Dynamics of reaction of ion pairs in aqueous solution: Racemization of the chiral ion pair intermediate of solvolysis of (S)-1-(4-methylphenyl)ethyl pentafluorobenzoate

Article abstract of DOI:10.1002/poc.611

The chiral ester (S)-1-(4-methylphenyl)ethyl pentafluorobenzoate, (S)-1-OC(O)C₆F₅, was prepared and a value of k _rac = 8.5 × 10^-7 s^-1 was determined for the rate constant for its racemization to (R)-1-OC(O)C₆F₅ in 50:50 (v/v) trifluoroethanol-water (I = 0.50, NaClO₄). This rate constant is 12 times smaller than k_solv = 1.06 × 10 ^-5 s^-1 for the stepwise solvolysis of 1-OC(O)C ₆F₅, which shows that inversion of the ion-pair intermediate is a relatively rare event during solvolysis in a largely aqueous solvent, and two times smaller than k_iso = 1.6 × 10 ^-6 s^-1 for degenerate isomerization which exchanges the position of the ester bridging and non-bridging oxygens of 1-OC(O)C ₆F₅. A simple relationship is derived between the empirical rate constant ratio k_rac/k_iso and the microscopic rate constants for reaction of the ion-pair intermediate within its solvation shell. Substitution of estimated rate constants for reactions of the 1-(4-methylphenyl)ethyl carbocation-pentafluorobenzoate anion pair into this equation gives k_i = 1.5 × 10¹⁰ s^-1 for inversion of the chiral ion pair, which is similar to k_-d = 1.6 × 10¹⁰ s^-1 for its separation to free ions. Copyright

Full text of DOI:10.1002/poc.611

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 2003; 16: 484–490
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/poc.611
Dynamics of reaction of ion pairs in aqueous solution:
racemization of the chiral ion pair intermediate of
solvolysis of (S)-1-(4-methylphenyl)ethyl
penta¯uorobenzoate
Yutaka Tsuji,1 Tetsuo Mori,1 Maria M. Toteva2 and John P. Richard2*
1Department of Biochemistry and Applied Chemistry, Kurume National College of Technology, Komorinomachi, Kurume 830-8555, Japan
2Department of Chemistry, University at Buffalo, SUNY, Buffalo, New York 14260-3000, USA
Received 28 October 2002; revised 12 December 2002; accepted 26 December 2002
ABSTRACT: The chiral ester (S)-1-(4-methylphenyl)ethyl pentafluorobenzoate, (S)-1-OC(O)C6F5, was prepared
and a value of krac = 8.5 Â 10À7 sÀ1 was determined for the rate constant for its racemization to (R)-1-OC(O)C6F5 in
50:50 (v/v) trifluoroethanol–water (I = 0.50, NaClO4). This rate constant is 12 times smaller than ksolv = 1.06 Â 10À5
sÀ1 for the stepwise solvolysis of 1-OC(O)C6F5, which shows that inversion of the ion-pair intermediate is a
relatively rare event during solvolysis in a largely aqueous solvent, and two times smaller than kiso = 1.6 Â 10À6 sÀ1
for degenerate isomerization which exchanges the position of the ester bridging and non-bridging oxygens of
1-OC(O)C6F5. A simple relationship is derived between the empirical rate constant ratio krac/kiso and the microscopic
rate constants for reaction of the ion-pair intermediate within its solvation shell. Substitution of estimated rate
constants for reactions of the 1-(4-methylphenyl)ethyl carbocation–pentafluorobenzoate anion pair into this equation
gives ki = 1.5 Â 1010
s
À1 for inversion of the chiral ion pair, which is similar to kÀd = 1.6 Â 1010
s
À1 for its separation
to free ions. Copyright  2003 John Wiley & Sons, Ltd.
KEYWORDS: reaction dynamics; chiral ion pairs; racemization; solvolysis; (S)-1-(4-methylphenyl)ethyl penta-
fluorobenzoate
INTRODUCTION
reactions that occur within the solvation shell of ion pair
intermediates of solvolysis:
The experimental protocols for the detection of the
reactions of carbocation–anion pair intermediates of
solvolysis reactions are well developed.1 However, there
is still great potential for using these protocols along with
modern analytical methods for product analysis to detect
ion-pair reactions during solvolysis. These experimental
results may then be combined with recent studies of ion
pair dynamics to obtain estimates for absolute rate
constants for reactions of ion pairs within their solvent
cages.2–5 For example, we have determined a value of
1. a value of kr = 1 Â 1011 sÀ1 found for reorganization
of the ion pair intermediate of solvolysis of ring-
substituted 1-phenylethyl thionobenzoates which is a
step in the O → S isomerization reaction of this
substrate;4
2. a value of kÀ1 = 7 Â 109 sÀ1 found for internal return
of the ion pair intermediate of solvolysis of 1-(4-
methylphenyl)ethyl pentafluorobenzoate to the neutral
ester, which is a step in the degenerate exchange
reaction of oxygen-18 label between the ester bridging
and non-bridging positions.5
k
Àd ꢀ 1.6 Â 1010 sÀ1 for the rate constant for separation
of ion pair intermediates to free ions during solvolysis of
ring-substituted 1-phenylethyl derivatives in 50:50 (v/v)
water–trifluoroethanol.6 This rate constant has served as a
‘clock’ to estimate the following rate constants for
Our set of rate constants for reactions of the ion-pair
intermediate of solvolysis of 1-(4-methylphenyl)ethyl
pentafluorobenzoate in 50:50 (v/v) water–trifluoroetha-
nol within the surrounding solvent cage is nearly
complete;4–6 all that is missing is the rate constant ki
for inversion of the chiral ion pair intermediate of
solvolysis of chiral 1-(4-methylphenyl)ethyl pentafluoro-
benzoate. We report here the synthesis (S)-1-(4-methyl-
phenyl)ethyl pentafluorobenzoate and the results of
experiments which show that this compound undergoes
*Correspondence to: J. P. Richard, Department of Chemistry,
University at Buffalo, SUNY, Buffalo, New York 14260-3000, USA.
E-mail: jrichard@acsu.buffalo.edu
Contract/grant sponsor: National Institutes of Health; Contract/grant
number: GM 39754.
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 484–490

DYNAMICS OF REACTION OF ION PAIRS IN AQUEOUS SOLUTION
485
racemization in 50:50 (v/v) water–trifluoroethanol 12-
times slower than solvolysis, but only two times slower
than scrambling of the 18O label between ester-bridging
and non-bridging positions determined in earlier work.5
Combining these rate data with results from earlier work
gives a value of ki = 1.5 Â 1010 sÀ1 for the rate constant
for interconversion of the S and R forms for the
carbocation–anion pair intermediate of solvolysis of
(S)-1-(4-methylphenyl)ethyl pentafluorobenzoate.
JEOL AL-400 FT-NMR spectrometer operating at
400 MHz. The chemical shifts are reported relative to a
value of ꢀ = 7.15 ppm for the solvent C6D6. The benzylic
protons of 1-OC(O)C6F5 and 1-OH were subjected to
homonuclear decoupling by irradiation of the corre-
sponding a-methyl protons during data acquisition.
Spectra (32 000 data points) were obtained using a
sweep width of 4000 Hz centered at the benzylic proton,
a 90° pulse angle, an acquisition time of 8 s and a 60 s
relaxation delay between pulses (200–400 transients) to
ensure complete relaxation of the benzylic proton.
Europium tris[3-(heptafluoropropylhydroxymethyl-
ene)-(‡)-camphorate (30 mM) and (S)-(‡)-2,2,2-tri-
fluoro-1-(9-anthryl)ethanol (0.1 M) were used as
chemical shift reagents to resolve the enantiomeric
protons of 1-OC(O)C6F5 and 1-OH, respectively. The
enantiomeric purities of 1-OC(O)C6F5 and 1-OH were
determined from the ratio of the integrated areas of the
spin-decoupled benzylic protons of the R- and S-
enantiomers determined for spectra in C6D6 that contains
the appropriate shift reagent.
EXPERIMENTAL
Materials. Pentafluorobenzoyl chloride, p-methylaceto-
phenone, europium tris[3-(heptafluoropropylhydroxy-
methylene)]-(‡)-camphorate and (S)-(‡)-2,2,2-trifluoro-
1-(9-anthryl)ethanol were purchased from Aldrich and
used without purification. All other chemicals were of
reagent grade and were used without purification.
1
Chemical syntheses. The chemical shifts of H NMR
spectra for routine characterization of the products of
chemical synthesis were referenced to CHCl3 at
7.27 ppm. Racemic 1-(4-methylphenyl)ethyl penta-
fluorobenzoate was prepared from 1-4-(methyl-
phenyl)ethyl alcohol and pentafluorobenzoyl chloride
by adaptation of a published method that is described in
greater detail below for the synthesis of the chiral ester.7
(S)-1-(4-Methylphenyl)ethyl alcohol [(S)-1-OH] was
prepared by reduction of p-methylacetophenone with
(À)-B-chlorodiisopinocamphenylborane [(À)-DIP-chlor-
ide].8 The crude reaction product was purified by silica
gel column chromatography eluting with hexane–diethyl
ether to give the final product in a yield of 90%. 1H NMR
(400 MHz, CDCl3), ꢀ 7.26, 7.15 (A2B2, 4H, J = 8 Hz,
C6H4Me) 4.86 (q, 1H, J = 6 Hz, CH), 2.34 (s, 3H,
ArCH3), 1.48 (d, 3H, J = 6 Hz, CH3).
Kinetic analyses. Solutions of (S)-1-OC(O)C6F5 in
50:50 (v/v) trifluoroethanol–water (I = 0.50, NaClO4)
were prepared by dissolving 30 mg of substrate in 1 ml of
acetonitrile and then adding this to 200 ml of 50:50 (v/v)
trifluoroethanol–water (I = 0.50, NaClO4) to give a final
substrate concentration of 0.5 mM. At a specified reaction
time the remaining substrate and reaction products were
extracted into 500 ml of diethyl ether and the ethereal
extract was washed with water, dried over MgSO4 and
evaporated. The substrate 1-OC(O)C6F5 was separated
from products 1-OH and 1-OTFE using a Recycling
Preparative LC-908 system from Japan Analytical
Industry, a JAIGEL-1H styrene polymer HPLC column
and eluting with CHCl3, and the solvent was stripped
from purified 1-OH and 1-OC(O)C6F5. These com-
pounds were dissolved in C6D6 and the appropriate shift
reagent was added for NMR analyses.
(S)-1-(Methylphenyl)ethyl pentafluorobenzoate [(S)-1-
OC(O)C6F5)] was prepared by adaptation of a published
procedure.7 (S)-1-OH (14.7 mmol) was reacted with 1.7
molar equiv. of pentafluorobenzoyl chloride in pyridine
with stirring at 0°C for 30 min, followed by 60 min at
room temperature. The reaction was quenched with cold
NaHCO3 (100 ml, 0°C) and the ester was extracted into
diethyl ether. The ethereal solution was washed with
dilute HCl, 5% aqueous NaHCO3 and saturated sodium
chloride and then dried over MgSO4. The crude reaction
product was purified by silica gel chromatography eluting
with 95:5 hexane–diethyl ether, and then recrystallized
twice from diethyl ether–hexane to give a final yield of
The rate constant for solvolysis of 1-OC(O)C6F5 was
determined by monitoring the disappearance of this
compound by HPLC.9 The pseudo-first-order rate con-
stant for the reaction was obtained from the slope of a
linear semi-logarithmic plot of reaction progress against
time.
RESULTS
A value of ksolv = 1.06 Â 10À5 sÀ1 for solvolysis of the
ester 1-OC(O)C6F5 at 25°C in 50:50 (v/v) trifluoroetha-
nol–water (I = 0.50, NaClO4) was determined by moni-
toring the disappearance of this compound by HPLC.
1
7%; m.p. 34–34.2°C; H NMR (400 MHz, CDCl3), ꢀ
7.31, 7.14 (A2B2, 4H, J = 8 Hz, C6H4Me), 6.12 (q, 1H,
J = 6 Hz, CH), 2.35 (s, 3H, ArCH3), 1.67 (d, 3H, J = 6 Hz,
CH3). Anal. (C16H11F5O2) C,H. Calc: H (3.36%), C
(58.19%) Found H (3.35%), C (58.25%).
1
Figure 1(A) shows the partial H NMR spectrum in
C6D6 of the benzylic proton of (S)-1-OC(O)C6F5 in
which coupling to the methyl protons has been eliminated
by using an inverse gated decoupling procedure (see
1
1H NMR analyses. H NMR spectra were recorded on a
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 484–490

486
Y. TSUJI ET AL.
Table 1. Normalized areas of the 1H NMR signals due to the
benzylic carbon during reaction of (S)-1-OC(O) C6F5 in 50:50
(v/v) tri¯uoroethanol±water at 25°C
Anor
b
c
c
Time (s)
(AR/AS)a
(Anor T
)
(Anor S
)
(Anor)R
1.00
43200
81000
129600
172800
0.0274
0.0846
0.106
0.14
0.633
0.426
0.253
0.16
0.616
0.392
0.229
0.14
0.0169
0.034
0.0242
0.0197
a
Ratio of the areas of the peaks for the benzylic protons of (R)-1-
OC(O)C6F5 and (S)-1-OC(O)C6F5 determined using a europium tris[3-
(heptafluoropropylhydroxymethylene)]-(‡)-camphorate shift reagent.
b
Normalized sum (Anor)T = (AR ‡ AS) of the 1H NMR peak areas due to the
benzylic carbon of (R)-1-OC(O)C6F5 and (S)-1-OC(O)C6F5, calculated as
(Anor)T = eÀkt, where k = ksolv = 1.06 Â 10À5 sÀ1 for solvolysis of the ester.
c
Normalized areas for the benzylic protons of (S)-1-OC(O)C6F5 [(Anor)S]
and (R)-1-OC(O)C6F5 [(Anor)R] calculated from (Anor)T = (AR ‡ AS) and
(AR/AS).
time t which is calculated as (Anor)T = eÀkt, where
k = ksolv = 1.06 Â 10À5
s
À1; (2) the normalized areas for
(S)-1-OC(O)C6F5 [(Anor)S] and (R)-1-OC(O)C6F5
[(Anor)R], which were determined from the total normal-
ized area (A)T = (AR ‡ AS) and the ratio (AR/AS).
h
i
ꢁAnor† ˆ 0:5 eÀk À eÀꢁ2k
ꢁ1†
solvt
‡ksolv†t
rac
Figure 1. Partial NMR spectra of the benzylic proton of 1-
OC(O)C6F5 in C6D6 in which coupling to the methyl protons
has been eliminated by using an inverse gated decoupling
procedure. (A) Spectrum of (S)-1-OC(O)C6F5 in the absence
of shift reagent. (B) Spectrum of (S)-1-OC(O)C6F5 in the
presence of 30 mM europium tris[3-(hepta¯uoropropylhy-
droxymethylene)]-(‡)-camphorate shift reagent. (C) Spec-
trum of 1-OC(O)C6F5 recovered after solvolysis for 36 h in
50:50 (v/v) tri¯uoroethanol±water at 25°C, also in the
presence of 30 mM shift reagent
R
Figure 2 shows the change in (Anor)R with time for the
reaction of (S)-1-OC(O)C6F5 at 25°C in 50:50 (v/v)
trifluoroethanol–water. The solid line shows the fit of the
data to Eqn. (1) derived for Scheme 1,10 using
ksolv = 1.06 Â 10À5 sÀ1 determined by experiment and
krac = 8.5 Â 10À7 sÀ1 determined by non-linear least-
squares analysis. The upper and lower dashed lines show
the fits obtained using values of krac that are 20% larger
Experimental). Figure 1(B) shows that 30 mM europium
tris[3-(heptafluoropropylhydroxymethylene)-(‡)-cam-
phorate causes about a 0.2 ppm downfield shift in the
decoupled signal for this benzylic proton. Figure 1(C)
shows the spectrum, determined in the presence of the
shift reagent, of 1-OC(O)C6F5 remaining after solvolysis
for 36 h at 25°C in 50:50 (v/v) trifluoroethanol–water at
(I = 0.50, NaClO4). Two signals are now observed, a
major peak for the starting substrate (S)-1-OC(O)C6F5
and a minor peak for (R)-1-OC(O)C6F5 which forms by
racemization of 1-OC(O)C6F5 during the course of its
solvolysis reaction.
Table 1 lists the ratio of the peak areas (AR/AS) for the
R- and S-isomers of 1-OC(O)C6F5 determined at four
different times during the solvolysis of (S)-1-OC(O)C6F5
in 50:50 (v/v) trifluoroethanol–water (I = 0.50, NaClO4).
Table 1 also lists the following normalized areas: (1) the
Figure 2. Time course for the appearance of the peak for the
benzylic proton of (R)-1-OC(O)C6F5 during solvolysis of
(S)-1-OC(O)C6F5 in 50:50 (v/v) tri¯uoroethanol±water at
25°C. The peak areas are normalized relative to a value of
1.0 for the benzylic proton of (S)-1-OC(O)C6F5 at t = 0. The
solid line shows the ®t of the data to ÀE1qn. (1) derived for
Scheme 1 using ksolv = 1.06 Â 10À5
s
and krac = 8.5 Â
10À7
s
À1. The upper and lower dashed lines show the ®ts
normalized area (Anor T
)
for total remaining 1-
obtained using values of krac that are 20% larger and 20%
smaller, respectively, than the best ®t value from least-
squares analysis
OC(O)C6F5 after solvolysis at 25°C in 50:50 (v/v)
trifluoroethanol–water (I = 0.50, NaClO4) at reaction
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 484–490

DYNAMICS OF REACTION OF ION PAIRS IN AQUEOUS SOLUTION
487
Experimental). Figure 3(B) shows that 0.1 M (S)-(‡)-
2,2,2-trifluoro-1-(9-anthryl)ethanol causes about
a
0.1 ppm upfield shift in the decoupled signal for the
benzylic proton. Figure 3(C) shows the spectrum of
1-OH isolated as the major product of a 183 h reaction of
(S)-1-OC(O)C6F5 (ꢂ10 half-times) at 25°C in 50:50
(v/v) trifluoroethanol–water (I = 0.50, NaClO4). Separate
signals are observed for (S)-1-OH and (R)-1-OH and the
ratio of the integrated areas of the signals for the S- and
R- isomers is 1.0.
Scheme 1
and 20% smaller, respectively, than the best fit value
from least-squares analysis. This shows that the variation
in the values of krac that provide an acceptable fit to these
data is Æ 20%.
DISCUSSION
Reaction conditions
The partial 1H NMR spectrum in C6D6 of the benzylic
proton of (S)-1-OH is shown in Fig. 3(A). Coupling to
the methyl protons was eliminated in this spectrum by
using an inverse gated decoupling procedure (see
The high instability and short lifetimes of carbocation–
anion pairs in aqueous solution create difficult problems
for the design of experiments to determine rate constants
for their reorganization in this solvent:
1. The products of reorganization of ion pairs can be
observed experimentally when the product is ‘trapped’
by internal return to reactant, but this requires that
trapping be very fast and competitive with separation
of the ion pair to free ions (kÀd = 1.6 Â 1010
s
À1). We
have presented strong evidence in earlier work that
internal return of the 1-(4-methylphenyl)ethylpenta-
fluorobenzoate ion intermediate of solvolysis of (S)-1-
OC(O)C6F5 occurs at a rate similar to its separation of
free ions.5,6
2. It is difficult to determine rate constants of >1010 sÀ1
by direct methods. However, it is possible to use the
value of kÀd = 1.6 Â 1010 sÀ1 for ion separation as a
clock for other reactions of these ion pairs which are
steps in the degenerate isomerization and racemization
reactions of neutral substrates.2,4,5
3. The neutral organic substrates used as precursors to
carbocation–anion pairs generally show low solubility
in aqueous solution. The good sensitivity of modern
high-resolution NMR spectroscopy in comparison
with polarimetric methods allowed us to work at a
low initial concentration 0.5 mM (S)-1-OC(O)C6F5,
which is soluble in 50:50 (v/v) water–trifluoroethanol
at 25°C. By comparison, a substrate concentration of
ꢂ40 mM was used in earlier studies in 60–90%
aqueous acetone of the reactions of ion pair inter-
mediates of chiral and O18-labeled 1-(4-methoxy-
phenyl)ethyl 4-nitrobenzoate.11
Figure 3. Partial NMR spectra of the benzylic proton of 1-OH
in C6D6 in which coupling to the methyl protons has been
eliminated by using an inverse gated decoupling procedure.
(A) Spectrum of (S)-1-OH in the absence of shift reagent. (B)
Spectrum of (S)-1-OH in the presence of 0.1 M (S)-(‡)-2,2,2-
tri¯uoro-1-(9-anthryl)ethanol. (C) Spectrum of 1-OC(O)C6F5
recovered after solvolysis for 183 h in 50:50 (v/v) tri¯uoro-
ethanol±water at 25°C, also in the presence of the shift
reagent
Racemization of (S)-1-OC(O)C6F5 during solvolysis
Solvolysis of 1-OC(O)C6F5 in 50:50 (v/v) trifluoroetha-
nol–water proceeds by a stepwise reaction mechanism.9
Figure 1(C) shows the partial NMR spectrum determined
in the presence of a chemical shift reagent of the substrate
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 484–490

488
Y. TSUJI ET AL.
recovered during the partial solvolysis reaction of (S)-1-
OC(O)C6F5 in this solvent. The appearance of a peak for
the benzylic proton of (R)-1-OC(O)C6F5 close to that for
the benzylic proton of (S)-1-OC(O)C6F5 shows that this
substrate undergoes racemization during the course of the
stepwise solvolysis reaction. A value of krac = 8.5 Â 10À7
sÀ1 for the rate constant for racemization of (S)-1-
OC(O)C6F5 (Scheme 1) was determined from the non-
linear least-squares fit of the experimental data from Fig.
2 to Eqn. (1), derived for Scheme 1. Racemization and
solvolysis of (S)-1-OC(O)C6F5 are assumed to proceed
through a common carbocation reaction intermediate,
because there is no precedent for the coupling of three
processes—bond cleavage, carbon inversion and bond
resynthesis—required for inversion of (S)-1-OC(O)C6F5
by a concerted reaction mechanism.12
ethanol–water, because of the stronger intermolecular
electrostatic interactions in the solvent of lower dielectric
constant and this allows for more substantial migration of
4-nitrobenzoate ion anion leaving group between the si
and re faces of the 1-(4-(methoxyphenyl)ethyl carboca-
tion.
Absolute rate constant for inversion of a chiral ion
pair
The value of kiso = 1.6 Â 10À6 sÀ1 determined in earlier
work for the scrambling of 18O label between ester-
bridging and non-bridging positions at 1-OC(O)C6F5 is
about two times larger than krac = 8.5 Â 10À7 sÀ1 for the
racemization of (S)-1-OC(O)C6F5 determined in this
work.5 Both racemization and 18O scrambling require
ionization of (S)-1-OC(O)C6F5 to form a carbocation–
anion pair intermediate (k1, Scheme 2) and resynthesis of
1-OC(O)C6F5 with an estimated rate constant of
This value of krac = 8.5 Â 10À7 sÀ1 for inversion of
(S)-1-OC(O)C6F5 is 12 times smaller than ksolv
=
1.06 Â 10À5 sÀ1 for the solvolysis reaction of this
substrate. This shows that inversion of the ion-pair
reaction intermediate followed by collapse to regenerate
neutral substrate is a relatively uncommon event during
solvolysis. The strong electrostatic interactions, which
favor ‘conducted-tour’ type mechanisms of ion pairs in
organic solvents,13,14 are strongly attenuated in water and
these ion pairs ‘seldom’ remain together long enough to
allow for migration of the anion between the carbocation
si and re faces. The result is that ion-pair rearrangement
in a mostly aqueous solvent is difficult to detect because
it requires that the reactivity of the ion pair fall within a
window where both its inversion and its internal return to
neutral product are competitive with separation to free
k
À1 = 7 Â 109 sÀ1.5 This two-fold difference in kiso and
krac is due to the difference in the very large rate constant
kr = 1 Â 1011
s
À1 for the fast ‘reorganization’ of the ion-
pair intermediate which exchanges the position of the
reactive carboxylate oxygens during the 18O scrambling
reaction of 1-18OC(O)C6F5,4,5 and the smaller rate
constant ki for the slower inversion of the ion pair
(Scheme 2) during the racemization of (S)-1-
OC(O)C6F5.
k1kik
À1
krac
ˆ
ˆ
ˆ
ꢁ2†
ꢁ3†
ꢁ4†
0
0
ꢁk ‡ k ‡ ks †ꢁ2ki ‡ k ‡ k ‡ ks †
À1
Àd
À1
Àd
ions (kÀd = 1.6 Â 1010
s
À1).
k1k
There are, by comparison, better opportunities to
detect and monitor the reactions of ion pairs in organic
solvents, because their separation to free ions is slower
À1
kiso
krac
0
2ꢁk ‡ k ‡ ks †
À1
Àd
than in water.1 For example, similar values of ksolv
=
2ki
0
1.0 Â 10À6 sÀ1 and krac = 1.2 Â 10À6 sÀ1 were reported
for the solvolysis and racemization of chiral 1-(4-
methoxyphenyl)ethyl 4-nitrobenzoate in 90% aqueous
acetone at 60°C.11 The half-time for separation of the
ion-pair intermediate to free ions in this mostly organic
solvent is longer than for separation in 50:50 trifluoro-
À1
Àd
kiso 2ki ‡ k ‡ k ‡ ks
The relationship between krac = 8.5 Â 10À7 sÀ1 and the
microscopic rate constants from Scheme 2 is given by
Eqn. (2), and the relationship between kiso = 1.6 Â 10À6
and the microscopic rate constants for a scheme similar to
Scheme 2
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 484–490

DYNAMICS OF REACTION OF ION PAIRS IN AQUEOUS SOLUTION
489
Scheme 2, except that inversion step (ki) has been
substituted by ion-pair ‘reorganization’ (kr = 1 Â 1011
(R)-1-phenylethanol in 18O-labeled water result from
direct addition of 18O-labeled water, and only a barely
detectable 1–3% result from migration of the 16O-labeled
water leaving group from the re to the si face of the
carbocation followed by collapse of the ion–dipole pair to
form (S)-1-phenylethanol.16 Similarly, internal return to
reactant that accompanies the stepwise solvolysis of
1-phenylethyl 4-nitrobenzoate in 70% aqueous acetone
does not result in detectable racemization of the
remaining ester.11
s
À1), is given by Eqn. (3).5 Equation (3) was derived with
the assumption that kr ꢃ k's ‡ kÀd.5 Equation (4) gives
the relationship between the rate constant ratio krac
/
kiso = 0.53 and the rate constants for partitioning of the
ion pair reaction intermediate from Scheme 2. Equation
(4) shows that racemization and isomerization would
have proceeded at similar rates if inversion of the ion pair
were very much faster than other reactions of the ion pair
(2ki ꢃ k's ‡ k ‡ kÀ1). Our observation that krac/kiso
Àd
drops to
a
value of about 0.50 requires that
2ki ꢀ k's ‡ k ‡ k
.
k
Stereochemical course of solvolysis
Àd
À1
Substitution of
Àd = 1.6 Â 1010 sÀ1 (Ref. 6),
k's = 6 Â 109 sÀ1 (Ref. 6) and kÀ1 = 7 Â 109 sÀ1 (Ref. 5)
The selectivity for reaction of 1-OC(O)C6F5 with
methanol and trifluoroethanol, kMeOH/kTFE = 5.3, in a
solvent of 50:45:5 water–trifluoroethanol-methanol is
significantly smaller than the selectivity determined for
the acid-catalyzed reaction of 1-OH, kMeOH/kTFE = 7.5,6
under the same conditions. This result shows that
partitioning of the ion pair is influenced by the presence
of the pentafluorobenzoate leaving group. The smaller
value of kMeOH/kTFE for reaction of 1-OC(O)C6F5
compared with reaction of 1-OH is consistent with more
effective general base catalysis by the pentafluorobenzo-
ate leaving group of addition of trifluoroethanol than
addition of methanol to the 1-(4-methylphenyl)ethyl
carbocation reaction intermediate, which was observed in
earlier studies on bimolecular catalysis by alkyl carbox-
ylate ions of the addition of methanol and trifluoroethanol
to the 1-(4-methoxyphenyl)ethyl carbocation.17,18 This
pentafluorobenzoate leaving group is also expected to
provide steric hindrance to frontside addition of sol-
vent.19,20 Therefore, the observation that essentially
racemic 1-OH is obtained from the reaction of (S)-1-
OC(O)C6F5 in 50:50 (v/v) water–trifluoroethanol sug-
gests that steric hindrance by the pentafluorobenzoate
leaving group to frontside addition of solvent water to
form products of retained configuration is balanced by
pentafluorobenzoate leaving group catalysis of the front-
side water, so that there are similar values for the rate
constants for these reactions.
into Eqn. (4) gives ki = 1.5 Â 1010
s
À1 for inversion of the
ion pair in 50:50 (v/v) water–trifluoroethanol. In other
words, the rate constant for separation of this ion pair to
free ions, which involves complete loss of intermolecular
ionic interactions (kÀd ꢀ 1.6 Â 1010
s
À1),6 is similar to
the rate constant for inversion of the ion pair, which
involves substantial reorganization of the ions but
maintains the intermolecular interactions between these
ions.
The value of krac/kiso = 0.54 reported for the reaction of
chiral and 18O-labeled 1-(4-methoxyphenyl)ethyl 4-
nitrobenzoate in 90% aqueous acetone is similar to the
value krac/kiso = 0.53 determined here for the reaction of a
different chiral and 18O-labeled substrate in a different
solvent. By comparison, the same change in substrate and
reaction conditions results in about a 12-fold decrease in
the values of both krac/ksolv (see above) and kiso/ksolv for
the reaction of the 1-OC(O)C6F5 in 50:50 (v/v)
trifluoroethanol–water. A detailed analysis of the sig-
nificance of these results is not possible because of
uncertainties about the absolute rate constants for
reactions of the ion pair intermediate of solvolysis of
1-(4-methoxyphenyl)ethyl 4-nitrobenzoate in 90% aque-
ous acetone. However, they suggest that the relative
values of the absolute rate constants for the reactions of
ion pairs that lead to 18O scrambling and inversion show a
low sensitivity to these specific changes in solvent and
substrate structure.
Inversion of the ion pair becomes more difficult and
then impossible to detect as the carbocation is stabilized
Acknowledgements
and k
becomes very much smaller than k
s
=
À1
Àd
1.6 Â 1010
À1. Similarly, inversion of the ion pair
This work was supported by grant GM 39754 from the
National Institutes of Health. It is important to acknowl-
edge that these experiments and much other work resulting
in a wealth of important data are the result of the untiring
efforts of Shinjiro Kobayashi to foster interactions
between Japanese chemists and their international collea-
gues through the organization of numerous conferences at
Kyushu University. The authors wish to dedicate this
paper to Professor Kobayashi in order to express their
gratitude to him for making this collaboration possible,
and for many other kindnesses large and small.
becomes more difficult and then impossible to detect as
the carbocation is destabilized and k's increases to the
point where the ion pair undergoes quantitative addition
of solvent before significant inversion can occur (k's
ꢃki). This is the case for solvolysis of (R)-1-phenyletha-
nol, which proceeds by a stepwise mechanism, through a
1-phenylethyl carbocation intermediate which is captured
by water with an estimated rate constant of k's ꢀ 1 Â 1011
s .
À1 9,15 This 1-phenylethyl carbocation is so reactive that
nearly all of the products of its formation from unlabelled
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 484–490

490
Y. TSUJI ET AL.
11. Goering HL, Briody RG, Sandrock G. J. Am. Chem. Soc. 1970; 92:
REFERENCES
7401–7407.
12. Dewar MJS. J. Am. Chem. Soc. 1984; 106: 209–219.
13. Chu KC, Cram DJ. J. Am. Chem. Soc. 1972; 94: 3521–3531.
14. Wong SM, Fischer HP, Cram DJ. J. Am. Chem. Soc. 1970; 93:
2235–2243.
1. Harris JM. Prog. Phys. Org. Chem. 1974; 11: 89–173.
2. Richard JP. J. Org. Chem. 1992; 57: 625–629.
3. Richard JP, Yeary PE. J. Am. Chem. Soc. 1993; 115: 1739–1744.
4. Richard JP, Tsuji Y. J. Am. Chem. Soc. 2000; 122: 3963–3964.
5. Tsuji Y, Mori T, Richard JP, Amyes TL, Fujio M, Tsuno Y. Org.
Lett. 2001; 3: 1237–1240.
6. Richard JP, Jencks WP. J. Am. Chem. Soc. 1984; 106: 1373–1383.
7. Richard JP, Amyes TL, Bei L, Stubblefield V. J. Am. Chem. Soc.
1990; 112: 9513–9519.
8. Brown HC, Chandrasekharan J, Ramachandran PV. J. Am. Chem.
Soc. 1988; 110: 1539–1546.
9. Richard JP, Rothenberg ME, Jencks WP. J. Am. Chem. Soc. 1984;
106: 1361–1372.
10. Alberty RA, Miller WG. J. Phys. Chem. 1957; 26: 1231–1237.
15. Amyes TL, Richard JP, Novak M. J. Am. Chem. Soc. 1992; 114:
8032–8041.
16. Merritt MV, Anderson DB, Basu KA, Chan I-W, Cheon H-J,
Mukundan NE, Flannary CA, Kim AY, Vaishampayan A, Yens
DA. J. Am. Ch. Soc. 1994; 116: 5551–5559.
17. Richard JP, Jencks WP. J. Am. Chem. Soc. 1984; 106: 1396–1401.
18. Ta-Shma R, Jencks WP. J. Am. Chem. Soc. 1986; 108: 8040–8050.
19. Grunwald E, Heller A, Klein FS. J. Chem. Soc. 1957; 2604–2613.
20. Muller P, Rossier J-C. J. Chem. Soc., Perkin Trans. 2 2000; 2232–
2237.
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 484–490