Racemization and hydrolysis of (S)-naproxen 2,2,2-trifluoroethyl ester in non-polar solvents by strong neutral bases: Implication for ion-pair kinetic basicity and hydrolysis

Article abstract of DOI:10.1002/poc.741

By using strong neutral bases as catalyst, a detailed investigation of the racemization of (S)-naproxen 2,2,2-trifluoroethyl ester was conducted in the non-polar solvents isooctane, cyclohexane and n-hexane. The second-order interconversion constant k_int* as representing the ion-pair kinetic basicity in isooctane was first estimated and correlated with the equilibrium ion-pair basicity pK_ip in tetrahydrofuran, giving slopes of 0.768 and 0.689 for non-phosphazene and phosphazene bases, respectively, in the Bronsted correlations. The result was further compared with that for (S)-naproxen 2,2,2-trifluoroethyl thioester, showing about a 1-2 orders of magnitude enhancement of k_int* for the corresponding thio-containing analogue. A smaller influence of non-polar solvents (i.e. isooctane, n-hexane and cyclohexane) on k_int* was found. Kinetic analysis of the racemization and hydrolysis of (S)-naproxen 2,2,2-trifluoroethyl ester in isooctane and n-hexane containing 7-methyl-1,5,7-triazabicyclo[4.4.0]dec-5-ene and water suggests nucleophilic hydrolysis by the base, where the breakdown of tetrahedral intermediates I _R1 and I_S1 is the rate-limiting step and the hydrolysis constant k_hy is in proportion to the product of base and ion-pair concentrations. Copyright 2004 John Wiley & Sons, Ltd.

Full text of DOI:10.1002/poc.741

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 2004; 17: 387–392
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/poc.741
Racemization and hydrolysis of (S)-naproxen
2,2,2-trifluoroethyl ester in non-polar solvents
by strong neutral bases: implication for ion-pair
kinetic basicity and hydrolysis
Han-Yuan Lin, Eddy Lay, Wen-Yen Wen, Hamza Dewi, Yu-Chi Cheng and Shau-Wei Tsai*
Department of Chemical Engineering, National Cheng Kung University, Tainan 70101, Taiwan
Received 19 August 2003; revised 5 November 2003; accepted 8 December 2003
ABSTRACT: By using strong neutral bases as catalyst, a detailed investigation of the racemization of (S)-naproxen
epoc
2,2,2-trifluoroethyl ester was conducted in the non-polar solvents isooctane, cyclohexane and n-hexane. The second-
order interconversion constant kiꢀnt as representing the ion-pair kinetic basicity in isooctane was first estimated and
correlated with the equilibrium ion-pair basicity pKip in tetrahydrofuran, giving slopes of 0.768 and 0.689 for non-
phosphazene and phosphazene bases, respectively, in the Brønsted correlations. The result was further compared with
that for (S)-naproxen 2,2,2-trifluoroethyl thioester, showing about a 1–2 orders of magnitude enhancement of kiꢀnt
for the corresponding thio-containing analogue. A smaller influence of non-polar solvents (i.e. isooctane, n-hexane
and cyclohexane) on kiꢀnt was found. Kinetic analysis of the racemization and hydrolysis of (S)-naproxen 2,2,2-
trifluoroethyl ester in isooctane and n-hexane containing 7-methyl-1,5,7-triazabicyclo[4.4.0]dec-5-ene and water
suggests nucleophilic hydrolysis by the base, where the breakdown of tetrahedral intermediates IR1 and IS1 is the rate-
limiting step and the hydrolysis constant khy is in proportion to the product of base and ion-pair concentrations.
Copyright # 2004 John Wiley & Sons, Ltd.
Additional material for this paper is available in Wiley Interscience
KEYWORDS: racemization; hydrolysis; (S)-naproxen 2,2,2-trifluoroethyl ester; strong neutral bases; ion-pair basicity
INTRODUCTION
containing neutral organic bases, but the results were
disappointing owing to the lower racemization rate of the
remaining (R)-ester compared with the enzymatic rate of
fast-reacting (S)-ester.3 In order to explore more efficient
racemization catalysts for racemic esters containing a
chiral acid moiety in a non-polar solvent, isooctane
containing (S)-naproxen 2,2,2-trifluoroethyl ester and
strong neutral bases was first selected as a model system
for estimating the second-order interconversion constant.
It was then applied as a representation of the ion-pair
kinetic basicity of the base and compared with that for the
corresponding thio-containing analogue.4
When enzymatic biotransformations take place in non-
polar solvents, certain water retention in the enzyme
aggregate and dissolution in the solvent are needed to
maintain the enzyme in a catalytically active confor-
mation. It is then imperative to have non-enzymatic
hydrolysis catalyzed by the base in the dynamic kinetic
resolution of esters, thioesters or amides by employing
enzymatic methods. This may have profound effects on
decreasing the enantiomeric excess for the desired pro-
duct. Therefore, isooctane, n-hexane and cyclohexane
containing (S)-naproxen 2,2,2-trifluoroethyl ester, water
and 7-methyl-1,5,7-triazabicyclo[4.4.0]dec-5-ene (MTBD)
were designed as model systems for studying the kinetics
Base-catalyzed racemization has been employed for com-
pounds with a stereogenic center bearing an acidic proton
adjacent to an electron-withdrawing group, such as a
ketone or ester, via an enolate intermediate. Various types
of chiral ꢀ-substituted carboxyl acids, e.g. ꢀ-(hetero)
arylcarboxyl acids, ꢀ-aryloxypropionic acids, ꢀ-alkyl-
carboxyl acids, ꢀ-halocarboxyl acids and ꢀ-amino acids,
as valuable pharmaceuticals, agrochemicals, nutrients
and intermediates for organic synthesis have been pre-
pared via kinetic resolution processes. However, only
special carboxyl acid derivatives of Ketorolac esters,
hydantoins, substituted oxazol-5(4H)-ones, substituted
thiazolin-5-ones and ꢀ-substituted propionic thioesters
possessing relatively high ꢀ-proton acidity have been
employed in dynamic kinetic resolution using enzymatic
methods.1,2 Recently, lipase-catalyzed resolutions of ꢀ-
arylpropionic esters were demonstrated in organic solvents
*Correspondence to: S.-W. Tsai, Department of Chemical Engineer-
ing, National Cheng Kung University, Tainan 70101, Taiwan.
E-mail: t62647@mail.ncku.edu.tw
Contract/grant sponsor: Chinese National Science Council; Contract/
grant number: NSC 91-2214-E-006-025.
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 387–392

388
H.-Y. LIN ET AL.
of racemization and hydrolysis, where effects of ion-pair
formation between MTBD and water on the amine and
ion-pair concentrations were considered in deriving the
rate equations.
Chemical shifts in ppm from tetramethylsilane were as
follows: ꢂ1.61 (3H, t), 3.92 (3H, s), 3.95–4.00 (1H, q),
4.36–4.57 (2H, m), 7.12–7.17 (2H, q), 7.37–7.41 (1H, q),
7.67–7.73 (3H, m).
Racemization of (S)-naproxen 2,2,2-trifluoroethyl es-
ter. To 10 ml of anhydrous isooctane were added 1 mM
(S)-naproxen 2,2,2-trifluoroethyl ester and various bases
of different concentration with stirring at 45 ꢂC. Samples
were removed and injected into the above HPLC system
at different time intervals for analysis. From the time-
course variations of the enantiomeric excess for the ester
{i.e. eeS ¼ [(AS) ꢁ (AR)]/[(AS) þ (AR)] where (AS) and
(AR) are the concentrations of (S)- and (R)-naproxen
2,2,2-trifluoroethyl ester, respectively}, the first-order
interchange constant kint at each base concentration and
hence the second-order interchange constant kiꢀnt for the
base were estimated. Similar experiments with MTBD as
the base in anhydrous n-hexane and cyclohexane were
carried out.
EXPERIMENTAL
Materials. Optically pure (S)-naproxen [(S)-2-(6-methoxy-
2-naphthyl)propionic acid] was obtained from Sigma.
The strong neutral bases trioctylamine (TOA), 1,4-diaza-
bicyclo[2.2.2]octane (DABCO), 1,8-diazabicyclo[5.4.0]
undec-7-ene (DBU), MTBD, tert-butyliminotris-
(dimethylamino)phosphorane (P1), tert-butyliminotris
(pyrrolidino)phosphorane (P1-tris) and 1-tert-butyl-4,
4,4-tris(dimethylamino)-2,2-bis[tris(dimethylamino)pho-
sphoranylidenamino]-2ꢁ5,4ꢁ5-catenadi(phosphazene) (P4)
were purchased from Aldrich and Fluka. Other chemicals
of analytical grade were commercially available and
employed without further purification. Anhydrous isooc-
tane, cyclohexane and n-hexane were prepared by adding
molecular sieve (Type 3A, J. T. Baker) to the solvent for
more than 3 days.
Racemization and hydrolysis of (S)-naproxen 2,2,2-
trifluoroethyl ester. The racemization and hydrolysis of
1 mM (S)-naproxen 2,2,2-trifluoroethyl ester in isooctane,
n-hexane and cyclohexane containing different water
contents and MTBD were carried out. From the time-
course variations of (R)- and (S)-naproxen 2,2,2-trifluor-
oethyl ester concentrations and hence eeS, one could first
estimate the hydrolysis constant khy and kint, and then
the equilibrium constant Keq for the ion-pair formation
between MTBD and water and the kinetic constants for
hydrolysis.
Analysis. The racemization and hydrolysis of (S)-napro-
xen 2,2,2-trifluoroethyl ester in the organic solvent were
monitored by HPLC using a chiral column of (S,S)-
WHELK-01 from Regis capable of separating the internal
standard 2-nitrotoluene, (R)- and (S)-naproxen, (R)- and
(S)-ester with retention times of 4.5, 12.2, 21.5, 7.2 and
8.9 min, respectively. The mobile phase was a mixture of
n-hexane–propan-2-ol–acetic acid (80:20:0.5, v/v/v) at a
flow-rate of 1.0 ml minꢁ1. UV detection at 270 nm was
used for quantification at a column temperature of 25 ꢂC.
Synthesis of (S)-naproxen 2,2,2-trifluoroethyl ester.5 Fol-
lowing a standard procedure, the acid chloride of (S)-
naproxen was prepared by refluxing 20 ml of benzene
containing 3.45 g of the acid and 3.20 g of thionyl
chloride for 1.5 h. The resultant solution was evaporated
to dryness under vacuum, 30 ml of benzene containing
2.70 g of 2,2,2-trifluoroethanol and 1.19 g of pyridine
were added and the mixture was refluxed for 4 h. After
cooling the reaction solution, an aqueous solution (50 ml)
containing 6 mM sodium carbonate and deionized water
(100 ml) were successively employed four times and twice,
respectively, to extract the excess alcohol and remaining (S)-
naproxen. The organic layer was separated, dried over
magnesium sulfate, filtered and concentrated under vacuum.
After purification by silica gel chromatography with the
mobile phase n-hexane–ethyl acetate (2:1, v/v) and concen-
tration by vacuum, the desired (S )-naproxen 2,2,2-trifluor-
oethyl ester was obtained as a white powder and confirmed
by HPLC using authentic products from lipase-catalyzed
esterification of (R,S )-naproxen with 2,2,2-trifluoroethanol
in isooctane. 1H NMR spectra were also recorded at
400 MHz on a Bruker spectrometer in deuteriochloroform
solutions with tetramethylsilane as an internal standard.
RESULTS AND DISCUSSION
Racemization of (S)-naproxen ester in isooctane
The mechanism for the racemization of (S)-naproxen
2,2,2-trifluoroethyl ester involves ꢀ-proton abstraction
by the base to give a planar enolate. In most cases, the
ꢀ-proton abstraction is the rate-limiting step and the
mechanism can be expressed as follows:6
kint
(
kint
As
AR
ð1Þ
+
where AR and AS represent (R)- and (S)-naproxen 2,2,2-
trifluoroethyl ester, respectively. Therefore, the first-order
interconversion constant kint can be estimated from the
time-course data for eeS coupled with the theoretical
equation ln(eeS/eeS0) ¼ ꢁ 2tkint, where eeS0 and t are
the initial eeS and time, respectively.
Figure 1 demonstrates typical time-course variations of
ln(eeS/eeS0) at various DBU concentrations in isooctane
at 45 ꢂC, from which kint was determined and is presents
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 387–392

RACEMIZATION AND HYDROLYSIS OF (S)-NAPROXEN TRIFLUOROETHYL ESTER
389
Figure 1. Time-course variations of ln(eeS/eeS0) in isooctane
at 45 ꢂC for 1 mM (S)-naproxen 2,2,2-trifluoroethyl ester and
DBU concentrations of (*) 1, (!) 10, (&) 20 and (^) 30 mM
in Fig. 2(A) for illustration. As indicated by the linear
dependence of kint on DBU concentration, the second-
order interconversion constant as the slope of the straight
line, kiꢀnt ¼ 2.56ꢃ 10ꢁ2 hꢁ1 mMꢁ1, was further estimated.
Similar results to those shown in Fig. 1 for other neutral
bases were obtained (data not shown), from which the
first-order interconversion constants varied with the base
concentration as illustrated in Fig. 2(A) and (B), and
hence the second-order interconversion constants were
obtained as listed in Table 1.
In general, the stronger is the base in terms of the
equilibrium acidity pKA of the conjugated acid in acet-
onitrile (no available data in isooctane), the greater is
the second-order interconversion constant in isooctꢀane.
About a five orders of magnitude enhancement of kint is
obtained when TOA is replaced by MTBD, or P1 by P4.
Brønsted plots have been proposed to elucidate the nu-
cleophilic and/or general base catalysis of the acyl group
transfer reactions of esters and thioesters. Although the
mechanism of based-catalyzed racemization is completely
Figure 2. Variations of first-order interconversion constants
with base concentrations in isooctane for various strong
neutral bases: (A) (&) P4, (!) MTBD and (*) DBU; (B) (&)
P1-tris, (^) P1, (*) DABCO and (!) TOA
different from that of acyl transfer reaction, the Brønsted
correlations can be employed to investigate the effect of
the substrate and base on kiꢀnt. A good linear relationship,
Table 1. Effect of strong neutral bases on the second-order interconversion constants for the racemization of (S)-naproxen
2,2,2-trifluoroethyl ester and (S)-naproxen 2,2,2-trifluoroethyl thioester in isooctane at 45 ꢂC
Parameter
TOA
DABCO
DBU
MTBD
P1
P1-tris
P4
a
pKAb
pKip
18.04
12.63
18.29
12.85
24.33
17.80
25.44
18.83
8.902(ꢁ2)
ndf
26.88
18.80
2.262(ꢁ5)
nd
28.40
20.10
1.781(ꢁ4)
nd
42.70
nae
c
kiꢀnt
7.467(ꢁ7)
2.48(ꢁ4)
8.315(ꢁ6)
4.00(ꢁ3)
2.470(ꢁ2)
2.20(0)
4.459(0)
nd
d
kiꢀnt
a
In acetonitrile.7
pKip for TOA, DABCO or MTBD estimated from pKA ¼ 1.195pKip þ 2.94 (R2 ¼ 0.97) for non-phosphazene bases in tetrahydrofuran.8
b
c
d
e
f
ꢁ1
Units hꢁ1 mM for (S)-naproxen 2,2,2-trifluoroethyl ester. Values in parentheses are exponents, e.g. (ꢁ7) ¼ ꢃ 10ꢁ7
.
ꢁ1
Units as hꢁ1 mM for (S)-naproxen 2,2,2-trifluoroethyl thioester from Ref. 4. Values in parentheses are exponents (see footnote c).
na, Not available.
nd, Not done.
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 387–392

390
H.-Y. LIN ET AL.
log kiꢀnt ¼ 0.639pKA ꢁ 17.2 (R2 ¼ 0.98), was estimated for
TOA, DABCO, DBU and MTBD. For the extremely
hindered phosphazene bases P1, P1-tris and P4, the cor-
relation became log kiꢀnt ¼ 0.324pKA ꢁ 13.2 (R2 ¼ 0.99).
The different slopes and intercepts in the Brønsted cor-
relations for phosphazene and non-phosphazene bases
implies that aside from different hindrance effects of
strong neutral bases in abstracting the ꢀ-proton, the con-
tact ion-pair formation in isooctane, but not free ions in
polar solvents such as acetonitrile, might affect the ion-
pair formation of enolate intermediate with the base
and hence the kiꢀnt value during racemization. Therefore,
when the equilibrium ion-pair basicity in tetrahydrofuran
(Table 1) is regarded as that in isooctane,8 the correlations
become log kiꢀnt ¼ 0.768pKip ꢁ 15.4 (R2 ¼ 0.96) for TOA,
DABCO, DBU and MTBD and log kiꢀnt ¼ 0.689pKip ꢁ
17.6 for P1 and P1-tris (no data available pKip for P4),
respectively. The similar slopes of 0.768 and 0.689 imply
that an increase of one unit in pKip yields a similar
enhancement for kiꢀnt, regardless of whether a tertiary
amine, amidine, guanidine or phosphazene base is used.
Moreover, the lower value of ꢁ17.6 in comparison with
ꢁ15.4 for non-phosphazene bases reflects the extremely
hindered nature of phosphazene bases, giving about a two
orders of magnitude lower kiꢀnt at the same pKip value.
Obtaining more data using other neutral bases as the
racemization catalyst and/or enantiomers as the substrate
in isooctane to test the validity of the above Brønsted
correlations awaits further studies.
Table 2. Effect of non-polar solvents on water content
saturated, (H2O)sat, equilibrium constant for ion-pair forma-
tion between MTBD and water, Keq, and second-order
interconversion constants, k* , for racemization of (S)-
int
naproxen 2,2,2-trifluoroethyl ester with MTBD at 45 ꢂC
Isooctane
(H2O)sat (mM)a 16.2
n-hexane
Cyclohexane
15.0
0.62 (0.98)
7.66 (0.98)
16.8
0.30 (0.99)
6.57 (0.99)
ꢁ1 b
Keq (mM
)
0.16 (0.98)
8.90 (0.96)
k* ꢃ 102
int
(hꢁ1 mM
ꢁ1 b
)
a
Estimated from Table 2 in Ref. 13.
Values in parentheses are error estimates (R2).
b
to the value of 0.768 for (S)-naproxen 2,2,2-trifluoroethyl
ester.
Inspection on the lipase-catalyzed rate for the fast-
reacting (S)-2,2,2-trifluoroethyl ester in isooctane and the
racemization rate calculated by using the kiꢀnt values
(Table 1) indicates that MTBD and other stronger neutral
bases are good candidates for developing related dynamic
kinetic resolution processes.3b Although an order of
magnitude difference in the log P value (where P is the
partition coefficient of solvent between octanol and water)
between isooctane and n-hexane (or cyclohexane) has
been reported,11 very similar values of kiꢀnt using MTBD
as the base are given in Table 2. In contrast, almost the
same dielectric constant of 2.02 for the solvent is found.
Therefore, more experiments to check the possible cor-
relation between kiꢀnt and the dielectric constant of the
solvent are required.
Several scales of the equilibrium acidity in non-polar
solvents such as tetrahydrofuran and heptane have been
reported.9 The difficulty in measuring the ion-pair dis-
sociation constant has impeded further development of
the equilibrium ion-pair acidity scale.10 Similarly, only
recently was the equilibrium ion-pair basicity scale
anchored to the pKA value of 12.5 for triethylamine in
tetrahydrofuran.8 Isotope exchange of the ꢀ-proton in
non-polar solvents by base reagents might be employed
to measure the kinetic basicity of the base. However,
unless the exchange rate is much greater than the internal-
returning rate of the intermediate, the observed rate will
not reflect the kinetic basicity. With a careful selection
of chiral compounds, racemization of the enantiomer in
principle can overcome the limitation of the internal-
returning effect and be employed to construct the kinetic
basicity scale for strong neutral bases in non-polar organic
solvents.
The racemization of (S)-naproxen 2,2,2-trifluoroethyl
thioester in isooctane at 45 ꢂC by using TOA, DABCꢀO
and DBU as the base was carried out to estimate kint
values (Table 1).4 In general,ꢀmore than a one order of
magnitude enhancement of kint for (S)-naproxen 2,2,2-
trifluoroethyl thioester is found, which indicates that the
2,2,2-trifluoroethanethiol moiety plays an important role
in increasing the ꢀ-proton acidity during racemization.1b
A Brønsted correlation of log(kiꢀnt)thioester ¼ 0.659pKip ꢁ
11.37 (R2 ¼ 0.94) is obtained, where the slope is similar
Racemization and hydrolysis of
(S)-naproxen ester
When (S)-naproxen 2,2,2-trifluoroethyl ester is employed
as the model compound in non-polar solvents containing
water and strong neutral bases, ion-pair formation be-
tween the base and water and hydrolysis and racemiza-
tion of (R)- and (S)-naproxen 2,2,2-trifluoroethyl ester
may occur simultaneously.
Let us assume rapid ion-pair formation between the
base and water as follows (see Supplementary Material):
Keq
(
+
ð2Þ
B þ H2O
Bꢀ
One may express the base concentration (B) in terms of
the initial base concentration (B)0, initial water concen-
tration (H2O)0 and equilibrium constant Keq as
h
eq
ðBÞ ¼ 0:5 ðBÞ0 ꢁ ðH2OÞ0 ꢁ Keꢁq1 þ f½ðBÞ0 þ ðH2OÞ0
i
2
0:5
ꢁ1
þK ꢄ ꢁ 4ðBÞ0ðH2OÞ0g
ð3Þ
Three mechanisms, i.e. stepwise, dissociative and
concerted, have been employed to describe the acyl group
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 387–392

RACEMIZATION AND HYDROLYSIS OF (S)-NAPROXEN TRIFLUOROETHYL ESTER
391
transfer reactions of esters.12a The switchover from one
mechanism to another depends on the nature of the leav-
ing group and nucleophile. A stepwise mechanism has
been proposed for base-catalyzed ester hydrolysis in
aqueous solution or a polar organic solvent.12b–e Depend-
ing on the substituent of the acyl moiety, the leaving
ability of the alcoholic group and the structure of the
base, the stepwise mechanism via the formation of a
tetrahedral intermediate is furthermore classified into two
kinds, i.e. nucleophilic catalysis and general base cata-
lysis. It is well known that usually esters possessing a
good leaving group, such as 4-nitrophenyl or 2,4-dini-
trophenyl, are subject to nucleophilic catalysis.12d Since
the 2,2,2-trifluoroethyl moiety is a good leaving group,
MTBD-catalyzed nucleophilic catalysis is proposed for
the hydrolysis of (R)- and (S)-naproxen 2,2,2-trifluor-
oethyl ester in this work.
k2ðBꢀÞ ꢅ kꢁ1], the formation of IR1 and IS1 is the rate-
limiting step, giving khy ¼ k1(B) from Eqn (4). On the
other hand, if the expulsion of the base proceeds much
faster than that of the leaving group, the breakdown of IR1
and IS1 to IR2 and IS2 is rate limiting [i.e. kꢁ1 ꢅ k2(B*)],
giving khy ¼ k1k2(B)(B*)/k
.
ꢁ1
When the hydrolysis and racemization for both ester
substrates are considered at the initial stage, the balance
equations for both enantiomers were solved (see Supple-
mentary Material) and rearranged as
lnðeeS=eeS0Þ ¼ ꢁ2tkint
ð5Þ
ð6Þ
lnf½ðARÞ þ ðASÞꢄ=½ðARÞ0 þ ðASÞ0ꢄg ¼ ꢁkhyt
Therefore, the first-order interconversion constants at
various combinations of (B)0 and (H2O)0 can be deter-
mined from the time-course data of eeS coupled with
Eqn (5). One can further estimate Keq from the relation-
ship kint ¼ kiꢀnt(B) coupled with Eqn (3), where kiꢀnt
has been determined in anhydrous solvents. Similarly,
from the time-course data for (AR) þ (AS) coupled with
Eqn (6), the hydrolysis constants are estimated at various
combinations of (B)0 and (H2O)0, and then the kinetic
constants in Eqn (4) by using Eqn (3) and the relationship
(B*) ¼ [(B)0 ꢁ (B)].
As illustrated in Scheme 1, the hydrolysis reaction of
AR and AS passes through the first tetrahedral intermedi-
ates IR1 and IS1, and then the second ones intermediates
IR2 and IS2 to the products PR and PS. Unlike the nucleo-
philic hydrolysis of esters in polar solvents such that alkyl
oxide anions could leave, the requirement of contact ion-
pair between MTBD and the alcohol product in non-polar
solvents suggested that the reaction rate from IR1 to IR2
(or IS1 to IS2) was proportional to k2(B*).
By employing steady-state assumptions for the inter-
mediates IR1, IR2, IS1 and IS2 in Scheme 1, the hydrolysis
constant for the (R)- and (S)-ester can be derived as
follows (see Supplementary Material):
Similar data (not shown) to those illustrated Fig. 1 at
various combinations of initial concentrations of MTBD
and water in n-hexane, cyclohexane and isooctane were
obtained, and kint values determined. Good linear rela-
tionships between kint and (B) are illustrated in Fig. 3 for
all solvents. The equilibrium constants were then esti-
mated and are presented iꢀn Table 2. Unlike the saturated
water concentration and kint, the inert solvent has a slight
influence on the Keq value and hence the base and ion-pair
concentrations for a given combination of (B)0 and
(H2O)0.
khy ¼ k1ðBÞðBꢀÞ=½ðBꢀÞ þ kꢁ1ðkꢁ2 þ k3Þ=k2k3ꢄ ð4Þ
In comparison with PR and PS, the intermediates IS2 and
IR2 are not stable and will rapidly transform into the
product (i.e. k3 ꢅ kꢁ2). Moreover, if the 2,2,2-trifluor-
oethyl moiety departs from IR1 and IS1 to IR2 and IS2 much
faster than the base to the substrates AR and AS [i.e.
Similar behavior to that in Fig. 1 (data not shown) for
the time-course variations of ln{[(AR) þ (AS)]/[(AR)0 þ
Scheme 1
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 387–392

392
H.-Y. LIN ET AL.
await further studies, e.g. by employing esters contain-
ing an electron-withdrawing group on the acyl moiety or
an alcoholic leaving group without electron-withdrawing
groups in which the mechanism of nucleophilic hydro-
lysis may change.
Supplementary material
The equilibrium base and ion-pair concentrations, rate
equations for the racemization and hydrolysis of (R)- and
(S)-naproxen 2,2,2-trifluoroethyl ester in non-polar sol-
vents were derived by considering the ion-pair formation
between the base and water. This material is available in
Wiley Interscience.
Figure 3. Variations of first-order interconversion constant
with (B) using MTBD as the base: (A) in hexane containing
(*) no water, (~) (H2O)0 ¼ 15.0 mM and (&) (H2O)0 ¼
7.5 mM; (B) in isooctane containing (^) no water and (^)
(H2O)0 ¼ 16.2 mM; (C) in cyclohexane containing (!) no
water and (!) (H2O)0 ¼ 16.8 mM
Acknowledgments
The financial support of the Chinese National Science
Council, NSC 91-2214-E-006-025, is appreciated.
REFERENCES
1. (a) Stecher H, Faber K. Synthesis 1997; 29: 1–16; (b) Um PJ,
Drueckhammer DG. J. Am. Chem. Soc. 1998; 120: 5605–5610.
2. (a) Tan DS, Gunter MM, Drueckhammer DG. J. Am. Chem. Soc.
1995; 117: 9093–9094; (b) Chang CS, Tsai SW, Kuo J. Biotech-
nol. Bioeng. 1999; 64: 120–126; (c) Lin CN, Tsai SW. Biotechnol.
Bioeng. 2000; 69: 31–38; (d) Chen CY, Cheng YC, Tsai SW.
J. Chem. Technol. Biotechnol. 2001; 79: 200–210; (e) Lu CH,
Cheng YC, Tsai SW. Biotechnol. Bioeng. 2002; 79: 200–210.
3. (a) Dinh PM, Williams JMJ, Harris W. Tetrahedron Lett. 1999; 40:
749–752; (b) Lin HY, Tsai SW. J. Mol. Catal. B 2003; 24–25:
111–120.
4. Chen CY, Chang YS, Lin SA, Wen HI, Cheng YC, Tsai SW.
J. Org. Chem. 2002; 67: 3323–3326.
5. Allenmark S, Ohlsson A. Chirality 1992; 4: 98–102.
6. Eliel EL, Wilen SH. Stereochemistry of Organic Compounds.
Wiley: New York, 1994; 426.
7. (a) Coetzee JF, Padmanabhan GR. J. Am. Chem. Soc. 1965; 87:
5005–5011; (b) Schwesinger R, Schlemper H, Hasenfratz C,
Willaredt J, Dambacher T, Breuer T, Ottaway C, Fletschinger
M, Boele J, Fritz H, Putzas D, Rotter HW, Bordwell FG, Satish
AV, Ji G-Z, Peters E-M, Peters K, von Schnering HG, Walz L.
Liebigs Ann. Chem. 1996; 1055–1081.
8. Rodima T, Kaljurand I, Pihl A, Maemets V, Leito I, Koppel IA.
J. Org. Chem. 2002; 67: 1873–1881.
Figure 4. Variations of the hydrolysis constant with (B)(B*)
using MTBD as the base in (*) isooctane containing
(H2O)0 ¼ 16.2 mM and (~) hexane containing (H2O)0 ¼
15.0 mM
9. (a) Leito I, Rodima T, Koppel IA, Schwesinger R, Vlasov VM.
J. Org. Chem. 1997; 62: 8479–8483; (b) Facchetti A, Kim YJ,
Streitwieser A. J. Org. Chem. 2000; 65: 4195–4197.
10. Abdur-Rashid K, Fong TP, Greaves B, Gusev DG, Hinman JG,
Landau SE, Lough AJ, Morris RH. J. Am. Chem. Soc. 2000; 122:
9155–9171.
(AS)0]} in n-hexane and isooctane at various combinations
of initial MTBD and water concentrations were obtained.
Good linear relationships between khy and (B)(B*), but
not (B), are illustrated in Fig. 4, from which slopes (i.e.
11. Laane C, Boeren S, Vos K, Veeger C. Biotechnol. Bioeng. 1987;
28: 81–87.
12. (a) Marlier JF. Acc. Chem. Res. 2001; 34: 283–290; (b) Neuvonen
H. J. Chem. Soc., Perkin Trans. 2 1987; 159–167; (c) Stefanidis D,
Jencks WP. J. Am. Chem. Soc. 1993; 115: 6045–6050; (d)
Neuvonen H. J. Chem. Soc., Perkin Trans. 2 1995; 951–954; (e)
Um IH, Han HJ, Ahn JA, Kang S, Buncel E. J. Org. Chem. 2002;
67: 8475–8480.
ꢁ2
k1k2/kꢁ1) of 4.95 ꢃ 10ꢁ4 and 1.46 ꢃ 10ꢁ4 hꢁ1 mM for
isooctane and n-hexane, respectively, are determined.
No elucidations are made for the slightly higher slope
and the hydrolysis rate in isooctane. More experiments to
test the validity of the proposed hydrolysis mechanism
13. Maczynski A, Goral M, Wisniewska-Goclowaka B, Skrzecz A,
Shaw D. Monatshe. Chem. 2003; 134: 633–653.
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 387–392