Effects of charge separation, effective concentration, and aggregate formation on the phase transfer catalyzed alkylation of phenol

Article

pubs.acs.org/JACS

Eﬀects of Charge Separation, Eﬀective Concentration, and Aggregate

Formation on the Phase Transfer Catalyzed Alkylation of Phenol

Scott E. Denmark,* Robert C. Weintraub, and Nathan D. Gould

Roger Adams Laboratory, Department of Chemistry, University of Illinois, Urbana, Illinois 61801, United States

S

* Supporting Information

ABSTRACT: The factors that inﬂuence the rate of alkylation

of phenol under phase transfer catalysis (PTC) have been

investigated in detail. Six linear, symmetrical tetraalkylammo-

nium cations, Me₄N⁺, Et₄N⁺, (n-Pr)₄N⁺, (n-Bu)₄N⁺, (n-

Hex)₄N⁺, and (n-Oct)₄N⁺, were examined to compare the

eﬀects of cationic radius and lipophilicity on the rate of

alkylation. Tetraalkylammonium phenoxide·phenol salts were

prepared, and their intrinsic reactivity was determined from

initial alkylation rates with n-butyl bromide in homogeneous

solution. The catalytic activity of the same tetraalkylammo-

nium phenoxides was determined under PTC conditions

(under an extraction mechanism) employing quaternary ammonium bromide catalysts. In homogeneous solution the range in

reactivity was small (6.8-fold) for Me₄N⁺to (n-Oct)₄N⁺. In contrast, under PTC conditions a larger range in reactivity was

observed (663-fold). The eﬀective concentration of the tetraalkylammonium phenoxides in the organic phase was identiﬁed as

the primary factor inﬂuencing catalyst activity. Additionally, titration of active phenoxide in the organic phase conﬁrmed the

presence of both phenol and potassium phenoxide aggregates with (n-Bu)₄N⁺, (n-Hex)₄N⁺, and (n-Oct)₄N⁺, each with a unique

aggregate stoichiometry. The aggregate stoichiometry did not aﬀect the PTC initial alkylation rates.

nucleophile activation, and nucleophile desolvation to catalytic

activity have remained an outstanding question for nearly three

decades. Accordingly, the a priori prediction of catalyst

eﬃciency for any speciﬁc reaction remains an elusive objective.

Recently, we initiated a research program aimed at

elucidating the structural features that govern the activity and

enantioselectivity of quaternary ammonium ion phase transfer

catalysts.⁴After evaluation of many quantitative structure−

activity relationship (QSAR) models for catalyst activity in an

enolate alkylation, we were led to the conclusion that the

activity of quaternary ammonium ion catalysts could be well

described with a combination of nonlinear equations with only

a few molecular parameters. Nevertheless, we encountered a

number of questions that could not be directly addressed in a

quantitative manner with the initial data set. For example,

throughout the study we had little to no indication that the

actual reactivity of the quaternary ammonium·substrate ion pair

was needed to quantitatively describe the catalytic activities.

Instead, the results were interpreted entirely in the context of

partition rates of the quaternary ammonium ion pairs (i.e.,

interfacial transport phenomena). As a corollary, the QSAR

modeling results indicated that the catalytic activity of

tetraalkylammonium ions in reactions with more acidic

substrates could potentially be well described by a nonlinear

INTRODUCTION

■

Phase transfer catalysis (PTC) embodies the ideals of organic

synthesis by providing a simple, cheap, and general protocol for

a myriad of reactions.¹Of the numerous types of reactions that

are amenable to PTC, the most synthetically useful are those

that make C−N, C−O, and C−C bonds.²Speciﬁcally, aliphatic

nucleophilic substitution reactions employing nitrogen, oxygen,

or carbon nucleophiles in combination with carbon electro-

philes constitute the majority of the most synthetically useful

PTC reactions. The preparative advantages of PTC for such

alkylation reactions derive from its operational simplicity, ease

of removal of byproducts, and enhanced reaction rates. In

addition, PTC facilitates the use of inexpensive hydroxide bases

and obviates the need for dipolar aprotic solvents. In fact, under

PTC conditions, comparable rates are observed with little or no

solvent in combination with the addition of a small amount

(1−5 mol %) of a quaternary ammonium ion phase transfer

catalyst.³

The preparative advantages of PTC are accompanied by

mechanistic ambiguity that arises from the transport

component and non-covalent ion pairing that make PTC a

diﬃcult system to study. The catalytic cycle of PTC necessarily

involves at least one physical phenomenon (or step),

distinguishing it from homogeneous catalysis wherein the

entire catalytic cycle is comprised of chemical steps.

Consequently, the selection of the optimum catalyst for a

particular application remains a time-consuming, largely

empirical exercise. The relative contributions of transport,

Received: May 17, 2012

Published: August 2, 2012

13415

dx.doi.org/10.1021/ja304808u | J. Am. Chem. Soc. 2012, 134, 13415−13429

Journal of the American Chemical Society

Article

QSAR model with a single molecular parameter, namely, the

thermodynamic partition coeﬃcient of the catalyst.

phenol alkylation operating under the extraction mechanism are

described herein.¹⁰

Thus, a new study was conceived to establish whether the

catalytic activity of quaternary ammonium ions was related

more to their stoichiometric rate constants or phase transfer

rate constants (Scheme 1). The alkylation of phenol was

identiﬁed as a reaction that would allow for this question to be

addressed in a quantitative manner. Described herein is a

comparison of the stoichiometric, catalytic, and phase transfer

rate constants for a homologous series of quaternary

ammonium ion phenolates along with an analysis of their

partition equilibria and the composition of the species present

in the organic phase.

2. Mechanisms of PTC.¹¹2.1. Extraction Mechanism of

PTC. In the simplest PTC process, a water-soluble nucleophile

such as sodium cyanide reacts by the mechanism outlined in

Figure 1.¹²The observed catalytic eﬀect is the physical

extraction (or transfer) of the anionic nucleophile to the

organic phase (K₂) and thus is termed the “extraction

mechanism”. In the extraction mechanism, the substrate

anion (NC⁻) and product anions (X⁻) are formally distributed

between the aqueous and organic phases as quaternary

ammonium (Q⁺) ion pairs. The irreversible, bond-forming

reaction (k₃) occurs in the organic phase and is the rate-

determining step.^12,13Lipophilic catalysts tend to perform

better than hydrophilic ones. Some modiﬁcations have been

proposed to account for the activity of catalysts that have no

appreciable aqueous solubility,¹⁴but in general, this mecha-

nistic outline has been supported by numerous subsequent

investigations.¹⁵

Scheme 1

1. Selection of Reaction To Study. The reasons that the

alkylation of phenol is the appropriate reaction for this study

are not obvious and require explanation. The identiﬁcation and

analysis of catalyst structure−activity relationships for PTC

reactions are complicated by the fact that two mechanisms can

be operative, extraction and interfacial. Table 1 contains three

representative alkylation reactions arranged in order of pK_aof

the substrate (lowest to highest), from cyanide (pK_a= 9.4)⁵to

phenol (pK_a= 18.0)⁶to an α-aryl nitrile (pK_a> 18).⁷As the

pK_aof the substrate increases, the dominant operative

mechanism shifts from an extraction mechanism where the

bond-forming reaction occurs in the organic phase, to an

interfacial mechanism where the bond-forming reaction occurs

from an interfacially adsorbed ion pair.^8,9

Figure 1. Extraction mechanism of phase transfer catalysis.

2.2. Hydroxide-Initiated PTC. The second example in Table

1 is the alkylation of phenol (pK_a= 18) under biphasic

conditions in the presence of an aqueous hydroxide base. The

use of a neutral substrate in combination with an aqueous base

under PTC conditions is termed “hydroxide-initiated PTC”.¹⁶

The mechanism of hydroxide-initiated PTC reactions is more

complex than the extraction mechanism and involves

distinguishing (1) where the substrate is deprotonated, (2)

whether the catalyst is involved in the deprotonation step, (3)

where the substrate·ammonium ion pair is generated, and (4)

how the substrate·ammonium ion pair is transferred to the

organic phase.^8b,9,15b,17Typically, an in situ deprotonation only

adds pre-equilibrium steps to the catalytic cycle (K₁and K₂,

Figure 2), and the PTC alkylation of phenols comports with

this model.^18,19As expected, the deprotonation step (k₂) is fast

since the pK_aof phenol is much less than that of water (pK_a=

Table 1. Representative Extraction and Interfacial Phase

Transfer Catalyzed Reactions

To systematically probe catalyst structure−activity relation-

ships for reactions following either an extraction or an

interfacial mechanism, it would be most convenient to study

a single reaction where the dominant mechanism can be

aﬀected by changing the reaction conditions. The alkylation of

phenol uniquely fulﬁlls this requirement. For this reason, the

alkylation of phenol was identiﬁed as a reaction for which a

comparison of the stoichiometric and catalytic rate constants as

a function of a homologous series of ammonium counterions

could serve to bridge the gap of understanding of the two

mechanistic regimes of PTC. Our results and analysis for a

Figure 2. Schematic representation of the hydroxide-initiated PTC

alkylation of phenols.

13416

dx.doi.org/10.1021/ja304808u | J. Am. Chem. Soc. 2012, 134, 13415−13429

Journal of the American Chemical Society

Article

Figure 3. Schematic representation of the interfacial mechanism(s) of phase transfer catalysis: (a) the Makosza interfacial mechanism and (b) the

modiﬁed interfacial mechanism.

32).²⁰However, considerable evidence has been accumulated

indicating that “other complicating factors” may become

important in the PTC alkylation of phenols, depending on

reaction conditions.²¹For example, the formation of a “third

liquid phase” (TLP) at the aqueous/organic interface (K₁′,

Figure 2) has been proposed.^21dOne study reports the

purposeful engineering of a TLP (termed “liquid−liquid−liquid

PTC”) by the addition of sodium chloride (25% by mass) to

the aqueous phase.²²In this way, a stable, reproducible TLP is

generated with the relative volumes of organic/TLP/aqueous =

50/4/55 cm³. However, the composition of the TLP during the

reaction and consequent mechanistic implications remain

largely unknown.

Because of the ambiguity about the exact nature of the space

between the organic and aqueous phases under PTC

conditions, the terms “interfacial space” and “interphase” will

be used throughout this disquisition. To summarize, the

alkylation of phenol is suﬃciently well-behaved to be explained

by the extraction mechanism, but some subtle characteristics of

this reaction class can arise from interfacial adsorption/

desorption equilibria (K₁′, Figure 2). In the present study,

the reaction conditions were chosen to ensure PTC catalysis

under the extraction mechanism.

hydroxide (Q⁺OH⁻), which acts as the base that actively

deprotonates the substrate.

BACKGROUND

■

1. Analyses of Catalyst Structure−Activity Relation-

ships for Extraction PTC Reactions. One of the most

thorough studies in support of the extraction mechanism of

phase transfer catalysis compared the catalytic activity of 13

straight-chain quaternary ammonium and phosphonium ions

under stoichiometric, homogeneous (1.0 equiv) and catalytic,

heterogeneous (0.1 equiv) phase transfer conditions.^14dIn that

study, comparison was made between the rates of displacement

of methyl octyl sulfonate with small, hydrophilic, weakly basic

−

nucleophiles including Cl⁻, Br⁻, I⁻, SCN⁻, N₃, and CN⁻

(Scheme 2, top).

Scheme 2

2.3. Interfacial Mechanism. The third example from Table 1

is the hydroxide-initiated PTC alkylation of an α-aryl nitrile (18

< pK_a< 25), which is meant to be representative of enolate

alkylations.^23a,bThe inability to explain the behavior of

hydroxide-initiated PTC reactions of this type by the extraction

mechanism led Makosza to propose an interfacial mechanism

(Figure 3a).^23cIn the initially proposed interfacial mechanism,

the substrate (S−H) and quaternary ammonium catalyst are in

equilibrium between the organic phase and the interfacial

region. Similarly, the inorganic base (M⁺OH⁻) is in equilibrium

with the aqueous phase and the interfacial region. The

inorganic base facilitates deprotonation of the substrate,

which may then exchange with the quaternary ammonium

(Q⁺) to generate the key active species (S⁻Q⁺). The

substrate·ammonium ion pair then dissociates from the

interface into the organic phase where it may react. Subsequent

kinetic studies have shown that quaternary ammonium ions can

inﬂuence the rate of substrate deprotonation,²⁴which provides

support for a second, modiﬁed interfacial mechanism (Figure

3b).⁹The major modiﬁcation is that the quaternary ammonium

halide (Q⁺X⁻) is in equilibrium with the quaternary ammonium

Under PTC conditions the observed rate constants diﬀered

by up to 2 orders of magnitude between diﬀerent ammonium

ion catalysts. Under homogeneous conditions the observed rate

constants diﬀered by only a factor of 2.5. Some attempts have

been made to conduct similar studies with more basic

substrates. For example, the catalytic activities of 13 quaternary

ammonium bromides have been determined for the alkylation

of thiophenoxide (Scheme 2, bottom).^14cUnfortunately,

attempts to compare the stoichiometric rate of reaction to

those under PTC conditions were complicated by irreprodu-

cibility of the alkylation of tetrabutylammonium thiophenoxide

in homogeneous media.

In those studies the catalyst activity increased linearly with

lipophilicity (log P), but no optimum was identiﬁed. Both

studies concluded that “the eﬀectiveness of a phase transfer

catalyst depends mainly on its organophilicity, with other

structural factors [being] much less important.” ^14cThus, the

13417

dx.doi.org/10.1021/ja304808u | J. Am. Chem. Soc. 2012, 134, 13415−13429

Journal of the American Chemical Society

Article

dominant eﬀect of the catalyst is to increase the eﬀective

concentration of the nucleophile in the organic phase. Given

the range of substrates studied thus far, it seems plausible that

for any PTC reaction operating under the extraction

mechanism, the same dominant eﬀect of the catalyst will be

observed.

were included.³³The most recent investigations focus on the

generation of a TLP because when a TLP does form it is

accompanied by a signiﬁcant rate enhancement.^34,35Depending

on the conditions, up to 3 orders of magnitude in rate

enhancement accompanies the formation of a TLP, which is

highly dependent on the organic solvent and ionic strength of

the aqueous phase, but the eﬀect of the quaternary ammonium

ion structure has not been extensively investigated. Thus, no

data are available that are suitable for application of QSAR

methods.

4. Solid-State and Solution Structural Data Relevant

to PTC Alkylation of Phenols. The study of the solid-state

structure and solution-state reactivity of phenols and

phenoxides is extensive.³⁶In the solid state, most alkali metal

phenoxides form extended aggregates with multiple bridging

(O−M−O)_nlinkages where the number of oxygen−metal

contacts varies from 2 to 4 (μ₂−μ₄).³⁷Characterization by

single-crystal X-ray diﬀraction is rare, and most of the solid-

state structural data for phenolates are derived from X-ray

powder diﬀraction data.³⁸A few exceptions are known wherein

cation complexing agents, such as 18-crown-6,³⁹or sterically

bulky groups have been incorporated⁴⁰to facilitate the

collection of single-crystal X-ray diﬀraction data.

Numerous studies on the pK_aof phenol derivatives have

been conducted.⁴¹The eﬀect of substitution is enhanced in

dipolar aprotic solvents (ρ = 5.3, DMSO) compared to protic

media (ρ = 1, H₂O).⁶A complicating factor in studying the

reactivity diﬀerences of phenol derivatives is the formation of

dimeric species in solution (Scheme 4). The dimeric species,

termed homo-hydrogen-bonded complexes, are signiﬁcantly

slower to deprotonate than the parent, uncharged phenol

derivatives. Thus during the course of a titration, the acidity of

the dominant species in solution can change.⁶

By contrast, in our previous reports on interfacial PTC,⁴as

much as a 3.8 order of magnitude rate enhancement was

attributed to eﬀective concentration, but the correlation was not

linearly related to lipophilicity (ClogP). This study is designed to

determine the range of catalyst activity over which a linear

relationship with catalyst log P holds, and whether an optimum

catalyst lipophilicity can be identiﬁed.

2. Structure−Reactivity Relationships of Phenolates.

The stoichiometric reactivity of phenoxides has been studied

extensively.²⁵In fact, recognition of the relationship between

the acidity and reactivity of substituted phenols is partially

responsible for the original concept of the free energy

relationship.⁶Alkylation rates of substituted sodium and

potassium phenoxides increase when the pendant groups are

electron donating (ρ = −1.1).²⁶These data are consistent with

the intuitive notion that, for ion pairs, the greater the charge

separation, the more reactive the anionic species will be, and

that the charge separation is greater for ammonium phenoxides

than for alkali metal phenoxides.²⁷The relative reactivities of

tetrabutylammonium and potassium phenoxide have been

compared by determining the ﬁrst-order rate constants of

alkylation with 1-bromobutane in multiple solvents (DMF,

CH₃CN, dioxane).²⁸A decrease of 10³in rate constant is

observed upon changing from dimethylformamide to acetoni-

trile to dioxane for the alkylation of potassium phenoxide. In

contrast, the rates of alkylation of tetrabutylammonium

phenoxide, under the same series of reaction conditions, varied

by only a factor of 6. Interpolation and molecular modeling

have expanded upon these observations. It has been suggested

that the eﬀective ionic radii of quaternary ammonium ions

increases in the homologous series Me₄N⁺≈ 2.85 Å, Et₄N⁺≈

3.48 Å, (n-Pr)₄N⁺≈ 3.98 Å, (n-Bu)₄N⁺≈ 4.37 Å, after which

point there is little or no change.^29,30Not enough data are

available to ascertain the certainty of this proposal or how

relevant it is to PTC reactions with substoichiometric amounts

of quaternary ammonium ion catalysts. However, the analysis is

consistent with many of the stoichiometric extractive alkylation

Scheme 4

The decreased acidity of homo-hydrogen-bonded phenolate

complexes compared to phenols has been exploited in the study

of the kinetics of proton transfer reactions. Nielsen and

Hammerich described the preparation of the homo-hydrogen-

bonded complex Bu₄N⁺[PhOHOPh]⁻(Scheme 5).⁴²Reetz and

Goddard modiﬁed the procedure for the preparation of X-ray-

quality crystals of the same complex.⁴³

The solid-state X-ray diﬀraction data show that the closest

contact between the ammonium ion α-CH atoms and the

anionic phenoxide complex is 2.73 Å, and the average distance

studies by Brandstrom, wherein no catalyst turnover is

̈

required.³¹

3. Phase Transfer Catalyzed Alkylation of Phenols.

The initial report on hydroxide-initiated PTC alkylations of

phenol revealed that a wide range of phenol derivatives and

electrophiles (alkyl halides and sulfonate esters) could be

employed (Scheme 3).¹⁹

Nearly all of the reported extensions of this preliminary study

focus on preparative aspects³²but also reconﬁrm the ﬁrst-order

kinetic behavior in both reacting components as well as the

catalyst.¹⁸A tour de force application of kinetic modeling to the

study of phenol alkylations has yielded a bounty of insights, but

only tetrabutylammonium or benzyltributylammonium cations

Scheme 5

Scheme 3

13418

dx.doi.org/10.1021/ja304808u | J. Am. Chem. Soc. 2012, 134, 13415−13429

Journal of the American Chemical Society

Article

is 3.02 Å. These data are in good agreement with the hypothesis

that ammonium cations generate ion pairs that are signiﬁcantly

more separated than alkali metal cations, where analogous bond

lengths are signiﬁcantly shorter (Na, ∼2.3 Å; K, ∼2.7 Å).⁴⁴

5. Computational Studies on the Structure and

Reactivity of Ammonium Salts. The scarcity of structural

data for comparison of alkali metal and ammonium ion pairs

has provided motivation for a number of computational

studies.⁴⁵These studies show that the calculated cation−

anion interaction energy is weaker for tetrabutylammonium ion

phenolates (2.5 kcal/mol) than for alkali metal cations (Na⁺,

4.5 kcal/mol).⁴⁶

Recently, two types of computational modeling techniques

have been applied to the investigation of nucleophile reactivity

under catalytic PTC conditions, namely, ab initio calculations

employing a polarizable continuum solvent model (PCM),⁴⁷

and molecular dynamics simulations employing an ensemble of

solvent molecules (10−100 molecules).⁴⁸Molecular dynamic

simulations suggest that the rate of displacement in a model

nucleophilic substitution reaction is highly dependent on the

degree of hydration of the nucleophile, which is in turn

dependent on where in the interfacial region the reaction takes

place. Within 5 Å of the Gibbs surface into the organic phase,

the hydration level is still suﬃcient to mimic the reaction in

bulk water. However, beyond this distance the activation barrier

continuously decreases as the hydration level decreases. The

conclusion is that the role of the PTC is to transport the

nucleophile deep into the organic phase such that desolvation

leads to increased nucleophilic reactivity.⁴⁸

Similar conclusions have been reached independently by

comparison of association constants of quaternary ammonium

sulfates in solution, both computationally and experimentally.⁴⁹

The computational results suggest that the diﬀerences in

nucleophilicity of quaternary ammonium ion pairs result from a

combination of factors related to solvation, including (1) a

decrease in electrostatic interactions between the ions, (2)

disruption of dispersion interactions between the anion and the

solvent, and (3) generation of solvent cavities in the vicinity of

the nucleophile. That is, whereas a greater charge separation

may accompany ion exchange from alkali metal to ammonium

counterion, the origin of catalytic activity, if any, is more likely a

consequence of shielding the nucleophile from protic media, a

transition state lowering eﬀect.

phenolate itself, or to separate the alkali bromide salt from the

reaction mixture. The alkali halide contaminant is known to

inﬂuence the rates of phase transfer catalyzed phenol

alkylations.^22,50Thus, some ambiguity remains about the

previous kinetic analyses and mandates that samples of

known purity be employed in this study.

1.1. Preparation of Tetraalkylammonium Hydroxides. On

the basis of our previous results and known data on PTC

cyanide alkylations, it was decided to limit the survey of

quaternary ammonium salts to R = Me, Et, n-Pr, n-Bu, n-Hex,

and n-Oct. To prepare the phenoxides in pure form required

the preparation of the corresponding ammonium hydroxides in

pure form, as no subsequent puriﬁcation would be possible.

The requisite quaternary ammonium hydroxides were prepared

from their bromides by ion exchange, closely following the

protocol of Harlow, Noble, and Wyld, with the exception that

anhydrous methanol was used as the solvent rather than

isopropanol (Table 2).⁵¹Multiple passes (3−6) through an

Amberlyst A-26 resin were required to fully exchange the halide

for hydroxide.⁵²After each pass, the hydroxide form was

regenerated by rinsing the column with a large excess of sodium

hydroxide. Sodium hydroxide was chosen because any residual

sodium is expected to be less detrimental to the kinetic analyses

than potassium. The resulting quaternary ammonium hydrox-

ide solutions were titrated for total base to a phenolphthalein

end point (3−6 determinations) and titrated for residual halide

with an ion-selective electrode (2−3 determinations).⁵³The

results for the preparation of the quaternary ammonium

hydroxide solutions are shown in Table 2. The ﬁnal solutions

were diluted (or concentrated) such that the total base

concentration was between 0.1 and 0.7 M and stored as such.

The extent of exchange was greater than 97% in all cases, and

the salts could be stored at −20 °C for an extended period (∼1

month) with no detectable change in total base titer.tbl1

Table 2. Preparation of Quaternary Ammonium Hydroxides

in Methanol

ammonium

cation

OH conc,

Br,

wt %

Br,

d

a

b

entry

[M]

std dev

mmol %

6. Objectives of This Study. Thus, it is still unresolved

what the nucleophile activation and nucleophile extraction

contribute to the net catalytic activity of phase transfer catalysts.

The central aim of this study is to determine which relationship

best describes phase transfer catalyst activity, a rate−rate

(nucleophile activation) linear free energy relationship (LFER)

or a rate−equilibrium (nucleophile extraction) LFER. The

method chosen was to compare the catalytic rate constant

(k_activity) to the stoichiometric rate constant (k_reactivity) under

conditions that are as analogous as possible.

1

2

3

4

5

6

methyl₄N⁺

ethyl₄N⁺

n-propyl₄N⁺

n-butyl₄N⁺

n-hexyl₄N⁺

n-octyl₄N⁺

0.127

0.668

0.229

0.588

0.685

0.426

0.0095

0.0014

0.0050

0.0054

0.0096

0.0013

0.16

0.33

1.56

0.61

0.34

1.82

c

<0.1

<0.10

0.56

0.31

0.41

1.19

a

Total base titrations are an average of 3−6 determinations to a

phenolphthalein end point. Bromide weight percentage is an average

of 2−3 determinations by ion-selective electrode. Prepared from the

quaternary ammonium iodide, and no residual iodide could be

detected. The total mmol of anions is taken as the mmol of hydroxide

plus the mmol of bromide.

b

c

d

RESULTS

■

1. Stoichiometric Reactivity of Ammonium Pheno-

lates (Rate−Rate LFER). The initial challenge in this project

was to identify a general procedure for the preparation of a

homologous series of quaternary ammonium phenolates in

homogeneous and halide-free form. In previous studies,

tetrabutylammonium phenolates were prepared by mixing

tetrabutylammonium bromide and sodium (or potassium)

phenoxide.²⁸No attempt was made to purify the ammonium

1.2. Preparation of Tetraalkylammonium Phenolates. A

number of orienting experiments were carried out to generate

monomeric ammonium phenolates of the type R₄N⁺PhO⁻, but

none were successful. Preliminary tests for stability with the

tetraethylammonium counterion revealed that Et₄N⁺PhO⁻,

once generated in anhydrous form, was stable as a solid for

only ∼8−10 h in a glovebox. Thus, to enable the study of the

13419

dx.doi.org/10.1021/ja304808u | J. Am. Chem. Soc. 2012, 134, 13415−13429

Journal of the American Chemical Society

Article

stoichiometric reactivity of ammonium phenolate complexes,

we resorted to the preparation and evaluation of homo-

hydrogen-bonded complexes.^42,43

phenoxide substrate, sodium phenoxide in DIPK (rate for

Na⁺[(PhOH)OPh]⁻= 5.09 × 10⁻⁸mol L⁻¹s⁻¹).⁵⁸

A variety of solvents were initially surveyed for the formation

of tetraalkylammonium phenolate·phenol complexes including

hydrocarbons (hexanes, cyclohexane), aromatic hydrocarbons

(benzene, toluene), halogenated (dichloromethane, dichloro-

ethane), polar aprotic (THF, dioxane, acetonitrile), and polar

protic (methanol, ethanol, isopropyl alcohol) solvents.

Ultimately, dioxane proved superior because of several

advantageous physical properties. First, its relative high melting

point (11−12 °C) allowed for the tetraalkylammonium

phenolate solutions to be conveniently stored in a frozen

dioxane matrix. Although prolonged storage was never

attempted, no signs of decomposition were observed after

1

storage overnight by H NMR analysis. Second, dioxane forms

substantive azeotropes with water (dioxane/water, 49/51)⁵⁴

and methanol (dioxane/methanol, 77/23),⁵⁵as well as ternary

mixtures,⁵⁶which allowed for removal of methanol from the

salts.

Thus, the methanolic, quaternary ammonium hydroxide

solutions were combined with dioxane solutions of phenol to

aﬀord quaternary ammonium phenoxides as homo-hydrogen-

bonded complexes, R₄N⁺[(PhOH)OPh]⁻. Three azeotropic

concentrations from dioxane aﬀorded the tetraalkylammonium

phenolate complexes (R₄N⁺[(PhOH)OPh]⁻) as ﬁnely pow-

dered solids that could be handled in a glovebox. The

stoichiometry of the ammonium phenoxide complexes was

determined by quenching a small sample into an acetonitrile

solution containing a large excess of benzyl bromide and a

standard (Table 3). After stirring the alkylation overnight, the

relative amounts of phenol and benzyl phenyl ether were

determined by gas chromatography (GC). In all cases the

stoichiometry was found to be 1:1 ( 0.04).

Figure 4. Tetraalkylammonium phenoxide alkylation rates (stoichio-

metric).

The reactions were monitored to typically 5−10%

conversion, and reaction rate is expressed as the mean of

triplicate reactions in terms of the change in n-butyl phenyl

ether concentration (mol/L) with respect to time (s).⁵⁹The

error bars in the summary plot represent one standard

deviation in reaction rate. Reactions performed in acetonitrile

were between a factor of 1.5 and 5 faster than those in DIPK

for all of the quaternary ammonium cations examined. Overall,

the reaction rates were more reproducible in DIPK than in

acetonitrile, perhaps because of the diﬃculty of removing all

adventitious water from the acetonitrile.

The rates of alkylation of the various phenoxide salts

exhibited a modest solvent dependence; DIPK displayed a

wider range in rate (a factor of 67.0) compared to acetonitrile

(a factor of 18.6). Closer inspection of the data reveals that the

greatest diﬀerence in rates was seen for the alkali metal (Na⁺

and K⁺) or Me₄N⁺cations. For acetonitrile, the alkylation rate

for the potassium phenoxide (K⁺[(PhOH)OPh]⁻) was 9-fold

faster than that for the sodium phenoxide (Na⁺[(PhOH)-

OPh]⁻), whereas in DIPK, the rate of alkylation of the

potassium salt was only 2-fold greater that that of the sodium

salt. Curiously, the alkylation rate of K⁺[(PhOH)OPh]⁻was

1.2-fold faster than that of Me₄N⁺[(PhOH)OPh]⁻in

acetonitrile, whereas it was 4-fold slower than the rate for

Me₄N⁺[(PhOH)OPh]⁻in DIPK.

Table 3. Preparation of Quaternary Ammonium Phenoxides

as Hydrogen-Bonded Dimers

PhOH/PhO⁻,

mol/mol

entry

R

1

2

3

4

5

6

methyl

ethyl

1.04

0.98

1.02

1.00

0.94

0.98

n-propyl

n-butyl

n-hexyl

n-octyl

Within the series of homologous tetraalkylammonium salts,

the range in rate was remarkably small. In acetonitrile the rates

varied by a factor of only 2.4; the highest reaction rate was

observed for (n-Pr)₄N⁺. In DIPK, the rates varied by a factor of

6.8, with the highest rate observed for (n-Bu)₄N⁺. The largest

rate eﬀect was the diﬀerence in reactivity between Me₄N⁺and

Et₄N⁺; in acetonitrile the diﬀerence was small (1.7-fold faster

for Et₄N⁺), whereas in DIPK a larger diﬀerence was observed in

(4.0-fold faster for Et₄N⁺).

This ﬁnding suggests that, for phenoxide alkylations, the

eﬀective ionic radius of quaternary ammonium cations is not

highly variable. The stoichiometric reactivity of the surveyed

quaternary ammonium cations had a poor correlation to their

1.3. Stoichiometric Reactivity of Ammonium Phenolates.

The reactivity of the tetraalkylammonium phenoxides was

determined by employing conditions directly analogous to the

most comprehensive set of data available, namely, reaction with

a slight excess of 1-bromobutane in acetonitrile at 0.016 M.²⁸

The least polar, aprotic solvent that was tested and could

completely dissolve the tetraalkylammonium phenoxides was

diisopropyl ketone (DIPK).⁵⁷Initial-rate kinetic data were

collected for each ammonium phenoxide as well as the

corresponding sodium and potassium homo-hydrogen-bonded

dimers in each solvent (Figure 4). The data from these

alkylation reactions have been normalized to the slowest

13420

dx.doi.org/10.1021/ja304808u | J. Am. Chem. Soc. 2012, 134, 13415−13429

Journal of the American Chemical Society

Article

cationic radii in the respective tetraalkylammonium phenoxide

ion pair.

whereas under PTC conditions the catalytic activity did not

vary for catalysts larger than TBAB. The catalyst survey under

PTC reaction conditions clearly demonstrates that the intrinsic

reactivity of R₄N⁺[(PhOH)OPh]⁻has little eﬀect on the

observed catalytic activity.

3. Conﬁrmation of Extraction Mechanistic Regime.

3.1. Order in n-BuBr. Although the PTC alkylation of phenols

is known to operate by an extraction mechanism,^18bwe are

unaware of a case in which such a clearly nonlinear dependence

of catalytic activity and catalyst lipophilicity has been

reported.⁶¹A break in an LFER-type plot (see Discussion for

detailed LFER/QSAR analysis) has two common interpreta-

tions, namely, a change in rate-determining step or a change in

mechanism.⁶²The clearly nonlinear relationship observed

herein prompted us to test if either of these interpretations

was the origin of the observed nonlinear dependence of

catalytic activity on catalyst lipophilicity.

2. Catalytic Activity of Ammonium Phenoxides under

Phase Transfer Conditions. The catalytic activity of the same

series of quaternary ammonium cations was evaluated under

directly analogous conditions (40 °C, DIPK, 0.27 M in n-BuBr)

starting from the quaternary ammonium bromide. Several

aqueous base concentrations were tested (10 M and 50% (w/

w) aqueous NaOH and KOH), and it was observed that 5 M

aqueous KOH aﬀorded the most reproducible results. Both

phases appeared homogeneous prior to and after the addition

of n-butyl bromide at ambient (∼20 °C) and elevated

temperatures (40 °C). The initial rate for the formation of n-

butyl phenyl ether was monitored by GC analysis.⁵⁹

Under these reactions conditions, the background reaction

was very slow; only 5% conversion to product was observed

after 5 days (Figure 5).⁶⁰All of the catalysts were evaluated at

0.05 equiv loadings and in addition at 0.10 equiv loadings for

(n-Pr)₄N⁺Br⁻(TPAB), (n-Bu)₄N⁺Br⁻(TBAB), (n-Hex)₄N⁺Br⁻

(THAB), and (n-Oct)₄N⁺Br⁻(TOAB).

The possibility of a change in reaction mechanism and rate-

determining step was excluded by determination of the reaction

order in electrophile. The order in n-butyl bromide was

determined by the method of initial rate kinetics for two

catalysts (TPAB and TOAB), one on either side of the break in

the LFER (Figure 6). Plotting the resultant initial rates versus

Figure 5. Catalyst survey for PTC phenol alkylation.

Tetramethylammonium bromide (TMAB) was found to be

an inactive catalyst at 0.05 equiv loading and aﬀorded an

alkylation rate no greater than the background rate. At the same

catalyst loading, tetraethylammonium bromide (TEAB) was

able to catalyze the reaction 4.6 times faster than background,

and TPAB catalyzed the reaction 167 times faster than

background. Another signiﬁcant jump in catalytic activity was

observed with TBAB, which aﬀorded a 663-fold rate increase

over background, the maximum rate observed for the series.

For salts more lipophilic than TBAB, only small changes in

catalytic activity were seen at 0.05 equiv. Virtually no diﬀerence

was seen among these salts at 0.10 equiv loading as well (Figure

5).

The range in catalytic activity among the homologous

tetraalkylammonium salts (a factor of 663) was approximately

100-fold greater than the range in reactivity observed for

R₄N⁺[(PhOH)OPh]⁻under homogeneous reaction conditions

(a factor of 6.8). A relatively large diﬀerence in catalytic activity

was seen with TEAB and TPAB under PTC reaction conditions

(25.9-fold) in contrast to the homogeneous reaction conditions

that displayed similar reactivities with the same cations (1.5-

fold). As was the case in the reactions of the tetraalkylammo-

nium phenoxides under homogeneous reaction conditions in

DIPK, the maximum rate was observed with the (n-Bu)₄N⁺

cation. However, in DIPK under homogeneous reaction

conditions, the reactivity decreased slightly after TBAB,

Figure 6. Order in n-butyl bromide for PTC alkylation of phenol.

the concentration of n-butyl bromide (0.5, 1.0, 2.0, and 5.0

equiv with respect to phenol) revealed a linear relationship for

both catalysts. This ﬁnding is indicative of a pseudo-ﬁrst-order

rate dependence in electrophile under PTC conditions,

conﬁrming that the rate-determining step is the bond-forming

reaction (k₄, Figure 2).

If a change in reaction mechanism were operative (i.e., a

change from intrinsic rate to transport rate limiting), then a

diﬀerent order in electrophile would be observed for the two

catalysts. If one of the catalysts led to a transport rate-limited

PTC reaction, then a zeroth-order dependence on n-butyl

bromide would be expected. We conclude therefore that a

change in rate-determining step or mechanism is not responsible

for the diﬀerences in catalytic activity between the ammonium

ions. The slope of the linear dependence is strongly dependent

on the catalyst employed, indicating that although the catalytic

rate is highly dependent on the catalyst, the rate diﬀerences do

13421

dx.doi.org/10.1021/ja304808u | J. Am. Chem. Soc. 2012, 134, 13415−13429

Journal of the American Chemical Society

Article

not arise from changes in the rate-determining step, but rather

from diﬀerences in the pre-equilibria (k₁, k₂, or k₃, Figure 2).

3.2. Order in Catalyst. Carrying out the alkylation with 0.10

equiv instead of 0.05 equiv of TPAB aﬀorded a 2-fold increase

in alkylation rate, consistent with an intrinsic rate-limited PTC

reaction that is ﬁrst order in catalyst. However, a similar

doubling of the catalyst loading for TBAB, THAB, and TOAB

led to increases in alkylation rate by factors of 2.6, 3.0, and 2.9,

respectively. These increases in alkylation rates suggest the

presence of higher order phenoxide aggregates in the organic

phase (vide infra). The emergence of an interfacial mechanism

may also be responsible for the alkylation rate augmentation.

3.3. Eﬀect of Stirring Speed on Alkylation Rate. To

determine whether an extraction mechanism or interfacial

mechanism was operative, a variable stirring speed experiment

was conducted with TBAB as the catalyst.^9,63The PTC reaction

had to be run at ambient temperature because of the design of

the electric motor. The stirring speed was varied from 500 to

2000 rpm, and no dependence of the reaction rate on the

stirring speed was observed above 500 rpm (Figure 7).

Therefore, this result supports an extraction mechanism and

an intrinsic reaction rate-limited PTC reaction.

atically, this too could be the origin of the diﬀerences in

catalytic activity of the ammonium ions.

A priori, the phase partition equilibrium (k₃) seemed to be

the most logical source of the trend because partition

coeﬃcients, P (in this case P = K₃= k₃/k₋₃), are known to

be directly proportional to the number of carbons in the

partitioning,⁶⁴while many known ionic equilibria of ammo-

nium ions are known not to be highly dependent on the

ammonium ion.²⁷Nonetheless, we sought to collect a set of

data to (1) show that the amount of extracted ammonium

phenolate increased along the homologous series of ammonium

ions and (2) analyze the composition and stoichiometry of the

extracted ammonium phenolate ion pairs. Several experiments

were conducted to quantify the amount of this reactive species

in the organic phase.

4.1. UV−Vis Spectroscopic Analysis of the Aqueous Phase.

The ﬁrst series of experiments is designed to analyze the pre-

equilibrium partitioning of phenol/phenoxide (k₄, Figure 2),

which could readily be analyzed by UV−vis spectroscopy. The

greatest shortcoming of this analysis is that the concentration in

the organic phase had to be determined by diﬀerence, because

the DIPK (organic phase solvent) is UV-active and its

absorption is coincident with the UV absorption of phenoxide.

Therefore, in this study, the concentrations of phenol/

phenoxide in the aqueous phase were used to calculate the

amount of phenoxide transferred to the organic phase. The

same tetraalkylammonium bromides were analyzed by this

method for the amount of phenoxide transferred to the organic

phase under catalytic conditions employing substoichiometric

(0.2 equiv) and stoichiometric (1.0 equiv) amounts of the

ammonium salts.

The results from this series of experiments (Figure 8) show

that the more lipophilic ammonium ions extracted more of the

Figure 7. Variable stirring speed under PTC reaction conditions.

All of the data collect thus far are consistent with those found

in previous studies and support an intrinsic reaction rate-limited

PTC process operating by an extraction mechanism. In the

context of a QSAR/LFER analysis of catalysis, the data are

most consistent with a third, less common rationale for a break

in a LFER plot, namely, a change in a pre-equilibrium position

and no change in rate-determining step or mechanism.

4. Analysis of Composition of the Organic-Phase

Ammonium Phenolate Ion Pair(s). With a change in

reaction mechanism ruled out as the origin of the plateau in

catalytic activity as a function of ammonium ion size, two pre-

equilibria remain wherein the catalyst is necessarily involved,

namely, anion exchange (k₃, Figure 2) and equilibration of the

ammonium phenolate ion pair between the aqueous and

organic phases (k₄, Figure 2). As a corollary, if the composition

of the extracted ammonium phenoxide species varied system-

Figure 8. Composition of phenol in the organic phase by UV−vis

analysis.

phenol/phenoxide to the organic phase, as expected. With 1.0

equiv of the tetraalkylammonium bromides, the more lipophilic

salts, TPAB to TOAB, transferred a large amount of phenoxide

to the organic phase, whereas the less lipophilic salts, TMAB

and TEAB, transferred only a small amount of phenoxide.

13422

dx.doi.org/10.1021/ja304808u | J. Am. Chem. Soc. 2012, 134, 13415−13429

Journal of the American Chemical Society

Article

Similarly, when 0.2 equiv of the ammonium salts was employed,

the same trend was observed with the exception that the

relative amount of phenol extracted by TPAB decreased

signiﬁcantly.

Notably, when a stoichiometric amount of TPAB was

employed, an amber-colored TLP was observable between

the aqueous and organic phases (Figure 9). Analysis of the TLP

Figure 10. Composition of phenol in the organic phase by titration.

phenol was found than phenoxide. However, the more

lipophilic ammonium salts transferred stoichiometric (TBAB,

9.35%) and superstoichiometric (THAB, 14.13%; TOAB,

16.77%) amounts of phenoxide to the organic phase. Moreover,

the increasing amount of phenoxide transferred to the organic

phase with increasing alkyl chain length is accompanied by a

corresponding decrease in the amount of neutral phenol

transferred. Thus, the total amount of phenol transferred (black

squares) roughly corresponds to the results obtained from the

UV−vis spectroscopic analysis, now with ﬁner resolution of the

constituent components.

The presence of phenol in the organic phase was unexpected

because the use of concentrated aqueous potassium hydroxide

should heavily favor (by several pK_aunits) the formation of

potassium phenoxide at equilibrium. Even more interesting was

that the concentration of organic-phase phenol was unique to

each ammonium salt examined, and therefore not simply a

consequence of the speciﬁc base and solvent. For TBAB,

THAB, and TOAB, the amount of neutral phenol transferred

was found to be 6.78, 4.92, and 2.30%, respectively. The less

lipophilic tetraalkylammonium salts transported a greater

amount of phenol in the organic phase. The transport of

neutral phenol implicates a still more complex aggregate

structure wherein phenol may serve to stabilize phenoxide

anions by hydrogen bonding.

Figure 9. Third liquid phase visible with 1.0 equiv of TPAB.

by UV−vis spectroscopy indicated that the phenoxide

concentration was a little more than twice the starting aqueous

phase concentration.⁶⁵In contrast, at 0.2 equiv catalyst loading,

a large diﬀerence in phenoxide transport was observed between

TEAB and TBAB (a factor of 16.5). The largest diﬀerence was

seen between TPAB and TBAB, which aﬀorded nearly a 7-fold

increase in the transport of phenoxide out of the aqueous

phase. At 0.2 equiv of TPAB, the amount of transferred

phenoxide was drastically reduced, and no TLP was observed.

Taking into account the relatively large alkylation rate observed

for TPAB (a factor of 120 over background), this ﬁnding

suggests that the observed TLP is rich in both ammonium and

phenoxide ions. The critical catalyst concentration required for

the formation of TLP with TPAB was not determined.

Perhaps the most startling ﬁnding of the survey of phenoxide

extraction coeﬃecients as a function of tetraalkylammonium

ion was that the amount of phenol transferred to the organic

phase could be greater than the amount of ammonium bromide

employed! With only 0.2 equiv (20 mol %) of tetraalkylammo-

nium bromides, 27.2%, 27.6%, and 29.3% of the available

phenoxide was transferred to the organic phase for TBAB, THAB,

and TOAB, respectively. This corresponds to a greater than

100% transfer of phenoxide per molecule of tetraalkylammo-

nium salt. This striking discovery necessitated a direct analysis

of the composition of the organic phase to conﬁrm such an

unexpected outcome.

4.2. Titration of Active Phenoxide in the Organic Phase.

The second series of experiments involved careful quantitative

analysis of the amount of phenoxide transported to the organic

phase. This determination was accomplished by allowing the

aqueous and organic phases containing phenol and tetraalky-

lammonium bromide to reach equilibrium, removing precisely

measured aliquots from the organic phase, and quenching these

aliquots with an excess of benzyl bromide. The quenched

reaction aliquots were analyzed for both phenol and benzyl

phenyl ether by GC with the aid of biphenyl as an internal

standard. The same tetraalkylammonium salts were surveyed at

0.1 equiv with respect to phenol (Figure 10).

4.3. Determination of the Amounts of R₄N⁺and K⁺

Phenoxide in the Organic Phase. The intriguing observation

of superstoichiometric transport of phenoxide to the organic

phase was further investigated to determine the relative amount

of tetraalkylammonium vs potassium phenoxide being trans-

ferred. Experiments to quantify the amount of potassium

phenoxide associated with [R₄N⁺PhO⁻] in the organic phase

were limited to TBAB, THAB, and TOAB. The protocol for

these experiments was the same as for the titrations described

in the previous section: 0.10 equiv of the tetraalkylammonium

bromide, quenching aliquots of the organic phase with an

excess of benzyl bromide. The potassium bromide that was

produced in the alkylation was then removed by an aqueous

extraction. The potassium concentration of the resulting

aqueous solution was determined by inductively coupled

plasma optical emission spectroscopy (ICP-OES) (Table 4).

The amount of ammonium phenoxide present in the organic

phase was calculated as the diﬀerence between the transferred

potassium phenoxide and the total active phenoxide, as was

determined by the benzyl bromide titrations in section 4.2.

The results presented in Figure 9 qualitatively conﬁrm the

striking conclusion from UV−vis spectroscopy. The hydrophilic

ammonium salts, TMAB, TEAB, and TPAB, transferred only

traces of phenoxide to the organic phase; in fact, more neutral

13423

dx.doi.org/10.1021/ja304808u | J. Am. Chem. Soc. 2012, 134, 13415−13429

Journal of the American Chemical Society

Article

that the longer alkyl chains in (n-Hex)₄N⁺and (n-Oct)₄N⁺can

introduce a signiﬁcant entropic component to the transition

structure of the alkylation reaction (Price−Hammett principle),

thus lowering the rate.⁶⁷

Table 4. ICP-OES Determination of Potassium

Concentrations

a

R₄N⁺Br⁻, R

(0.10 equiv)

transfer of K⁺PhO⁻, %

(standard deviation)

R₄N⁺PhO⁻, %

in the organic phase

n-Bu

0.41 (0.05)

4.40 (0.08)

6.95 (0.70)

89.4

97.3

98.2

n-Hex

n-Oct

a

All values shown are the mean of triplicate runs.

For TBAB only a small amount of potassium phenoxide was

found in the organic phase. However, 11- and 17-fold increases

in the amount of transferred potassium phenoxide were

observed for THAB and TOAB, respectively. Additionally,

the more lipophilic the ammonium cation, the greater the

amount of the quaternary ammonium phenoxide ion pair

present in the organic phase. These results clearly show that, for

TBAB, THAB, and TOAB, between 89.4 and 98.2 of the

available phenoxide is transported to the organic phase. Most

importantly, the unexplained surplus of phenoxide seen in both

UV−vis analysis and benzyl bromide titrations can be

accounted for by transport of potassium phenoxide into the

organic phase by THAB and TOAB. Since absolutely no

potassium phenoxide is transported to the organic phase in the

absence of the tetraalkylammonium salt, these results point to

the formation of aggregates containing both tetraalkylammo-

nium and potassium cations.

Figure 11. Comparison of relative rates of alkylation closest approach

parameter a. Extraction constant

= /

[R₄N⁺(picrate⁻)]_org

{[R₄N⁺]_aq[picrate⁻]_aq}.

1.2. Catalytic Activity. In contrast, under PTC reaction

conditions (in DIPK), from K⁺to (n-Bu)₄N⁺, the range in

reactivity spanned a factor of 663. For quaternary ammonium

cations larger than Me₄N⁺in the PTC series, the reactivity

range spanned a factor of 144. Thus, by comparison of the

stoichiometric and catalytic rates, it is clear that the intrinsic

reactivity of the tetraalkylammonium phenoxide ion pair does

not contribute signiﬁcantly to the reactivity of the quaternary

ammonium salts under PTC reaction conditions.

DISCUSSION

■

1. Comparison of Stoichiometric and Catalytic

Systems. 1.1. Stoichiometric Reactivity. The stoichiometric

reactivity and catalytic activity proﬁles of the quaternary

ammonium salts diﬀered signiﬁcantly. The data collected in

Figure 4 reveal a number of important trends. First, the

reactivity of all of the tetraalkylammonium phenolates was

greater than that of the sodium or potassium phenolates.

Second, under homogeneous reaction conditions the range of

reactivity of the tetraalkylammonium phenolates varied by a

factor of only 2.4 in acetonitrile and 6.8 in DIPK. If only the

quaternary ammonium cations larger than Me₄N⁺are

compared, then the reactivity range is reduced further to 1.4-

and 1.7-fold for acetonitrile and DIPK, respectively. Third, the

homo-hydrogen-bonded complexes Et₄N⁺, (n-Hex)₄N⁺, and (n-

Oct)₄N⁺reacted at nearly identical rates (1.1-fold diﬀerence).

However, the (n-Pr)₄N⁺and (n-Bu)₄N⁺complexes reacted at

slightly increased rates (1.5 and 1.7 times, respectively).

In the catalytic cycle, the reactive ion pair, R₄N⁺PhO⁻, is ﬁrst

generated in the aqueous or interphase layer and must traverse

into the organic phase for the alkylation to occur. For the

[Me₄N⁺PhO⁻] and [Et₄N⁺PhO⁻] ion pairs this transport is

unfavorable because of their lesser lipophilic character

compared to [(n-Bu)₄N⁺PhO⁻], [(n-Hex)₄N⁺PhO⁻], and [(n-

Oct)₄N⁺PhO⁻]. The tendency for quaternary ammonium salts

to enter the organic phase is accurately reﬂected in their

extraction constants. These equilibrium constants have been

measured for many diﬀerent quaternary ammonium salts of

varying alkyl lengths and for diﬀerent gegenions in organic/

aqueous biphases.⁶⁸Of relevance to this work, the extraction

constants for several quaternary ammonium picrates partition-

ing between methylene chloride and water were chosen for

comparison.⁶⁹Graphical presentation of the log of these

extraction constants with the log of rate constant ratios⁷⁰for

the PTC reaction shows an excellent correlation (Figure 12).

This correlation strongly suggests that the rate of the PTC

reaction is dominated by the change in eﬀective concentration

of the reactive quaternary ammonium phenoxide in the organic

phase.

These data clearly show that the eﬀect of charge separation

on rate appears to be small and does not reﬂect the ionic radii

of the corresponding ammonium ions; i.e. an increase in the

ionic radius is expected to correlate with an increase in

reactivity. However, a structural parameter more relevant to ion

pair reactivity is the “closest approach of the ion” (a), which

reﬂects the distance between the positive charge of the

ammonium cation and the negative charge of the phenolate.

The parameter a can be estimated by comparing ion pair

formation constants (K_ip) with predictions form Bjerrum

theory. Closest approach parameters have been determined

for the quaternary ammonium picrates (C_nH_2n+2)₄N⁺for n =

1−8 and are plotted against the relative rates of alkylation in

Figure 11.⁶⁶For Me₄N⁺through (n-Bu)₄N⁺, the reactivity of

the phenoxide ion pair correlates well with the closest contact

parameter, whereas for (n-Hex)₄N⁺and (n-Oct)₄N⁺the

predicted reactivity deviates. The origin of this deviation is

2. QSAR Analysis of Catalytic Activity.⁷¹Previous

studies have described catalytic activities of tetraalkylammo-

nium ions that span as much as 2−2.5 orders of magnitude in

certain PTC reactions. In these studies, the reactions operate by

an extraction mechanism, and the catalytic activity is linearly

dependent on catalyst lipophilicity, as measured by the number

of carbons in the catalyst or the catalyst log P. In our prior

investigations, we observed that PTC reactions operating by an

13424

dx.doi.org/10.1021/ja304808u | J. Am. Chem. Soc. 2012, 134, 13415−13429

Journal of the American Chemical Society

Article

Figure 12. PTC initial alkylation rate with log(extration constant)

overlayed.

interfacial mechanism exhibit a range of catalytic activity up to

3.8 orders of magnitude. Therefore, it was surprising to observe

such a stark break from linearity in the plot of catalyst activity

vs catalyst lipophilicity as illustrated in Figure 12.

The data in Figure 11 clearly show a good correlation

between catalytic rate and lipophilicity for the smaller

ammonium ions (Me₄N⁺, Et₄N⁺, (n-Pr)₄N⁺, and (n-Bu)₄N⁺).

In this series, the increase in the catalytic turnover frequency is

directly proportional to the number of carbons in the

ammonium ion catalyst. No further increase in catalytic activity

is observed for more lipophilic catalysts. These conclusions can

be readily gleaned from a graphical representation that plots

log k_catvs either the total number of carbons in the catalyst

(Figure 13a) or ClogP (Figure 13b). However, the

representation of lipophilicity as the log of the partition

coeﬃcient is more valuable, as it lends naturally to

interpretation of the data as a rate−equilibrium QSAR. The

calculated partition coeﬃcient between octanol and water

(ClogP_(o/w)) is used for the x-axis because computational

models for DIPK solvation are not as readily available.

To facilitate analysis by a LFER the data are ﬁtted to a linear

biexponential equation (eq 1), with regression statistics n = 6,

R²= 0.991, s = 0.190. To the best of our knowledge, this is the

ﬁrst application of this kind of analysis in catalysis; therefore,

clariﬁcation of the parts that make up the ﬁtted equation is

warranted.

Figure 13. (a) Plot of the catalytic activity of homologous quaternary

ammonium ions vs the number of carbons in each catalyst. (b) Log−

log plot of the catalytic activity of homologous quaternary ammonium

bromides vs the octanol−water partition coeﬃcient. The black line is a

partial least-squares ﬁt of eq 1.

bilinear model as x_o(purple dashed line, x_o= 0.198), represents

the point of intersection of the two LFERs. The point of

intersection indicates that the maximum catalytic turnover

frequency is not reached until the catalyst is lipophilic enough

to reside predominantly in the organic phase (ClogP > 0). The

second LFER is represented by the red dashed line with a slope

of zero (descending slope, b = 0). Finally, the constant c

corresponds to the maximum value of the catalytic data set (on

a log₁₀scale) and is coincident with the red dashed line, i.e., the

magnitude of rate acceleration over the background rate. The

value of η, which aﬀects the curvature at the intersection of two

LFERs, was not varied in the regression analysis. Therefore, η

will not be discussed further.

3. Mechanistic Analysis. A break in an LFER-type plot

describing reactivity has two common interpretations, namely, a

change in rate-determining step or a change in mechanism.

However, in studies of catalysis a third possibility exists, namely,

a change in a pre-equilibrium position and no change in rate-

determining step or mechanism. The data collected to date are

most consistent with this third possibility and can be readily

interpreted in the context of a continuum of the extraction and

interfacial mechanisms.¹¹Depending upon the lipophilicity of

the tetraalkylammonium salt, several equilibria operate prior to

the irreversible alkylation step for hydroxide-initiated PTC

reactions that follow an extraction mechanism (Figure 14). For

less lipophilic tetraalkylammonium salts the four equilibria

involve (1) partitioning of the substrate, (2) deprotonation of

the substrate, (3) ion exchange of the potassium and

ammonium salts, and (4) partitioning of the ammonium·sub-

strate ion pair into the organic phase (k₁−k₄). However, with

highly lipophilic salts having neglibigle water solubility, three

log(k_cat./k_o) = −η ln[e^{−a(x−x )/n}+ e^{b(x−x )/n}] + c

o

(1)

where η = 0.434, x_o= 0.198, a = 0.588, b = 0.00, and c = 2.184.

The linearized biexponential model, as ﬁrst proposed by

Buckwald, is a useful tool for the analysis of QSAR data

showing any type of bilinear distribution.⁷²Part of the utility of

the linearized biexponential model is the relative ease with

which the data are related to the two intersecting LFERs that

comprise the bilinear distribution. In this case, the ﬁrst LFER is

reﬂected in the slope of the regression over the ﬁrst four data

points (blue dashed line, rising slope, a = 0.588). This LFER

reﬂects an increase of 0.59 orders of magnitude in catalyst

activity per order of magnitude change in the partition

coeﬃcient and is also reﬂected in the bilinear model (a =

0.588). The LFER holds until the partition coeﬃcient is near

zero, at which point a sharp break is seen and no further change

in catalytic activity is observed. The break point, shown in the

13425

dx.doi.org/10.1021/ja304808u | J. Am. Chem. Soc. 2012, 134, 13415−13429

Journal of the American Chemical Society

Article

additional equilibria (k₆−k₈) operate at the interface to eﬀect

ion exchange of the aqueous phenoxide. As described above,

previous studies on the structure−activity relationship of

quaternary ammonium ion catalysts for reactions following an

extraction mechanism have concluded that the more lipophilic

the catalyst (more carbons), the greater will be the activity of

the catalyst. No maximum or “leveling-oﬀ” of catalytic activity is

on record.

contained a signiﬁcant quantity of potassium phenoxide and

phenol, particularly with the more lipophilic catalysts, and the

ratio of active phenoxide to phenol was unique to a particular

ammonium salt. These ﬁndings are striking for two reasons.

First, given the diﬀerence in pK_abetween water and phenol

(15.7 and 9.95, respectively), the observation that any phenol

was left protonated was unexpected. In a strongly basic,

aqueous medium, phenoxide should be highly favored at

equilibrium. However, at the interfacial region of a PTC

process, this equilibrium may not be as favorable because the

stability of homo-hydrogen-bonded dimers could signiﬁcantly

raise the remaining pK_aas the polarity of the medium

decreases. Indeed, Dehmlow and Sasson have observed that

quaternary ammonium monoalkoxides of diols exhibit greater

lipophilicity (larger extraction constants) than the alkoxides of

simple alcohols.⁷⁴This behavior is explained by the formation

of an intramolecular hydrogen bond within the mono-

deprotonated diol that is absent in the alcohol.

Second, despite the unique phenol/phenoxide ratios among

(n-Bu)₄N⁺, (n-Hex)₄N⁺, and (n-Oct)₄N⁺, their PTC initial

alkylation rates are nearly identical. This discrepancy implies

that, under PTC reaction conditions, the ratio of aggregate

components is inconsequential to the reactivity of the

ammonium phenoxide. The sole determinant of catalytic

activity is the eﬀective concentration of the ammonium

phenoxide in the organic phase. In hindsight, this result should

not be surprising, given the 25-fold greater rate of reaction of

(n-Bu)₄N⁺[(PhOH)OPh]⁻compared to K⁺OPh⁻in DIPK. A

word of caution is warranted here because the presence of

potassium alkoxides in the organic phase need not be innocent

in other reactions.

To understand the formation of the potassium/ammonium

phenoxide aggregates, a closer examination of the mechanism

of ammonium phenoxide transport is required. As an

ammonium cation approaches the interphase boundary

separating the bulk aqueous and organic phases, its alkyl chains

are oriented toward the more lipophilic organic phase due to

hydrophobic interactions with the aqueous phase. This

conformation change creates a greater charge accessibility of

the ammonium center. At the interphase the ammonium

phenoxide forms ﬁrst, and with this increased accessibility,

additional potassium phenoxide molecules may complex with

the ammonium phenoxide to form a dimeric ion pair. This ion

pair is highly charged, and without suﬃcient lipophilicity

provided by the ammonium alkyl chains it is unable to cross

into the organic phase (Figure 15). On basis of the benzyl

bromide titration data, the TBAB cation is not suﬃciently

lipophilic to transport the heteronuclear aggregate to the

organic phase, whereas the (n-Oct)₄N⁺cation can provide

suﬃcient lipophilicity to carry the dimeric ion pair into the

organic phase.⁷⁵

Figure 14. Schematic representation of the hydroxide-initiated PTC

alkylation of phenols.

The data collected herein are completely consistent with

those studies in that catalyst activity is linearly correlated with

catalyst lipophilicity (number of carbons or log P) over a range

of ∼2.5 orders of magnitude in catalyst activity. This trend is

easily interpreted in terms of a corresponding increase in the

concentration of the active species in solution as measured

from the titration experiments. However, in the alkylation of

phenol no further increase in catalytic activity is observed

beyond a rate enhancement of 2.8 orders of magnitude over the

background reaction for more lipophilic catalysts (e.g., (n-

Oct)₄N⁺ClogP = 9.56, #C = 32). The observed leveling-oﬀ of

catalytic activity can also be easily explained by the results from

the titration experiments. For the lipophilic quaternary

ammonium cations ((n-Bu)₄N⁺, (n-Hex)₄N⁺, and (n-

Oct)₄N⁺), the equilibrium concentrations of the reactive

species Q⁺PhO⁻in the organic phase (k₈/k₋₈, Figure 14)

reach their maximum obtainable values under these conditions.

It is important to note that the extraction constants of the

tetraalkylammonium cations have not leveled oﬀ, but under

these conditions (0.1 equiv of catalyst) the enhanced ClogPs

cannot be manifested.⁷³The alternative interpretation that the

break-point corresponds to a change in rate-determining step is

ruled out by the uniform kinetic order of the electrophile on

either side of the break point.

4. Analysis of the UV−Vis Spectroscopy and Benzyl

Bromide Titration Data. Although the UV−vis spectroscopy

data were not able to quantitatively determine the composition

of reactive species in the organic phase, they did reveal the

superstoichiometric transport of phenoxide species into the

organic phase. From the observed aqueous-phase concen-

trations of phenoxide with (n-Bu)₄N⁺, (n-Hex)₄N⁺, and (n-

Oct)₄N⁺, it was determined that ∼50% more phenoxide was

transferred to the organic phase than the theoretical maximum

as constrained by the ammonium salt loading.

These same interactions may also explain the inverse

relationship between organic-phase phenol content and

ammonium lipophilicity. Since the dimeric aggregate with (n-

Bu)₄N⁺is insuﬃciently lipophilic for transport, phenol may

displace a potassium phenoxide from the aggregate at the

interphase. The newly formed homo-hydrogen-bonded com-

plex would have less charge than the former aggregate and

would now be suﬃciently lipophilic for transport into the

organic phase. In the case of (n-Oct)₄N⁺, the dimeric aggregate

readily moves into the organic phase before phenol can displace

a potassium phenoxide. With the increased alkyl chain length of

The benzyl bromide titration experiments provided a ﬁner

resolution of the composition of the species being transported

to the organic phase. Much to our surprise, the aggregate

13426

dx.doi.org/10.1021/ja304808u | J. Am. Chem. Soc. 2012, 134, 13415−13429

Journal of the American Chemical Society

Article

(n-Oct)₄N⁺, this aggregate may also be sterically inaccessible for

formation of a homo-hydrogen-bonded complex.

potassium salt, which is likely shepherded into the organic

phase in a higher order aggregate along with neutral phenol.

The composition of these aggregates had no eﬀect on reaction

rate. Further studies on the behavior of PTC alkylation of

phenol under an interfacial mechanism will be reported in due

course.

ASSOCIATED CONTENT

■

S

* Supporting Information

Detailed experimental procedures for and characterization of

the ammonium phenoxides, including ¹H and ¹³C NMR

spectra, tables of all of the kinetic runs for alkylation of each

ammonium phenoxide, stirring rate experiments, and titrations.

This material is available free of charge via the Internet at

http://pubs.acs.org.

Figure 15. Proposed dimeric ion pair aggregates.

5. Tetrapropylammonium Bromide and the Third

Liquid Phase. The rate of TPAB-catalyzed alkylation did not

reﬂect the active phenoxide concentration in the organic phase.

All of the spectroscopic and thermodynamic partition data

suggest that this ammonium salt is hydrophilic and should lead

to a slow alkylation under PTC conditions. Yet, the alkylation

rate for TPAB was closer to those of the lipophilic ammonium

salts than to those of the hydrophilic ammonium salts.

AUTHOR INFORMATION

■

Corresponding Author

sdenmark@illinois.edu

Notes

The authors declare no competing ﬁnancial interest.

A reasonable explanation for the enhanced PTC alkylation

rate with TPAB is the formation of a TLP. Earlier studies have

conclude that the TLP in a PTC process is a catalyst-rich phase,

and therefore the presence of a TLP would explain the

enhanced alkylation rate for an intrinsic reaction that was ﬁrst

order in catalyst.^{22,34,76−78}The TLP was easily observed when a

full equivalent of TPAB with respect to phenol was introduced.

However, no TLP was visually observed when a substoichio-

metric amount (0.05 or 0.10 equiv) of TPAB was employed.

The inability to observe a TLP does not exclude the possibility

of its existence under the reaction conditions. Sasson et al.

noted the formation of a TLP in the elimination of phenethyl

bromide to styrene under PTC conditions.⁷⁶Upon closer

examination, this TLP was shown to exist as a membrane

around an aqueous droplet. If such a TLP were formed in

reaction with a low ammonium salt content, it would be

reasonable to assume that the total volume of the TLP was too

low to allow coalescence of TLP droplets into a visible layer.

ACKNOWLEDGMENTS

■

We are grateful to the American Chemical Society Petroleum

Research Fund (ACS PRF 49668-ND1) for generous ﬁnancial

support. N.D.-G. thanks Amgen for a Graduate Fellowship in

Synthetic Organic Chemistry.

REFERENCES

■

(1) (a) Anastas, P.; Warner, J. Green Chemistry: Theory and Practice;

Oxford University Press: Oxford, UK, 1998. (b) Weber, W. P.; Gokel,

G. W. Phase Transfer Catalysis in Organic Synthesis; Springer-Verlag:

Berlin/New York, 1977; Vol. 4. (c) Dehmlow, E. V.; Dehmlow, S. S.

Phase Transfer Catalysis; Verlag Chemie: Weinheim/Deerﬁeld Beach,

1983; Vol. 11. (d) Phase-Transfer Catalysis: New Chemistry, Catalysts,

and Applications; American Chemical Society: Washington, DC, 1985.

(e) Starks, C. M.; Liotta, C. L.; Halpern, M. Phase-Transfer Catalysis:

Fundamentals, Applications and Industrial Perspectives; Chapman and

Hall: New York, 1994. (f) Phase-Transfer Catalysis: Mechanisms and

Syntheses; Halpern, M. E., Ed.; American Chemical Society:

Washington, DC, 1997. (g) Makosza, M.; Fedorynski, M. In Interfacial

Catalysis; Volkov, A. G., Ed.; Marcel Dekker: New York, 2003; pp

150−201.

(2) Handbook of Phase Transfer Catalysis; Sasson, Y., Neumann, R.,

Eds.; Chapman & Hall: London, 1997.

(3) Jones, R. A. Quaternary Ammonium Salt: Their Use in Phase

Transfer Catalysis; Academic Press: London, 2001.

(4) (a) Denmark, S. E.; Gould, N. D.; Wolf, L. M. J. Org. Chem. 2011,

76, 4337−4357. (b) Denmark, S. E.; Gould, N. D.; Wolf, L. M. J. Org.

Chem. 2011, 76, 4260−4336.

(5) Ionization Constants of Organic Acids in Solution; IUPAC

Chemical Data Series No. 23, Sergent, E. P., Dempsey, B., Eds.;

Pergamon Press: Oxford, 1979.

(6) Bordwell, F. G.; McCallum, R. J.; Olmstead, W. N. J. Org. Chem.

1984, 49, 1424−1427.

(7) (a) Taft, R. W.; Bordwell, F. G. Acc. Chem. Res. 1988, 21, 463−

469. (b) Bordwell, F. G.; Harrelson, J. A. Can. J. Chem. 1990, 68,

1714−1718.

(8) (a) Halpern, M. Phase Trans. Catal. Commun. 1995, 1, 1−3.

(b) Yang, H.-M.; Wu, H.-S. Catal. Rev. 2003, 45, 463−540.

(9) Rabinovitz, M.; Cohen, Y.; Halpern, M. Angew. Chem., Int. Ed.

Engl. 1986, 25, 960−970.

(10) A companion study of the catalyst SAR for phenol alkylation

under the interfacial mechanism is planned.

CONCLUSIONS

■

In the PTC alkylation of phenol operating by an extraction

mechanism, the catalytic reactivity of linear, symmetrical

tetraalkylammonium salts is directly related to their eﬀective

concentration in the organic phase, in which the alkylation

reaction takes place. The range of catalytic rates spans a 663-

fold diﬀerence for the series TMAB to TOAB. By comparison,

the stoichiometric reactivity of the same tetraalkylammonium

salts of the phenoxide homo-hydrogen-bonded dimers diﬀers

by a factor of only 6.8. Under PTC reaction conditions, the

concentration of the active tetraalkylammonium phenoxide and

the initial alkylation rates reach a maximum for tetrabutylam-

monium salts and do not increase further with more lipophilic

salts. The leveling oﬀ in catalytic activity (i.e., a break in the

LFER) was not the result of a change in mechanism. Titration

experiments demonstrated that, for the more lipophilic salts

(TBAB, THAB, TOAB), a superstoichiometric amount of

phenoxide is transported to the organic phase. These

experiments further revealed that a near-theoretical maximum

of the ammonium phenoxide was transported to the organic

phase, and the additional active phenoxide arose from the

13427

dx.doi.org/10.1021/ja304808u | J. Am. Chem. Soc. 2012, 134, 13415−13429

Journal of the American Chemical Society

Article

(11) For the most recent summary of the state of mechanistic

understanding of phase transfer catalysis, see: Makosza, M.;

Fedorysnki, M. Catal. Rev. 2003, 45, 321−367.

(35) Mason, D.; Magdassi, S.; Sasson, Y. J. Org. Chem. 1991, 56,

7229−7232.

(36) Bordwell, F. G.; Hughes, D. L. J. Org. Chem. 1982, 47, 3224−

3232.

(12) Starks, C. M.; Owens, R. M. J. Am. Chem. Soc. 1973, 95, 3613−

3617.

(37) Caulton, K. G.; Chisholm, M. H.; Drake, S. R.; Folting, K.;

Huffman, J. C.; Streib, W. E. Inorg. Chem. 1993, 32, 1970−1976.

(38) (a) Dinnebier, R. E.; Pink, M.; Sieler, J.; Norby, P.; Stephens, P.

W. Inorg. Chem. 1998, 37, 4996−5000. (b) Dinnebier, R. E.; Pink, M.;

Sieler, J.; Stephens, P. W. Inorg. Chem. 1997, 36, 3398−3401.

(39) Fraser, M. E.; Fortier, S.; Markiewicz, M. K.; Rodrigue, A.;

Bovenkamp, J. W. Can. J. Chem. 1987, 65, 2558−2563.

(40) Caulton, K. G.; Chisholm, M. H.; Drake, S. R.; Folting, K.;

Huffman, J. C.; Streib, W. E. Inorg. Chem. 1993, 32, 1970−1976.

(41) See, for example: (a) Chantooni, M. K., Jr.; Kolthoff, I. M. J.

Phys. Chem. 1976, 80, 1306−1310. (b) Arnett, E. M.; Moriarity, T. C.;

Small, L. E.; Rudolph, J. P.; Quirk, R. P. J. Am. Chem. Soc. 1973, 95,

1492−1495. (c) Arnett, E. M.; Small, L. E.; Oancea, D.; Johnston, D. J.

Am. Chem. Soc. 1976, 98, 7346−7350. (d) Arnett, E. M.;

Venkatasubramanium, K. G. J. Org. Chem. 1983, 48, 1569−1578.

(e) Silva, P. J. J. Org. Chem. 2009, 74, 914−916.

(42) Nielsen, M. F.; Hammerich, O. Acta Chem. Scand. 1989, 43,

269−274.

(43) Goddard, R.; Reetz, M.; Herzog, M. Tetrahedron 2002, 58,

7847−7850.

(44) Based on a search of the Cambridge Structural Database on

April 27, 2011, in combination with energy minimization of

monomeric sodium and potassium phenolates (6-31G*//RHF).

(45) Viout, P. J. Mol. Catal. 1981, 10, 231−240.

(13) Wang, M.-L. In Handbook of Phase Transfer Catalysis; Sasson, Y.,

Neumann, R., Eds.; Chapman & Hall: London, 1997; Chapter 2, pp

36−107.

(14) (a) Herriot, A. W.; Picker, D. Tetrahedron Lett. 1972, 13, 4521−

4524. (b) Gordon, J. E.; Kutina, R. E. J. Am. Chem. Soc. 1977, 99,

3903−3909. (c) Herriot, A.; Picker, D. J. Am. Chem. Soc. 1975, 97,

2345−2349. (d) Landini, D.; Maia, A.; Montanari, F. J. Am. Chem. Soc.

1978, 100, 2796−2801. (e) Landini, D.; Maia, A.; Montanari, F. J.

Chem. Soc., Chem. Commun. 1977, 112−113.

(15) (a) Dehmlow, E. V. Angew. Chem., Int. Ed. Engl. 1977, 16, 493−

505. (b) Brandstrom, A. Adv. Phys. Org. Chem. 1977, 15, 267−330.

̈

(16) For an insightful analysis of hydroxide-initiated PTC reactions,

see ref 1e, pp 89−108.

(17) Naik, S. D.; Doraiswamy, L. K. Am. Inst. Chem. Eng. J. 1998, 44,

612−646.

(18) (a) Wang, M.-L.; Yang, H.-M. Chem. Eng. Sci. 1991, 46, 619−

627. (b) Wang, M.; Yang, H. Ind. Eng. Chem. Res. 1990, 29, 522−526.

(19) McKillop, A.; Fiaud, J.-C.; Hug, R. P. Tetrahedron 1974, 30,

1379−1382.

(20) Taft, R. W.; Bordwell, F. G. Acc. Chem. Res. 1988, 21, 463−469.

(21) (a) Wang, D. H.; Weng, H. S. J. Chin. Inst. Chem. Eng. 1996, 27,

129−139. (b) Weng, H. S.; Wang, D. H. J. Chin. Inst. Chem. Eng. 1996,

27, 419−426. (c) Wang, D. H.; Weng, H. S. J. Chin. Inst. Chem. Eng.

1995, 26, 147−156. (d) Wang, D. H.; Weng, H. S. J. Eng. Sci. 1995, 50,

3477.

(46) Minot, C.; Trong, N. A. Tetrahedron Lett. 1975, 16, 3905−3908.

(47) (a) Johnson, M. L.; Rodriguez, C.; Benjamin, I. J. Phys. Chem. A

2009, 113, 2086−2091. (b) Nelson, K. V.; Benjamin, I. Chem. Phys.

Lett. 2011, 508, 59−62.

(22) Yadav, G. D.; Desai, N. M. Org. Process Res. Dev. 2005, 9, 749−

756.

(23) (a) Makosza, M. Tetrahedron Lett. 1966, 7, 4261−4624.

(b) Makosza, M. Pure Appl. Chem. 1975, 43, 439−462. (c) Makosza,

M.; Bialecka, E. Tetrahedron Lett. 1977, 18, 183−186.

(24) (a) Halpern, M.; Sasson, Y.; Willner, I.; Rabinovitz, M.

Tetrahedron Lett. 1981, 22, 1719−1722. (b) Halpern, M.; Sasson, Y.;

Rabinovitz, M. J. Org. Chem. 1983, 48, 1022−1025. (c) Halpern, M.;

Sasson, Y.; Rabinovitz, M. J. Org. Chem. 1984, 49, 2011−2012.

(d) Halpern, M.; Feldman, D.; Sasson, Y.; Rabinovitz, M. Angew.

Chem. 1984, 96, 79−80. (e) Halpern, M.; Zahalaka, H. A.; Sasson, Y.;

Rabinovitz, M. J. Org. Chem. 1985, 50, 5088−5092. (f) Feldman, D.;

Halpern, M.; Rabinovitz, M. J. Org. Chem. 1985, 50, 1746−1749.

(25) Synthetic and Natural Phenols; Tyman, J. H. P., Ed.; Elsevier:

Amsterdam, 1996; Vol. 52.

(26) Lewis, E. S.; Vanderpool, S. J. Am. Chem. Soc. 1977, 99, 1946−

1949.

(27) For an overview and discussion of theoretical derivations of

ammonium ion pair reactivity, see : Ions and Ion Pairs in Organic

Reactions; Szwarc, M., Ed.; John Wiley & Sons: New York, 1974; Vols.

1 and 2.

(48) (a) Nelson, K. V.; Benjamin, I. J. Phys. Chem. C 2010, 114,

1154−1163. (b) Nelson, K. V.; Benjamin, I. J. Phys. Chem. C 2011,

115, 2290−2296.

(49) Pradines, V.; Poteau, R.; Pimienta, V. Eur. J. Chem Phys. Phys.

Chem 2007, 8, 1524−1533.

(50) Coleman, M. T.; Leblanc, G. Org. Process Res. Dev. 2010, 14,

732−736.

(51) Harlow, G. A.; Noble, C. M.; Wyld, G. E. A. Anal. Chem. 1956,

28, 787−791.

(52) Abrams, I. M.; Millar, J. R. React. Funct. Polym. 1997, 35, 7−22.

(53) For tables of the titration data, see Supporting Information.

(54) Section 6: Fluid Properties. CRC Handbook of Chemistry and

Physics, 92nd ed.; Haynes, W. M., Lide, D. R., Eds.; CRC Press: Boca

Raton, FL, 2011−2012; pp 212 − 228.

(55) Hopkins, R. N.; Yerger, E. S.; Lynch, C. C. J. Am. Chem. Soc.

1939, 61, 2460−2461.

(56) Schneider, C. H.; Lynch, C. C. J. Am. Chem. Soc. 1943, 65,

1063−1066.

(57) The solubility of DIPK in H₂O at 20 °C is 45.5 mM: Section 5:

Aqueous Solubility and Henry’s Law Constants of Organic

Compounds. CRC Handbook of Chemistry and Physics, 92nd ed.;

Haynes, W. M., Lide, D. R., Eds.; CRC Press: Boca Raton, FL, 2011−

2012.

(28) Ugelstad, J.; Ellingsen, T.; Berge, A. Acta Chim. Scand 1966, 20,

1593−1598.

(29) Starks, C. M. In Phase-Transfer Catalysis; Halpern, M. E., Ed.;

American Chemical Society: Washington, DC, 1997; Vol. 659,

Chapter 2, pp 2−10.

(58) For sodium and potassium homo-hydrogen-bonded phenoxide

complex synthesis, see the Supporting Information.

(59) For the the raw data and kinetic plots, see the Supporting

Information.

(30) Reference 1e, p 86.

(31) (a) Brandstrom, A.; Junggren, U. Acta Chem. Scand. 1969, 23,

̈

2203−2204. (b) Brandstrom, A; Junggren, U. Acta Chem. Scand. 1969,

̈

23, 2204−2205. (c) Brandstrom, A.; Junggren, U. Acta Chem. Scand.

̈

(60) The background alkylation rate was found to be 2.99 × 10⁻⁸

1969, 23, 2536−2537. (d) Brandstrom, A.; Junggren, U. Acta Chem.

̈

mol L⁻¹s⁻¹.

Scand. 1969, 23, 3585−3586. (e) Brandstrom, A. Pure Appl. Chem.

̈

(61) For a recent instructive analysis of nonlinear membrane

permeability−lipophilicity models, see: Balaz, S. Chem. Rev. 2009, 109,

1793−1899.

(62) Anslyn, E. V.; Dougherty, D. Modern Physical Organic Chemistry;

University Science Books: Sausalito, CA, 2006; Chapters 7−9.

(63) For experimental details (calibration, raw data, etc.), see the

Supporting Information.

1982, 54, 1769−1782.

(32) Wilson, K.; Adams, D. J.; Rothenberg, G.; Clark, J. H. J. Mol.

Catal. A: Chem. 2000, 159, 309−314.

(33) Wang, M.-L. In Handbook of Phase Transfer Catalysis; Sasson, Y.,

Neumann, R., Eds.; Chapman & Hall: London, 1997; Chapter 2, pp

36−107.

(34) Wang, D.-H.; Weng, H.-S. Chem. Eng. Sci. 1988, 43, 2019−2024.

13428

dx.doi.org/10.1021/ja304808u | J. Am. Chem. Soc. 2012, 134, 13415−13429

Journal of the American Chemical Society

Article

(64) Studies in medicinal chemistry have shown repeatedly that the

contribution of a methylene unit to ClogP_(o/w)is additive for

homologous series like quaternary ammonium ions. See: (a) Hansch,

C. Acc. Chem. Res. 1969, 2, 232−239. (b) Hansch, C.; Kurup, A.; Garg,

R.; Gao, H. Chem. Rev. 2001, 101, 619−672.

(65) The eﬀective phenoxide concentration of the TLP was

calculated at 1620 mM; the starting aqueous phase concentration

was 759 mM.

(66) Sawada, K.; Chigira, F. J. Mol. Liq. 1995, 65/66, 265−268.

(67) Price, F. P., Jr.; Hammett, L. P. J. Am. Chem. Soc. 1941, 63,

2387−2393.

(68) Reference 1c, p 13.

(69) (a) Gustavii, K.; Schill, G. Acta Pharm. Suec 1966, 3, 241−258.

(b) Gustavii, K.; Schill, G. Acta Pharm. Suec 1966, 3, 259−268.

(c) Gustavii, K. Acta Pharm. Suec 1967, 4, 233−246.

(70) For this comparison, the rate constants are normalized to the

rate for the uncatalyzed reaction (i.e., K⁺OPh]⁻.)

(71) For recent reviews of the application of QSAR methodology to

catalytic acitivty, see: (a) Burello, E.; Rothenberg, G. Int. J. Mol. Sci

2006, 7, 375−404. (b) Maldonado, A. G.; Rothenberg, G. Chem. Soc.

Rev. 2010, 39, 1891−1902.

(72) Buchwald, P. J. Pharm. Sci. 2005, 94, 2355−2379.

(73) This assertion is supported by the demonstration that the

concentration of the reactive species Q⁺PhO⁻continues to increase in

partitioning experiments with 0.6 equiv of tetraalkylammonium salts:

(n-Bu)₄N⁺, 47.1% transport, and (n-Oct)₄N⁺, 63.2% transport.

(74) (a) Dehmlow, E. V.; Thieser, R.; Sasson, Y.; Pross, E.

Tetrahedron 1985, 41, 2927−2932. (b) Zerda, J.; Sasson, Y. J. Chem.

Soc., Perkin Trans. 2 1987, 1147−1151.

(75) Such supramolecular structures have been identiﬁed for

transport of hydrogen peroxide in ruthenium-catalyzed oxidations.

Sasson et al. have identiﬁed the formation of vesicles as the mechanism

of transport of hydrogen peroxide: Rothenberg, G.; Barak, G.; Sasson,

Y. Tetrahedron 1999, 55, 6301−6310.

(76) Mason, D.; Magdassi, S.; Sasson, Y. J. Org. Chem. 1991, 56,

7229−7232.

(77) Wang, D.-H.; Weng, H.-S. Chem. Eng. Sci. 1995, 50, 3477−3486.

(78) Mąkosza, M.; Kryłowa, I. Tetrahedron 1999, 55, 6395−6402.

13429

dx.doi.org/10.1021/ja304808u | J. Am. Chem. Soc. 2012, 134, 13415−13429

Article Doi

Effects of charge separation, effective concentration, and aggregate formation on the phase transfer catalyzed alkylation of phenol

DOI: 10.1021/ja304808u

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja304808u

Full text of DOI:10.1021/ja304808u

Products guided by the article

R&D Labs maybe for 1392488-18-4

Relevant to this article

Hot Product