(11)

−

]

org

= K

QOH

−

]

;

(12)

J. A. B. Satrio, L. K. Doraiswamy / Chemical Engineering Science 57 (2002) 1355–1377

1361

OCl

−

]

org

= K

QOCl

OCl

−

]

;

(13)

−

]

org

= K

QCl

−

]

;

(14)

PhCH

−

]

org

= K

QPhCH

PhCH

−

]

(15)

Introducing the total amount ofPT catalyst in the system as

, we compute N

as the sum ofthe catalyst amounts in

the organic and aqueous phases.

= V

]

+ [Q

OAc

−

]

+ [Q

−

]

+ [Q

OCl

−

]

+ [Q

−

]

+ [Q

PhCH

−

]

+ V

org

OAc

−

]

org

+ [Q

−

]

org

+[Q

OCl

−

]

org

+ [Q

−

]

org

+[Q

PhCH

−

]

org

(16)

In terms oftotal catalyst concentration per organic phase

volume, [Q

], and by assuming equal organic and aqueous

phase volumes, Eq. (16) can be rewritten as

] = [Q

]

+ [Q

OAc

−

]

+ [Q

−

]

+ [Q

OCl

−

]

+ [Q

−

]

+ [Q

PhCH

−

]

+ [Q

OAc

−

]

org

+ [Q

−

]

org

+ [Q

OCl

−

]

org

+ [Q

−

]

org

+ [Q

PhCH

−

]

org

;

(17)

where [Q

] = N

org

Expressing free Q

and Q

−

in the aqueous phase in

terms ofthe corresponding salt ion pairs in the organic phase

by incorporating Eqs. (6)–(15) into Eq. (17), we obtain

] =

OAc

−

]

org

QOAc

−

]

org

QOH

OCl

−

]

org

QOCl

−

]

org

QCl

PhCH

−

]

org

QPhCH

OAc

−

]

org

QOAc

−

]

org

QOH

OCl

−

]

org

QOCl

−

]

org

QCl

PhCH

−

]

org

QPhCH

+ [Q

OAc

−

]

org

+ [Q

−

]

org

+ [Q

OCl

−

]

org

+ [Q

−

]

org

+ [Q

PhCH

−

]

org

(18)

] =

comp=i




[Q

−

]

org

−

]

org

−

]

org




 ;

(19)

where i = OAc

−

, OH

−

, OCl

−

, Cl

−

, and PhCH

−

In order to use Eq. (19) to obtain [Q

OAc

−

]

org

, for in-

stance, the concentration ofany other ion pair in the organic

phase must be expressed in term of[Q

OAc

−

]

org

. This is

done by introducing the overall equilibrium equation ofthe

ion-exchange reaction between OAc

−

and i

−

anions.

OAc

−

)

org

+ (i

−

)

OAc

↔ (Q

−

)

org

+ (OAc

−

)

OAc

−

]

org

[OAc

−

]

OAc

−

]

org

−

]

QOAc

(20)

Incorporating Eq. (20) into Eq. (19) and rearranging, we

obtain the following expression:




[Q

OAc

−

]

org

OAc

−

]

org

QOAc

OAc

−

]

org

QOAc






=[Q

] −

comp=i






OAc

−

]

org

OAc

−

]

[OAc

−

]

OAc

−

]

org

OAc

−

]

[OAc

−

]

OAc

−

]

org

OAc

−

]

[OAc

−

]




 :

(21)

By knowing the concentrations ofall the electrolyte salts in

the aqueous phase and using the values of K

and K

ofeach PT catalyst salt ion pair involved in the system, the

concentration of(Q

OAc

−

) in the organic phase can be

estimated by using Eq. (21). The same approach is used for

the other PTC salt ion pairs.

From calculations based on experimental data, it is noted

that the values ofthe equilibrium association constant ofthe

PT catalyst cations with the inorganic anions in the aqueous

phase are generally quite large, which indicates that they

tend to form ion pairs. When the aqueous phase contains

high amounts ofelectrolytes, it is probably safe to assume

that all the PT catalyst cations in the aqueous phase exist

as ion pairs. With this assumption the square root terms in

1362

J. A. B. Satrio, L. K. Doraiswamy / Chemical Engineering Science 57 (2002) 1355–1377

Eq. (21) which account for the concentration of free Q

can be neglected. Thus the non-linear equation reduces to a

simpler form that can be solved directly to give

OAc

−

]

org

]

1 + 1=K

QOAc

comp=i

OAc

−

]=[OAc

−

](1 + 1=K

))

(22)

By the same approach, similar expressions for the concentra-

tions ofthe PT catalyst cation attached to the OH

−

, OCl

−

, (and PhCH

−

) anions in the organic phase can also

be obtained. These expressions then are used along with

the main reaction rate equations (Eqs. (1)–(5)) to calculate

the concentrations ofthe products and reactants during the

course ofthe reactions. Note that, when not supplied exter-

nally, the OH

−

(and PhCH

−

) anions can be present in the

aqueous phase via the interaction ofwater with the OAc

−

and OCl

−

anions and the interaction ofdissolved benzyl al-

cohol with the OH

−

anion.

[OH

−

] =

OAc

O][OAc

−

]

OAc

−

]

OCl

O][OCl

−

]

OCl

−

]

OAc

[OAc

−

] +

OCl

[OAc

−

];

(23)

[PhCH

−

]

PhCH

[PhCH

OH]

[OH

−

]

(24)

5.3. PT catalyst phase-distribution coe7cients, K

In order to incorporate the ionic e7ect on the phase dis-

tribution ofindividual PT catalyst ion pairs, the equilib-

rium phase-distribution coeAcients, K

, used in Eq. (22)

will need to be expressed as functions of ionic strength of

the aqueous solution based on the electrolyte components

present in the phase. To determine K

as a function of

the aqueous phase ionic strength (aqueous phase contains

MX salt), values of K

at various MX salt concentrations

in the aqueous phase are obtained. K

values are related

to the ionic strength ofthe aqueous phase by the following

correlation:

log

QX;w

= k

;

(25)

where K

QX;w

is the phase-distribution coeAcient ofQ

−

ion pair in the absence ofelectrolyte in the aqueous phase,

is the salting-out parameter for the catalyst-electrolyte

MX solution system, and I

is the ionic strength ofthe

electrolyte MX solution. Based on this correlation, a plot

oflog(K

QX;w

) vs. I

is expected to show a linear

relationship with a slope of k

. This correlation is an

extension to PTC systems ofan earlier correlation developed

by Van Krevelen and Hoftijzer (1948) for gas solubility in

single-electrolyte solution.

When the aqueous phase is a mixed electrolyte so-

lution, the salting out e7ect ofeach ofthe electrolytes

must be considered. The phase-distribution coeAcient of

−

) then can be predicted from a correlation used by

Danckwerts and Gillham (1966) for gas solubility in a

multi-electrolyte solution, which was also based on the cor-

relation ofvan Krevelen and Hoftijzer and which has been

shown by Asai, Nakamura, Tanabe, and Sakamoto (1993),

Asai, Nakamura, and Furuichi (1991) to work well for PT

catalyst phase distribution.

log

QX;w

;

(26)

where j

−

and Mj are an individual anion and its correspond-

ing salt in the aqueous phase, respectively. By knowing the

values of k

and I

for each of the electrolytes present

in the aqueous phase, the phase-distribution coeAcient of

the Q

−

salt ion pair can be determined. Eq. (26) is then

incorporated in Eq. (22).

6. Model veri)cation

To test the kinetic model developed above, the several

routes to convert benzyl chloride to benzaldehyde shown

previously in Fig. 2 were used. Model equations to describe

one- and two-step, and three-step reaction systems are shown

in Tables 1 and 2, respectively. In order to utilize these model

equations, the organic-phase reaction rate constants, k

org

PTC

and k

org

base

, the PT catalyst’s equilibrium association coef-

5cients, K

, and the equilibrium phase-distribution co-

eAcients, K

, were measured separately. For measur-

ing k

org

it is assumed that the presence ofPT catalyst does

not a7ect the interfacial area which is important for the

non-catalyzed reaction, while for measuring K

it is as-

sumed that the reactants in the organic phase do not a7ect

the distribution ofPT catalyst. Further, since the model does

not incorporate the e7ect ofmass transfers, in order to re-

move the mass transfer e7ect, all the experimental data were

taken at the same high agitation speed, i.e. 700 rpm. More

comprehensive models should incorporate the e7ect ofmass

transfer, i.e by including the e7ect of agitation on model

parameters

6.1. Experimental procedure

Experimental data were obtained to determine the

organic-phase reaction rate constants and the equilibrium

phase-distribution and association coeAcients ofthe PT

catalyst required to test the models. Unless stated other-

wise, all reagents were ofanalytical grade and not further

puri5ed. The soluble PT catalysts used in this study were:

A.
B

.Satrio,

L.
K.
Doraiswamy

/Chemical

Engineering

Science

57
(2002)

1355–1377

1363

Table 1

Final equations to be solved for benzyl chloride/NaOCl and benzyl chloride/NaOCl/NaOH systems

Organic phase reactions

Aqueous phase reactions

d[PhCH

OH]

org

= [PhCH

Cl]

org=int

PTC

−

]

org=int

+ k

base

[OH

−

]

0:45

aq=int

)

−[PhCH

OH]

org=int

PTC

OCl

−

]

org=int

+ k

base

[OCl

−

]

aq=int

)

d[Na

−

]

d[PhCHO]

org

d[PhCH

Cl]

org

−

d[PhCH

Cl]

org

= [PhCH

Cl]

org=int

PTC

−

]

org=int

+ k

base

[OH

−

]

0:45

aq=int

)

−

d[Na

OCl

−

]

d[PhCHO]

org

−

d[PhCHO]

org

= [PhCH

OH]

org=int

PTC

OCl

−

]

org=int

+ k

base

[OCl

−

]

aq=int

)

[OH

−

] =

OCl

[OCl

−

] (benzyl chloride/NaOCl system)

d[Na

−

]

= −

d[PhCH

Cl]

org

(benzyl chloride/NaOCl/NaOH system)

Amounts of active PT catalyst in organic phase

OCl

−

]

org

]

QOCl

i=Cl;OH

OCl

−

]

[OCl

−

]

−

]

org

]

QOH

i=OCl;Cl

−

]

[OH

−

]

−

]

org

]

QCl

i=OCl;OH

−

]

[Cl

−

]

Equilibrium phase distribution constants

Ion exchange equilibrium constants

log

QCl

QCl;w

j=OCl;OH;Cl

Na+j−

OCl

QCl

QOCl

log

QCl

QOCl;w

j=OCl;OH;Cl

Na+j−

OCl

QOH

QOCl

log

QOH

QOH;w

j=OCl;OH;Cl

Na+j−

QCl

QOH

1364

A.
B

.Satrio,

L.
K.
Doraiswamy

/Chemical

Engineering

Science

57
(2002)

1355–1377

Table 2

Final equations to be solved for benzyl chloride/NaOAc/NaOCl/NaOH system

Organic phase reactions

Aqueous phase reactions

Amounts of (Q

−

)

org

d[PhCH

Cl]

org

= −[PhCH

Cl]

org=int

PTC

OAc

−

]

org=int

base

[OAc

−

]

0:45

aq=int

−[PhCH

Cl]

org=int

PTC

−

]

org=int

base

[OH

−

]

0:45

aq=int

d[Na

−

]

d[PhCHO]

org

d[PhCH

Cl]

org

OAc

−

]

org

]

QOAc

OAc

−

]

[OAc

−

]

d[PhCH

OAc]

org

= [PhCH

Cl]

org=int

PTC

OAc

−

]

org=int

+ k

base

[OAc

−

]

0:45

aq=int

−[PhCH

OAc]

org=int

PTC

−

]

org=int

+ k

base

[OH

−

]

0:15

aq=int

d[Na

OAc

−

]

d[PhCH

OAc]

org

−

]

org

]

1 +

QOH

−

]

[OH

−

]

1 +

d[PhCH

OH]

org

= [PhCH

Cl]

org=int

PTC

−

]

org=int

+ k

base

[OH

−

]

0:45

aq=int

+[PhCH

OAc]

org=int

PTC

−

]

org=int

+ k

base

[OH

−

]

0:15

aq=int

−[PhCH

OH]

org=int

PTC

OCl

−

]

org=int

+ k

base

[OCl

−

]

aq=int

d[Na

OCl

−

]

d[PhCHO]

org

−

]

org

]

1 +

QCl

−

]

[Cl

−

]

1 +

d[PhCHO]

org

= −[PhCH

OH]

org=int

PTC

OCl

−

]

org=int

+ k

base

[OCl

−

]

aq=int

−

d[Na

−

]

d[PhCH

OH]

org

d[PhCHO]

org

OCl

−

]

org

]

1 +

QOCl

i=Cl;OH

OCl

−

]

[OCl

−

]

1 +

d[PhCH

OCH

Ph]

org

= [PhCH

Cl]

org=int

PTC

[OH

−

]

0:67

org=int

[OH

−

] =

OAc

[OAc

−

] +

OCl

[OCl

−

] (cases with combined NaOAc/NaOCl without external NaOH)

[OH

−

] =

OAc

[OAc

−

] (cases with separate esteri5cation step)

Equilibrium phase distribution constants

Ion exchange equilibrium constants

log

QCl

QCl;w

j=OCl;OH;Cl;OAc

Na+j−

log

QOH

QOH;w

j=OCl;OH;Cl;OAc

Na+j−

OCl

QCl

QOCl

OCl

QOH

QOCl

OAc

OCl

QOAc

QOCl

log

OCl

QOCl;w

j=OCl;OH;Cl;OAc

Na+j−

log

OAc

QOAc;w

j=OCl;OH;Cl;OAc

Na+j−

QOH

QCl

OAc

QOH

QOAc

OAc

QCl

QOAc

J. A. B. Satrio, L. K. Doraiswamy / Chemical Engineering Science 57 (2002) 1355–1377

1365

tetrabutylammonium chloride (TBAC) (obtained from

Fluka Chemical Corp.) and tetrabutylammonium acetate

(TBAA) and tetrabutylammonium hydroxide (TBAOH)

(obtained from Aldrich Chemicals). The organic reac-

tants (benzyl chloride, benzyl alcohol, and benzyl acetate)

were dissolved in toluene and the solid inorganic reactants

(sodium acetate, sodium hydroxide, and sodium hypochlo-

rite) in deionized water.

The kinetic experiments were carried out in a 300-ml

stainless steel reactor from PARR Instruments. The vessel

was lined with a TeSon insert and equipped with a two-blade

paddle. The temperature could be controlled to within 1

◦

For all reactions, the total volume ofthe mixture was 140 ml

with equal volume fractions of the liquid phases. Also, un-

less otherwise noted, the agitation speed was maintained

at 700 rpm in order to maintain a high mass transfer rate.

The concentrations ofthe organic reactants and products

were measured by using a Perkin Elmer gas chromatograph

(Model 3000 Autosystem with FID). A packed column (Car-

bopack, 10% SP-2250 from Supelco Inc.) with a length of

2:0 m and a diameter of1=8 in was used for the analysis.

An external standard was used.

In determining the PT catalyst phase-distribution and

association coeAcients, a measured amount ofthe cat-

alyst was dissolved in the aqueous phase. The solu-

tion was brought into contact with the second phase of

equal volume and stirred vigorously for 1 h. The con-

centrations ofthe catalyst cation in the aqueous phase

were then determined by the ion pair extraction-titration

method with tetrabromo-phenolphthalein ethyl ester as in-

dicator developed by Sakai, Tsubouchi, Nakagawa, and

Tanaka (1977).

6.2. Estimation of K

QX;w

and K

and k

values

TBAC, TBAA, and TBAOH catalysts were used to ob-

tain the phase-distribution and association coeAcients of

−

), (Q

OAc

−

), and (Q

−

), respectively. TBAC

catalyst and NaOCl were used to obtain the coeAcients for

OCl

−

). Experimental measurements at six levels ofcat-

alyst concentrations ranging from 0.05 to 0:3 mol=l

org

were

made with two replications ofeach level. The correspond-

ing sodium salts were used to estimate the salting-out pa-

rameter values. The salts were used in amounts to provide

aqueous phase solutions ofionic strengths ranging from

Table 3

Values of K

, K

QX;w

, E

QX;w

and k

NaX

for the nucleophiles in-

volved in the reactions

Anion, X

QX;w

= K

QX;w

NaX

=mmol)

Chloride

1:150 × 10

−2

5:635 × 10

−2

6:480 × 10

−4

4:750 × 10

−5

Acetate

9:991 × 10

−3

1:831 × 10

−2

1:929 × 10

−4

9:590 × 10

−5

Hypochlorite

3:721 × 10

−2

1:008 × 10

−1

3:751 × 10

−3

1:229 × 10

−4

Hydroxide

6:202 × 10

−3

1:489 × 10

−4

9:235 × 10

−7

8:538 × 10

−5

1 to 8 mol=l

. By using the methods explained in Appendix

A, the values of K

D;QX;w

, K

, and k

were obtained

for each type of catalyst, Q

−

, and its corresponding salt,

NaX, as shown in Table 3. This table also shows values

ofthe stoichiometric extraction constant, E

QX;w

, which is a

combination of K

and K

QX;w

. The trend ofthese val-

ues agrees well with the trend ofrelative aAnities ofthe

anions to the PT catalyst cation as reported in the litera-

ture (Starks et al., 1994; Dehmlow & Dehmlow, 1993), i.e.

hypochlorite ¿ chloride ¿ acetate ¿ hydroxide.

6.3. Estimation of organic reaction rate constants

Experimental data were obtained from the following in-

dividual reactions with and without the PT catalyst in order

to estimate their organic-phase reaction rate constants:

(1) Esteri5cation ofbenzyl chloride by NaOAc (for the rate

constant ofbenzyl acetate formation, k

(2) Hydrolysis ofbenzyl chloride by NaOH (for the rate

constants ofbenzyl alcohol formation, k

, and ofbenzyl

ether formation, k

(3) Hydrolysis ofbenzyl acetate by NaOH (for the rate

constant ofbenzyl alcohol formation, k

(4) Oxidation ofbenzyl alcohol by NaOCl (for the rate

constant ofbenzaldehyde formation, k

For each reaction: the operating conditions are shown in

Table 4; data were obtained by varying the concentrations

ofthe inorganic reactant while holding the concentrations

ofthe organic reactant and PT catalyst constant; and at least

one replication was obtained. Depending on the reaction be-

ing studied, between 6 and 10 samples were drawn during

the course ofthe reaction. For each data point, the concen-

trations ofproducts and reactants were measured.

To obtain the reaction rate constants for the base reac-

tions, power law analysis was applied since the e7ect of

product ions would be negligible at initial conditions. The

linear plots obtained for −ln(1 − X) vs. time showed that

this assumption was valid. The observed pseudo-5rst-order

constants at di7erent concentration levels then were plotted

against the nucleophile concentration to obtain the actual

power law reaction rate constant. Results ofthe analysis are

tabulated in Table 5.

To obtain the rate constants for the PTC-enhanced reac-

tions, again, initially the power law analysis was applied.

1366

J. A. B. Satrio, L. K. Doraiswamy / Chemical Engineering Science 57 (2002) 1355–1377

Table 4

Reaction conditions ofindividual reactions to obtain rate constants

Reaction

Organic reactant

Inorganic reactant

PT catalyst

Temp.

Rate constant

(mol=l

org

)

(mol=l

)

(mmol)

(

◦

obtained

Esteri5cation ofBenzyl chloride,

Sodium acetate,

TBAC

PTC

and k

base

benzyl choride

0.5

1, 1.5, 2

1.75

Hydrolysis ofBenzyl chloride,

Sodium hydroxide

TBAC

PTC

; k

base

benzyl chloride

0.5

0.6, 1, 1.5

1.75

PTC

and k

base

Hydrolysis ofBenzyl acetate,

Sodium hydroxide

TBAC

40, 90

PTC

and k

base

benzyl acetate

0.5

0.6, 1, 1.5

1.75

Oxidation ofBenzyl alcohol,

Sodium hypochlorite

TBAC

23, 30,

PTC

and k

base

benzyl alcohol

0.5

1.75

40, 50

Table 5

Rate constants ofthe steps involved in benzaldehyde synthesis from benzyl chloride

Reaction step

Reaction rate constant

PTC enhanced reaction

base reaction

Formation ofbenzyl acetate

6:48

−4

1:82

−5

from benzyl chloride, k

; 90

◦

org

=mmol(Q

OAc

−

)

org

)=min

=mmol OAc

−

)

0:45

=min

Formation ofbenzyl alcohol

1:61

−1

9:22

−5

from benzyl chloride, k

; 90

◦

org

=mmol(Q

−

)

org

)=min

=mmol OH

−

)

0:45

=min

Formation ofbenzyl alcohol

from benzyl acetate, k

◦

5:23

5:28

−2

◦

2:60

1:05

−3

org

=mmol(Q

−

)

org

)=min

=mmol OH

−

)

0:15

=min

Formation ofbenzaldehyde

from benzyl alcohol, k

◦