The effect of solvent polarity on the rate of the Mitsunobu esterification reaction

Article abstract of DOI:10.1016/j.tet.2015.04.035

The rate of the Mitsunobu esterification reaction of ethanol or isopropanol with benzoic acid was found to be much faster in non-polar solvents. The logarithm of the rate constant was inversely proportional to the solvent polarity, as defined by E_T values. Typically, the rate constant for ethyl benzoate formation in THF was 100 times greater than that in MeCN. The presence of either sodium benzoate or excess benzoic acid resulted in a decrease in rate. Each of the main species involved in the Mitsunobu esterification reaction, the alcohol starting material, dialkoxyphosphorane, alkoxyphosphonium salt and ester product, was detected by proton NMR analysis. The possible role of ion pair aggregates or clusters, prior to rate-determining S_N2 attack of carboxylate on the alkoxyphosphonium ion, is discussed. An explanation is provided as to why the yield in the Mitsunobu reaction is often higher in non-polar solvents.

Full text of DOI:10.1016/j.tet.2015.04.035

Tetrahedron xxx (2015) 1e7
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
The effect of solvent polarity on the rate of the Mitsunobu
esterification reaction
y
*
David Camp, Peta J. Harvey , Ian D. Jenkins
Eskitis Institute, Griffith University, Brisbane, QLD, 4111, Australia
a r t i c l e i n f o
a b s t r a c t
Article history:
The rate of the Mitsunobu esterification reaction of ethanol or isopropanol with benzoic acid was found
to be much faster in non-polar solvents. The logarithm of the rate constant was inversely proportional to
the solvent polarity, as defined by ET values. Typically, the rate constant for ethyl benzoate formation in
THF was 100 times greater than that in MeCN. The presence of either sodium benzoate or excess benzoic
acid resulted in a decrease in rate. Each of the main species involved in the Mitsunobu esterification
reaction, the alcohol starting material, dialkoxyphosphorane, alkoxyphosphonium salt and ester product,
was detected by proton NMR analysis. The possible role of ion pair aggregates or clusters, prior to rate-
determining SN2 attack of carboxylate on the alkoxyphosphonium ion, is discussed. An explanation is
provided as to why the yield in the Mitsunobu reaction is often higher in non-polar solvents.
Crown Copyright Ó 2015 Published by Elsevier Ltd. All rights reserved.
Received 16 January 2015
Received in revised form 19 March 2015
Accepted 13 April 2015
Available online xxx
Keywords:
Mitsunobu
Rate
Ion-pair
Solvent-polarity
Mechanism
1. Introduction
Hughes and Reamer33 reported that the betaine 1, reacts with
carboxylic acids to form acylhydrazines, and that this reaction was
faster in more polar solvents.
The Mitsunobu reaction was discovered 47 years ago1,2 and re-
mains one of the most useful reactions in organic synthesis, par-
ticularly for the esterification of alcohols where inversion of
configuration of the hydroxyl group is required. A number of re-
views are available.3e12 A SciFinder search of ‘Mitsunobu Reaction’
gave more than 2600 hits, with over half of these in the last 10
years. Although the mechanism of the Mitsunobu reaction (Scheme
1) has been the subject of a large number of investigations13e47
including DFT,46 as far as we are aware there have been no spe-
cific studies of the effect of solvent polarity on the rate of the
Mitsunobu esterification reaction. Yet, the solvent can sometimes
have a dramatic effect on the outcome of the Mitsunobu reaction.
For example, Dodge et al.48 have shown that the Mitsunobu in-
version of menthol with 4-nitrobenzoic acid in THF gives an 83%
yield, but when the reaction is carried out in CH2Cl2 the yield is only
3%. Similarly, Loibner and Zbiral49 obtained a 73% yield of an
inverted steroid benzoate when the reaction was carried out in
benzene, but no product was observed in THF. Hughes et al.22 have
studied the rate of the Mitsunobu esterification reaction as a func-
tion of acid pKa, but in only one solvent (CH2Cl2). In a later study,
In this paper, we report a study of the relative rate of the Mit-
sunobu esterification reaction as a function of solvent polarity. The
choice of solvent and its effect upon the rate of a homogeneous
chemical reaction can often provide an insight into the mechanistic
pathway of the reaction.
2. Results and discussion
A simple system was chosen for this study, the esterification of
ethanol (or isopropanol) by benzoic acid with triphenylphosphine/
diisopropyl azodicarboxylate (TPP/DIAD). It should be noted that
the acids of choice for the stereochemical inversion/esterification of
hindered secondary alcohols are 4-nitrobenzoic acid27,36 and
chloroacetic acid.27,50 We chose benzoic acid for this study to en-
sure that the rate-determining step was the final SN2 displacement
(Scheme 1). With stronger acids, especially when excess acid is
used, the alcohol activation step (1/2, Scheme 1) can become rate-
determining.22
The procedure initially followed (Protocol A) involved the
dropwise addition of DIAD (0.5 mmol) to a pre-mixed solution of
TPP (0.5 mmol), benzoic acid (0.34 mmol), benzophenone
(0.08 mmol as internal standard) and alcohol (0.34 mmol) in dry
solvent (5 mL) at 0 ꢀC under a nitrogen atmosphere. At recorded
time intervals, small aliquots were removed from the reaction
mixture and added to a known volume of an acetonitrile/water
* Corresponding author. Tel.: þ61 (07) 3735 6025; fax: þ61 (07) 3735 6001;
e-mail address: i.jenkins@griffith.edu.au (I.D. Jenkins).
y
Present address: Institute of Molecular Bioscience, University of Queensland,
Australia.
http://dx.doi.org/10.1016/j.tet.2015.04.035
0040-4020/Crown Copyright Ó 2015 Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Camp, D.; et al., Tetrahedron (2015), http://dx.doi.org/10.1016/j.tet.2015.04.035

2
D. Camp et al. / Tetrahedron xxx (2015) 1e7
scales such as Z-values53 but that there is no correlation whatsoever
between these two empirical parameters and the dielectric
constant.
Table 2 presents the half-lives for alkyl benzoate synthesis and
demonstrates the influence of temperature and solvent. The entries
are arranged in order of increasing solvent polarity (as listed in
Table 1) and the results indicate that ester formation is very fast
regardless of reaction medium. Due to the speed and complexity of
the reaction, a detailed investigation of the reaction kinetics and
the determination of absolute rate constants was not attempted.
However, a comparison of the half-lives clearly shows that an in-
crease in solvent polarity results in a significant reduction in the
rate of esterification. In chloroform at 0 ꢀC, for example, the syn-
thesis of ethyl benzoate was essentially complete within 1 min of
DIAD being added. Use of acetonitrile, on the other hand, resulted
in a decrease in reaction rate to give a half-life of several minutes. At
0
ꢀC, ethyl benzoate synthesis in acetonitrile is thusꢁ15 times
slower than in chloroform.
Replacement of ethanol as substrate by a secondary alcohol
resulted in a significant rate decrease as expected. Hence, the re-
action half-lives at 0 ꢀC in both chloroform and acetonitrile were
increased by a factor of>20 when isopropanol was used. This ob-
servation is consistent with reports in the literature25,26,54e58 that
regioselective esterification is generally obtained with poly-
hydroxylic compounds. For example, the primary position of 1,3-
diols is the least hindered and therefore, the favored reaction site.59
The data for the esterification of the more sterically hindered
isopropanol followed the same trend observed for ethanol (i.e.,
a decrease in esterification rate with increasing solvent polarity, as
indicated by ET values, was evident). The major difference noted
was when each alcohol was also used as the solvent. With ethanol
as solvent/reactant, esterification was very much slower than when
isopropanol was the solvent/reactant. We attribute this to a com-
bination of a higher solvent polarity (ET 51.9 vs 48.4) and lower pKa
(15.85 vs 16.48)60 for ethanol compared with isopropanol. Both
factors would reduce the rate of the reaction. Hughes et al.22 have
shown how sensitive the reaction is to the pKa of the acid com-
ponent and the amount of acid present. In the more acidic ethanol,
the nucleophilicity of the benzoate ion would be reduced, thereby
slowing down the SN2 step of the reaction (5/6, Scheme 1).
Scheme 1. Mechanism of the Mitsunobu reaction.
solution to quench the reaction prior to analysis by HPLC. A stan-
dard curve was created using mixtures of benzophenone and ethyl
(or isopropyl) benzoate at known concentrations. From this cali-
bration plot, a linear relationship was calculated between the in-
tegral ratio and the molar ratio of the product versus internal
standard, thus providing an indication of the experimental yield at
a given reaction time. Half-lives were estimated by plotting the
experimental yield against reaction time (for very fast reactions) or
from the apparent first-order rate constants when these could be
determined.
A total of eight different solvents were studied (Table 1).
Table 2
The dielectric constant (εR) provides a rough measure of a sol-
vent’s polarity and ability to dissolve ionic salts, but is generally not
as useful for organic reactions as the more comprehensive ET sol-
vent polarity scale determined by Dimroth and co-workers.52 The
ET scale is an empirical scale of solvent polarity derived from the
change of color of a solvatochromic dye in response to solvent
polarity. This empirical scale and others such as Kosower’s Z-
values,53 generally provide a more useful scale of solvent polarity
than any one physical characteristic, such as the dielectric constant.
Note that the order of ET values correlates well with other empirical
Influence of temperature and solvent on ester synthesisa,b
Solvent
t1/2 (min)c
Ethyl
t1/2 (min)c
Ethyl
benzoate
20 ꢀC
t1/2 (min)c
Isopropyl
benzoate
t1/2 (min)c
Isopropyl
benzoate
20 ꢀC
benzoate
0
ꢀC
0
ꢀC
THF
CHCl3
CHCl3(25 mL)
HMPA
DMF
DMF (2 equiv
NaOCOPh)
DMSO
<0.1
<0.1
0.22
w0.2
w0.3
1.0
0.3
1.0
2.2
*
*
1.5
*
*
1.5
*
3.4
9.5
Table 1
MeCN
1.5
2.2
3.5
*
51
*
Some parameters of solvent polaritya (25 ꢀC)
MeCN (10 mL)
MeCN (2 equiv
acid)
MeCN (2 equiv
NaOCOPh)
i-PrOH
ET (kcal.molꢂ1
)
Z (kcal.molꢂ1
)
εR
7.58
*
Solvent
THF
37.4
39.1
40.9
43.8
45.1
45.6
48.4
51.9
58.8
63.2
62.8
68.4
71.1
71.3
76.3
79.6
2.2
CHCl3
HMPA
DMF
DMSO
MeCN
iPrOH
EtOH
4.81
29.60
36.71
46.45
35.94
19.92
24.55
52
11.2
EtOH
315
a
Standard reaction conditions: alcohol (0.34 mmol), benzoic acid (0.34 mmol),
TPP (0.50 mmol) and benzophenone (0.08 mmol) pre-mixed in solvent (5 mL) be-
fore final addition of DIAD (0.50 mmol) at 0 ꢀC.
b
* indicates the result of at least two separate determinations.
Error in half-life determination is approximately ꢃ 20%.
a
Data obtained from Ref. 51.
c
Please cite this article in press as: Camp, D.; et al., Tetrahedron (2015), http://dx.doi.org/10.1016/j.tet.2015.04.035

D. Camp et al. / Tetrahedron xxx (2015) 1e7
3
The definitive mechanistic work of Hughes and co-workers22
established that the rate of esterification is reduced in the pres-
ence of an excess of carboxylic acid. This effect was confirmed in the
esterification of benzoic acid with ethanol. Thus, in acetonitrile at
latter reactions, the SN2 reaction rate increases with increasing
polarity of the solvent.51
There is evidence that the mechanism of the Mitsunobu ester-
ification reaction may involve the formation of ion pair
aggregates,20,25,66e70 wherein a positive phosphorus ion in one ion
pair is in part electrically neutralized by the negative carboxylate
moiety of another ion pair (Fig. 2, shown as a linear array for il-
lustrative purposes). A duplex mechanism involving a 12-mem-
bered ring (Fig. 3) is also possible.25 Polar solvents and salts could
easily interfere with the formation of such ion pair aggregates (or
break the chain depicted in Fig. 2), thereby inhibiting the SN2
process.
0
ꢀC, the half-life for formation of ethyl benzoate increased from
1.5 min to 3.5 min with the use of two equivalents of acid. As
suggested by Hughes and co-workers,22 increasing the amount of
benzoic acid results in a decrease in the kinetic reactivity of the
benzoate anion thereby reducing the rate of the final SN2 dis-
placement reaction.
We also investigated the effect of added sodium benzoate on the
rate of esterification. The addition of sodium benzoate has pre-
viously been found by Walker et al.20 to dramatically accelerate the
synthesis of trifluoroacetate esters via the Mitsunobu reaction.
Under the present experimental conditions in DMF and acetonitrile,
the addition of 2 M equivalents of sodium benzoate resulted in
a significant decrease in the rate of ethyl benzoate formation. This
apparently anomalous result is readily explained as Walker and co-
workers20 employed a much stronger acid (trifluoroacetic acid) in
their Mitsunobu reaction. Benzoate, being a much stronger base
than trifluoroacetate, results in acceleration of the alcohol activa-
tion step (1/2, Scheme 1) rather than the SN2 step as discussed by
Hughes et al.22 The reduction in rate observed in the present study
is consistent with a kinetic salt effect where the addition of an
external salt affects the rate of the reaction in the same way as an
increase in the solvent polarity.51,61
It is clear from Table 1 that increasing the solvent polarity results
in a significant reduction in the rate of Mitsunobu esterification
with an approximately linear relationship between the logarithm
of the reaction half-life and the solvent polarity (Fig. 1). This de-
crease in rate with increasing polarity is readily explained by the
Hughes-Ingold rules62e64 if we assume that the rate-determining
step is the final SN2 reaction (5/6, Scheme 1). As summarized
by Reichardt,51 an increase in solvent polarity leads to a rate de-
crease for those reactions in which the activated complex has
a lower charge density than the reactant molecules. In the present
case, the charge in the transition state relative to the reactants (the
alkoxyphosphonium cation and benzoate anion) will be decreased.
For example, if the transition state resembles the products, by the
Hammond postulate65 the charge would be close to zero as the
ester and triphenylphosphine oxide are both neutral. If, on the
other hand, the transition state is early as suggested by the kinetic
results of Hughes et al.,22 and recent theoretical calculations,46 then
there will still be a reduction in charge in the transition state, but it
will be a more modest reduction. The decrease in rate of the Mit-
sunobu reaction with increasing polarity of the solvent contrasts
with analogous nucleophilic substitution reactions where the
leaving group is halide, mesylate or triflate for example. In these
2.1. Apparent first-order kinetics
In an attempt to slow the reaction (being too fast to measure
accurately in non-polar solvents), an excess of benzoic acid was
employed (Protocol B). DIAD (0.66 mmol) was added dropwise to
a stirred solution of TPP (0.66 mmol), ethanol (0.34 mmol) and
benzophenone (0.05 mmol) in solvent (4 mL) under a nitrogen
atmosphere at ꢂ15 ꢀC. An excess of benzoic acid (1.31 mmol, four-
fold relative to alcohol; two-fold relative to betaine) in solvent
(1 mL) was then added and the solution maintained at 0 ꢀC. Under
these conditions, all of the alcohol was consumed to form dia-
lkoxytriphenylphosphorane (3, R¼Et). Addition of benzoic acid led
to rapid formation of the corresponding alkoxyphosphonium
benzoate (2, R¼Et, R0¼Ph). Both of these steps were confirmed by
proton NMR experiments (described later). The excess acid solvates
the benzoate nucleophile slowing the reaction as described by
Hughes.22 The kinetics were roughly first-order (i.e., first-order in
alkoxyphosphonium carboxylate), consistent with the results ob-
tained by Hughes et al.22 for the esterification of a secondary al-
cohol with formic acid. It should be noted, however that although
the rate of formation of ester appeared to follow first-order kinetics,
at least in the early stages of the reaction, the half-life was not in-
dependent of the concentration of alkoxyphosphonium salt. For
example, in MeCN using protocol A, halving the concentration of
ethoxyphosphonium benzoate resulted in an increase in the half-
life from 1.5 min to 2.2 min (Table 1). Similarly, in DMF using
protocol B, halving the concentration of ethoxyphosphonium
benzoate resulted in an increase in the half-life from 6.6 min to
14 min. As the half-life of a first-order reaction is independent of the
initial concentration, then the reaction cannot be a true first-order
reaction. The data is more consistent with a second-order reaction
(as expected for an SN2 process), where the half-life is inversely
proportional to the initial concentration.
Hughes and co-workers attributed this unusual rate de-
pendency (SN1 kinetics, first-order in alkoxyphosphonium ion,
zero-order in carboxylate) to salt effects. However, apparent first-
order kinetics could also be a result of ion pair clustering. Thus, if
the alkoxyphosphonium carboxylate were to form ion pair aggre-
gates prior to (rate-determining) SN2 displacement as in Fig. 2 or
Fig. 3, each benzoate ion is associated with one alkox-
yphosphonium ion. External benzoate ion would have no effect on
the concentration of benzoate within the ion pair aggregates. The
rate of ester formation would be dependent on the concentration of
ion pair aggregates that, in turn, would depend on the concentra-
tion of alkoxyphosphonium benzoate. Conversely, in the presence
of a swamping electrolyte (n-Bu4NBF4 was used by Hughes et al.22),
Fig. 1. Correlation between ET values and log half-lives for the Mitsunobu esterification
of benzoic acid with isopropanol at 20 ꢀC.
Fig. 2. Possible mechanism of SN2 step via an ion-pair aggregate.
Please cite this article in press as: Camp, D.; et al., Tetrahedron (2015), http://dx.doi.org/10.1016/j.tet.2015.04.035

4
D. Camp et al. / Tetrahedron xxx (2015) 1e7
Fig. 3. Alternative mechanism of SN2 step via an ion-pair duplex.
the ion pair aggregates would be broken down and the rate of ester
formation would revert to a normal second-order reaction (first-
order in alkoxyphosphonium ion, first-order in carboxylate as ob-
served by Hughes et al.22). We believe that the ion pair clustering
concept provides a more mechanistically satisfying explanation for
the unusual rate dependency than attributing it to ‘salt effects’. It is
also interesting to note that the formation of ion pair aggregates
would be enhanced in non-polar solvents.71 Very polar or protic
solvents on the other hand would favour the formation of solvent-
separated ion pairs. Whether ion pair aggregate formation might
under some circumstances (high polarity solvent, low concentra-
tions) become rate-limiting is not clear.
Fig. 4. Correlation between ET values and log rate constants for the Mitsunobu es-
terification of benzoic acid with ethanol at 0 ꢀC. (Protocol B).
decreased in intensity with time as the ester methylene resonance
at
d 4.35 ppm grew. This quintet has been assigned to the methy-
lene protons of the ethoxytriphenylphosphonium salt intermediatþe
(2, R¼Et, R0¼Ph) by comparison with literature values (Ph3POEt
ClOꢂ4 ,
d
4.41, quintet, J¼7 Hz;72 Ph3POEtþ BFꢂ4 d 4.44, quintet,
J¼7Hz73ethe benzoate salt would be expected to be slightly upfield
of these salts due to equilibrium with the alkoxyacylox-
yphosphorane 4). The use of ten equivalents of acid slowed the
reaction rate further and proton spectra for this reaction are shown
in Fig. 5.
The influence on the Mitsunobu esterification by four different
solvents using Protocol B (2 equiv of benzoic acid relative to betaine
1) can be seen in Table 3.
Using Protocol B (slower reaction), it was possible to obtain
initial first-order reaction rate constants in each of the four sol-
vents. It is again evident from the data that the rate of reaction is
inversely proportional to the polarity of the solvent. Fig. 4 illus-
trates the linear relationship between the logarithm of the rate
constant and the solvent polarity, as defined by ET values.
The effect of using an excess of carboxylic acid was also studied
with Protocol B. Table 4 presents the results with four, six, and ten
equivalents of benzoic acid (relative to ethanol) in chloroform.
It is clear that the rate of esterification drops significantly as the
concentration of benzoic acid increases, consistent with the results
of Hughes et al.22
Unfortunately, no accurate integrations of the ester and salt
methylene resonances could be obtained since they were almost
coincident. Deuterated benzene was briefly examined as a solvent
but did not provide any more resolution of the two peaks.
In order to provide further evidence that the quintet close to the
ester methylene resonance was due to the ethoxyphosphonium
carboxylate species, ethanol was replaced by neopentyl alcohol.
DIAD was added to an equimolar mixture of TPP and neopentyl
alcohol in CDCl3 at 0 ꢀC. The methylene resonance of the alcohol at
d
3.20 ppm was replaced by a doublet at
d
2.10 ppm, JPOCH¼4.0 Hz,
indicative of the dialkoxyphosphorane (3, R¼neopentyl, lit.18
d
2.30 ppm, JPOCH¼4.3 Hz). Complete reaction required several
2.2. Ester formation by 1H NMR
hours, as shown in Fig. 6.
Upon addition of benzoic acid, the dialkoxyphosphorane was
observed to decompose immediately to form the alkox-
yphosphonium benzoate salt (2, R¼neopentyl, R0¼Ph) and one
equivalent of alcohol. The methylene resonance of the salt was
We also examined the use of 1H NMR to study the rate of ester
formation. Initially, DIAD (two equivalents) was added to a CDCl3
solution of TPP (two equivalents), ethanol and benzoic acid at 0 ꢀC.
Immediate formation of ester (i.e., within 40 s) was evidenced by
the loss of ethanol resonances and the appearance of a quartet at
visible as a doublet at
d
3.80 ppm, JPOCH¼4.2 Hz (lit.25,26
d 4.30 ppm,
JPOCH¼4.2 Hz; lit.67 for triflate salt at
d
3.95 ppm, JPOCH¼4.0 Hz; lit.69
d
4.35 ppm and a triplet at d1.36 ppm. These shifts were confirmed
for chloride salt at
d
4.17 ppm, JPOCH¼4.3 Hz) Fig. 7.
by comparison with an authentic sample of ethyl benzoate in CDCl3.
In order to slow the rate of the esterification, an excess of benzoic
acid was used. Four equivalents of acid again gave immediate ester
formation, but when six equivalents of acid were used, a new
At room temperature, this salt resonance was slowly replaced by
that of neopentyl benzoate at
d
3.94 ppm. Ester formation at 0 ꢀC
was significantly slower, with a reaction time of over 4 h necessary
before any ester was observed.
The mechanism of esterification using the sterically hindered
neopentyl alcohol is probably via slow SN2 decomposition of the
quintet at
d
4.25 ppm (J¼7 Hz) was observed. This quintet rapidly
Table 3
Influence of solvent on ester synthesis by protocol Ba,b
Table 4
Solvent
t1/2 (min)c
103 k (sꢂ1
)
Influence of excess acid on the rate of ester synthesis in CHCl3 by protocol Ba
THF
0.5
2.2
6.6
52.9
*
*
*
*
22.5ꢃ5
5.2ꢃ1.4
1.7ꢃ0.1
Acid equivalents
T1/2 (min)b
103 k (sꢂ1
)
CHCl3
DMF
MeCN
4.0
6.0
10.0
2.2
5.1
11.1
5.2
2.3
1.0
0.22ꢃ0.01
a
Standard reaction conditions: addition of DIAD (0.66 mmol) to ethanol
(0.34 mmol), TPP (0.66 mmol) and benzophenone (0.05 mmol) pre-mixed in solvent
(4 mL) at ꢂ15 ꢀC, followed by final addition of benzoic acid (1.32 mmol) in solvent
(1 mL). Reaction mixture then maintained at 0 ꢀC.
a
Standard reaction conditions: addition of DIAD (0.66 mmol) to ethanol
(0.34 mmol), TPP (0.66 mmol) and benzophenone (0.05 mmol) pre-mixed in solvent
(4 mL) at ꢂ15 ꢀC, followed by final addition of benzoic acid (as shown) in solvent
(1 mL). Reaction mixture then maintained at 0 ꢀC.
b
* indicates the result of at least two separate determinations.
Error in half-life determination is approximately ꢃ 20%.
c
b
Error in half-life determination is approximately ꢃ 20%.
Please cite this article in press as: Camp, D.; et al., Tetrahedron (2015), http://dx.doi.org/10.1016/j.tet.2015.04.035

D. Camp et al. / Tetrahedron xxx (2015) 1e7
5
Fig. 7. 1H NMR spectra for the reaction of dineopentyloxytriphenylphosphorane with
benzoic acid in CDCl3 at 24 ꢀC.
Fig. 5. 1H NMR spectra for the reaction of TPP (two equivalents), ethanol (one
equivalent) and benzoic acid (ten equivalents) with DIAD (two equivalents) in CDCl3 at
0
ꢀC.
formation was found to be too rapid and the intermediates in-
sufficiently resolved for accurate kinetic data for the esterification
reaction to be obtained by 1H NMR.
alkoxyphosphonium salt (2, R¼neopentyl, R0¼Ph) although, acyl-
ation could also occur via the small equilibrium concentration of
alkoxyacyloxyphosphorane species 4. By adopting a conformation
with an equatorial neopentyloxy group and an apical benzoate
group, the oxygen of the neopentyloxy group is in close proximity
to the acyloxy carbonyl carbon (Fig. 8), thus providing a pathway to
esterification with retention of configuration of the alcohol, as de-
scribed by Enders et al.46
3. Conclusions
The rate of the Mitsunobu esterification reaction was found to
be much faster in non-polar solvents. This provides an explanation
as to why the yield in the Mitsunobu reaction, particularly with
sterically hindered alcohols, is often higher in non-polar solvents:
side reactions, such as acylation of the hydrazine, are much faster in
polar solvents,33 and therefore slower and less likely to be com-
petitive in non-polar solvents.
The logarithm of the rate constant was inversely proportional to
the solvent polarity, as defined by ET values. The presence of excess
benzoic acid resulted in a rate decrease, as did the presence of
sodium benzoate.
The 1H NMR results obtained for neopentyl benzoate formation
parallel those obtained for ethyl benzoate formation. The methy-
lene signal of ethyl benzoate (
d 4.35 ppm) is approximately 0.1 ppm
downfield of the methylene signal of the ethoxyphosphonium salt
(
(
d
d
4.25 ppm); similarly, the methylene signal of neopentyl benzoate
3.94 ppm) is also approximately 0.1 ppm downfield of the
methylene signal of the neopentyloxyphosphonium salt
(d
Each of the main species involved in the Mitsunobu esterifica-
tion reaction, including the alcohol starting material, dialkox-
yphosphorane, alkoxyphosphonium salt and ester product, was
detected by proton NMR analysis. The NMR data support rapid
formation of the alkoxyphosphonium salt intermediate with simple
alcohols such as ethanol, followed by rate-determining SN2 attack
of benzoate on the alkoxyphosphonium ion even in the presence of
a large excess of benzoic acid. The unusual kinetic data (SN2 re-
action but apparent first-order kinetics) can be explained by the
formation of ion pair aggregates or clusters, prior to rate-
determining SN2 attack of benzoate on the alkoxyphosphonium
ion.
3.80 ppm). The relatively rapid build-up of the alkoxyphosphonium
salt signal, followed by its slow decline with concomitant increase
of ester signal, is consistent with rate-determining SN2 attack of
benzoate on the alkoxyphosphonium ion, even in the presence of
excess benzoic acid.
These results confirm that the alkoxyphosphonium benzoate is
an intermediate in ethyl benzoate formation and that it is possible
for this salt to be identified by proton NMR. In fact, each of the main
species involved in alkyl benzoate synthesis via the Mitsunobu
reaction can be detected by proton NMR analysis. These include the
alcohol
starting
material,
dialkoxyphosphorane,
alkox-
yphosphonium salt and ester product. However, ethyl benzoate
4. Experimental section
4.1. General
Analytical grade THF was refluxed over sodiumepotassium alloy
ꢀ
under nitrogen, fractionally distilled and stored over 4 A molecular
sieves and sodium wire. Chloroform was pre-dried over calcium
Fig. 6. 1H NMR spectra for the reaction of TPP/DIAD with neopentyl alcohol in CDCl3 at
0
Fig. 8. Esterification with retention of configuration via an alkoxyacyloxyphosphorane
species.
ꢀC.
Please cite this article in press as: Camp, D.; et al., Tetrahedron (2015), http://dx.doi.org/10.1016/j.tet.2015.04.035

6
D. Camp et al. / Tetrahedron xxx (2015) 1e7
ꢀ
chloride, refluxed over P2O5, distilled, and stored over 4 A molec-
0.66 mmol) was then added dropwise. The mixture was maintained
at ꢂ15 ꢀC until the yellow color characteristic of DIAD had dissi-
pated (approximately 1 min) after, which time benzoic acid
(160.0 mg, 1.31 mmol) in solvent (1 mL) was added. The reaction
flask was transferred to an ice bath and maintained at 0 ꢀC. Aliquots
were removed as described above. Integral ratios and calculated
yields of ester for each reaction are tabulated in the Supplementary
Data.
ular sieves. DMF was pre-dried over barium oxide for twelve hours,
distilled at reduced pressure and stored over 4 A molecular sieves.
DMSO was refluxed over calcium oxide for several hours, distilled
at reduced pressure and stored over 4 A molecular sieves. Aceto-
nitrile was refluxed over calcium hydride, distilled and stored over
4 A molecular sieves. Analytical grade isopropanol was pre-dried
over calcium oxide and stored over 4 A molecular sieves. Dry eth-
ꢀ
ꢀ
ꢀ
ꢀ
anol was prepared by distillation from magnesium ethoxide and
ꢀ
stored over 4 A molecular sieves. Deuterated chloroform was stored
4.3. NMR experiments
ꢀ
over 4 A sieves. HPLC grade acetonitrile and distilled water were
filtered through 0.22 mm Millipore filter paper. Molecular sieves
4.3.1. NMR protocol A: standard reaction conditions. To a 5 mm
NMR tube was added triphenylphosphine (43.2 mg, 0.16 mmol),
were activated in a furnace at approximately 300 ꢀC for 12 h.
Triphenylphosphine, DIAD, benzoic acid and sodium benzoate
were commercially available and were not further purified. Au-
thentic samples of ethyl benzoate and isopropyl benzoate were
prepared by the method of Vogel.74
ethanol (5
chloroform-d (1.25 mL) under a nitrogen atmosphere. The sample
was cooled to 0 ꢀC and DIAD (32
L, 0.16 mmol) then added. Proton
mL, 0.08 mmol), benzoic acid (10.0 mg, 0.08 mmol) and
m
NMR spectra were acquired as described above.
Analytical HPLC was performed using a Dynamax-60A 8
mm
4.6 mm ID X 200 mm reverse phase C18 analytical column fitted
with a Whatman C18 reverse phase guard column. An ETPKortec
K35M HPLC Pump provided a flow of 0.7 mL/min of the eluent (60%
acetonitrile, 40% water) and an ETPKortec K95 Variable Wavelength
HPLC UV Detector was set to a wavelength of 230 nm for detection.
Data was recorded with a Shimadzu C-R6A Chromatopac. A cali-
bration curve was constructed from mixtures of internal standard
and authentic product samples in known concentrations.
4.3.2. NMR protocol B: standard reaction conditions. To a 5 mm
NMR tube was added triphenylphosphine (43.2 mg, 0.16 mmol),
ethanol (5
m
L, 0.08 mmol) and chloroform-d (1.25 mL) under a ni-
trogen atmosphere. The sample was cooled to 0 ꢀC and DIAD (32
mL,
0.16 mmol) then added. The mixture was shaken briefly and ben-
zoic acid (40.0 mg, 0.33 mmol) was added. Proton NMR spectra
were acquired as described above.
All NMR spectra were acquired at Griffith University Magnetic
4.3.3. Use of neopentyl alcohol. To a 5 mm NMR tube was added
triphenylphosphine (43.2 mg, 0.16 mmol), neopentyl alcohol
(15 mg, 0.17 mmol) and chloroform-d (1.25 mL) under a nitrogen
Resonance Facility, on
a Varian Unity-400 spectrometer at
400.016 MHz using a 45ꢀ pulse width and 1.66 s acquisition time. 4
to 64 scans were acquired. 1H chemical shifts are reported in
atmosphere. The sample was cooled to 0 ꢀC and DIAD (32
mL,
d
(parts per million) relative to internal TMS.
0.16 mmol) then added. Proton NMR spectra were acquired as de-
scribed above. After several hours, benzoic acid (21.0 mg,
0.17 mmol) was added and proton NMR spectra were acquired as
described above at 24 ꢀC.
4.2. Rate experiments
4.2.1. Protocol A: standard reaction conditions. Triphenylphosphine
(132.0 mg, 0.50 mmol), benzoic acid (40.5 mg, 0.34 mmol), ethanol
Acknowledgements
(20
zophenone (15.0 mg, 0.08 mmol) were stirred in solvent (5 mL) at
ꢀC under a nitrogen atmosphere. DIAD (98
L, 0.50 mmol) was
then added dropwise. At recorded time intervals, 20 L aliquots
mL, 0.34 mmol), or isopropanol (26 mL, 0.34 mmol), and ben-
We thank Griffith University for financial support.
0
m
m
Supplementary data
were removed from the reaction mixture and added to an aceto-
nitrile/water solution (1.0 mL, 60:40). This also served to quench
the reaction. A minimum of two 5 mL injections were performed for
each of these 1 mL volumes. The integral ratio of the ethyl benzoate
peak (retention time: 13.0 min) versus the benzophenone peak
(retention time: 15.4 min) was compared to a calibration curve to
determine the percentage yield of product as a function of time.
Integral ratios and calculated yields of ester for each reaction are
tabulated in the Supplementary Data.
Supplementary data (Integral ratios and calculated yields of
ester as a function of time. Sample calculation of first-order rate
constant and half-life.) related to this article can be found at http://
dx.doi.org/10.1016/j.tet.2015.04.035.
References and notes
1. Mitsunobu, O.; Yamada, M. ,; Mukaiyama, T. Bull. Chem. Soc. Jpn. 1967, 40,
935e939.
4.2.2. Protocol A: reaction performed at 20 ꢀC. Reagents (as de-
2. Mitsunobu, O.; Yamada, M. Bull. Chem. Soc. Jpn. 1967, 40, 2380e2383.
3. Mitsunobu, O. Synthesis 1981, 1e28.
4. Castro, B. R. Org. React. 1983, 29, 1e162.
scribed above) were stirred in solvent (5 mL) at 12 ꢀC under a ni-
trogen atmosphere. DIAD (98 mL, 0.50 mmol) was added dropwise
5. Hughes, D. L. Org. React. 1992, 42, 335e656.
and the reaction mixture warmed to 20 ꢀC. Aliquots were removed
as described above.
6. Jenkins, I. D.; Mitsunobu, O. In Encyclopaedia of Reagents for Organic Synthesis;
Paquette, L. A., Ed.; Wiley: New York, NY, 1995; Vol. 8, pp 5379e5390.
7. Hughes, D. L. Org. Prep. Proced. Int. 1996, 28, 127e164.
8. Dandapani, S.; Curran, D. P. Chem.dEur. J. 2004, 10, 3130e3138.
9. Dembrinski, R. Eur. J. Org. Chem. 2004, 69, 2763e2772.
10. But, T. Y. S.; Toy, P. H. Chem.dAsian J. 2007, 2, 1340e1355.
11. Reynolds, A. J.; Kassiou, M. Curr. Org. Chem. 2009, 13, 1610e1632.
12. Swamy, K. C. K.; Kumar, N. N. B.; Balaraman, E.; Kumar, K. V. P. Chem. Rev. 2009,
109, 2551e2651.
13. Brunn, E.; Huisgen, R. Angew. Chem., Int. Ed. Engl. 1969, 8, 513e515.
14. Guthrie, R. D. G.; Jenkins, I. D. Aust. J. Chem. 1982, 35, 767e774.
15. Heesing, A.; Steinkamp, H. Chem. Ber. 1982, 115, 2854e2864.
16. Grochowski, E.; Hilton, B. D.; Kupper, R. J.; Michejda, C. J. J. Am. Chem. Soc. 1982,
104, 6876e6877.
4.2.3. Protocol A: addition of sodium benzoate. Reagents (as de-
scribed above) and sodium benzoate (95.0 mg, 0.66 mmol) were
stirred in solvent (5 mL) at 0 ꢀC under a nitrogen atmosphere. DIAD
(98 mL, 0.50 mmol) was added dropwise and aliquots were removed
as described as above.
4.2.4. Protocol B: standard reaction conditions. Triphenylphosphine
(174.0 mg, 0.66 mmol), ethanol (20
mL, 0.34 mmol) and benzo-
17. Von Itzstein, M.; Jenkins, I. D. Aust. J. Chem. 1983, 36, 557e563.
18. Adam, W.; Narita, N.; Nishizawa, Y. J. Am. Chem. Soc. 1984, 106, 1843e1845.
19. Von Itzstein, M.; Jenkins, I. D. J. Chem. Soc., Perkin Trans. 1 1986, 437e445.
phenone (10.0 mg, 0.05 mmol) were stirred in solvent (4 mL) under
a nitrogen atmosphere at ꢂ15 ꢀC in an ice-salt bath. DIAD (130
mL,
Please cite this article in press as: Camp, D.; et al., Tetrahedron (2015), http://dx.doi.org/10.1016/j.tet.2015.04.035

D. Camp et al. / Tetrahedron xxx (2015) 1e7
7
20. Varasi, M.; Walker, K. A. M.; Maddox, M. L. J. Org. Chem. 1987, 52, 4235e4238.
49. Loibner, H.; Zbiral, E. Helv. Chim. Acta 1977, 60, 417e425.
21. Von Itzstein, M.; Jenkins, I. D. J. Chem. Soc., Perkin Trans. 1 1987, 2057e2060.
22. Hughes, D. L.; Reamer, R. A.; Bergan, J. J.; Grabowski, E. J. J. J. Am. Chem. Soc.
1988, 110, 6487e6491.
50. Saiah, M.; Bessodes, M.; Antonakis, K. Tetrahedron Lett. 1992, 33, 4317e4320.
51. Reichardt, C. Solvents and Solvent Effects in Organic Chemistry, 2nd ed.; VSH:
Weinheim, Germany, 1988; chapter 2 and chapter 7.
23. Crich, D.; Dyker, H.; Harris, R. J. J. Org. Chem. 1989, 54, 257e259.
24. Pautard-Cooper, A.; Evans, S. A., Jr. J. Org. Chem. 1989, 54, 2485e2488.
25. Camp, D.; Jenkins, I. D. J. Org. Chem. 1989, 54, 3045e3049.
26. Camp, D.; Jenkins, I. D. J. Org. Chem. 1989, 54, 3049e3054.
27. Martin, S. F.; Dodge, J. A. Tetrahedron Lett. 1991, 32, 3017e3020.
28. Camp, D.; Jenkins, I. D. Aust. J. Chem. 1992, 45, 47e55.
29. Wilson, S. R.; Perez, J.; Pasternak, A. J. Am. Chem. Soc. 1993, 115, 1994e1997.
30. Macor, J. E.; Wehner, J. M. Heterocycles 1993, 35, 349e365.
31. Kodaka, M.; Tomohiro, T.; Okuno, H. J. Chem. Soc., Chem. Commun. 1993, 81e82.
32. Camp, D.; Hanson, G. R.; Jenkins, I. D. J. Org. Chem. 1995, 60, 2977e2980.
33. Hughes, D. L.; Reamer, R. A. J. Org. Chem. 1996, 61, 2967e2971.
34. Moravcova, J.; Rollin, P.; Lorin, C.; Gardon, V.; Capkova, J.; Mazac, J. J. Carbohydr.
Chem. 1997, 16, 113e127.
35. Harvey, P. J.; von Itzstein, M.; Jenkins, I. D. Tetrahedron 1997, 53, 3933e3942.
36. Dodge, J. A.; Jones, S. A. Recent Dev. Org. Chem. 1997, 1, 273e283.
37. Eberson, L.; Persson, O.; Svensson, J. O. Acta Chem. Scand. 1998, 52, 1293e1300.
38. Watanabe, T.; Gridnev, I. D.; Imamoto, T. Chirality 2000, 12, 346e351.
39. Kumar, N. S.; Kommana, P.; Vittal, J. J.; Swamy, K. C. K. J. Org. Chem. 2002, 67,
6653e6658.
52. Dimroth, K.; Reichardt, C.; Siepmann, T.; Bohlmann, F. Liebigs Ann. Chem. 1963,
661, 1e37.
53. Kosower, E. M. J. Am. Chem. Soc. 1958, 80, 3253e3260.
54. Mitsunobu, O.; Kimura, J.; Iiizumi, K.; Noboru, Y. Bull. Chem. Soc. Jpn. 1976, 49,
510e513.
55. Bottle, S.; Jenkins, I. D. J. Chem. Soc., Chem. Commun. 1984, 385.
56. Jenkins, I. D.; Thang, S. Aust. J. Chem. 1984, 37, 2447e2451.
57. Guthrie, R. D.; Jenkins, I. D.; Thang, S.; Yamasaki, R. Carbohydr. Res. 1988, 176,
306e308.
58. Jenkins, I. D.; Goren, M. B. Chem. Phys. Lipids 1986, 41, 225e235.
59. Pautard, A. M.; Evans, S. A., Jr. J. Org. Chem. 1988, 53, 2300e2303.
60. Reeve, W.; Erikson, C. M.; Aluotto, P. F. Can. J. Chem. 1979, 57, 2747e2754.
61. Smith, B. M.; March, J. March’s Advanced Organic Chemistry, 5th ed.; Wiley: New
York, NY, 2001; pp 450e454.
62. Hughes, E. D.; Ingold, C. K. J. Chem. Soc. 1935, 244e255.
63. Cooper, K. A.; Dhar, M. L.; Hughes, E. D.; Ingold, C. K.; MacNulty, B. J.; Woolf, L. I.
J. Chem. Soc. 1948, 2043e2049.
64. Ingold, C. K. Structure and Mechanism in Organic Chemistry, 2nd ed.; Cornell
Univ.: Ithaca, 1969; pp 457e680.
40. Ahn, C.; Correia, R.; DeShong, P. J. Org. Chem. 2002, 67, 1751e1753 (addendum).
Ahn, C.; Correia, R.; DeShong, P. J. Org. Chem. 2003, 68, 1176.
41. Elson, K. E.; Jenkins, I. D.; Loughlin, W. A. Org. Biomol. Chem. 2003, 1,
2958e2965.
42. McNulty, J.; Capretta, A.; Laritchev, V.; Dyck, J.; Robertson, A. J. J. Org. Chem.
2003, 68, 1597e1600.
65. Hammond, G. S. J. Am. Chem. Soc. 1955, 77, 334e338.
66. Ramos, S.; Rosen, W. Tetrahedron Lett. 1981, 22, 35e38.
67. Ramos, S.; Rosen, W. J. Org. Chem. 1981, 46, 3530e3533.
68. Weiss, R. G.; Snyder, E. I. J. Org. Chem. 1970, 35, 1627e1632.
69. Jones, L. A.; Sumner, C. E., Jr.; Franzus, B.; Huang, T.-S.; Snyder, E. I. J. Org. Chem.
1978, 43, 2821e2827.
43. McNulty, J.; Capretta, A.; Laritchev, V.; Dyck, J.; Robertson, A. J. Angew. Chem.,
70. Slagle, J. D.; Huang, T.-S.; Franzus, B. J. Org. Chem. 1981, 46, 3526e3530.
71. See for example: MacMillar, S.; Fang, Y.; Westaway, K. C.; Matsson, O.; Beronius,
P. J. Phys. Org. Chem. 2008, 21, 237e241.
72. Maeda, H.; Koide, T.; Maki, T.; Ohmori, H. Chem. Pharm. Bull. 1995, 43,
1076e1080.
Int. Ed. 2003, 42, 4051e4054.
44. Shi, J. Y.; Hughes, D. L.; McNamara, J. M. Tetrahedron Lett. 2003, 44, 3609e3611.
45. Kumar, N. S.; Kumar, K. P.; Kumar, K. V. P.; Kommana, P.; Vittal, J. J.; Swamy, K. C.
K. J. Org. Chem. 2004, 69, 1880e1889.
46. Schenk, S.; Weston, J.; Anders, E. J. Am. Chem. Soc. 2005, 127, 12566e12576.
47. Swamy, K. C. K.; Kumar, K. P.; Kumar, N. N. B. J. Org. Chem. 2006, 71, 1002e1008.
48. Dodge, J. A.; Trujillo, J. I.; Presnell, M. J. Org. Chem. 1994, 59, 234e236.
73. Crich, D.; Dyker, H. Tetrahedron Lett. 1989, 30, 475e476.
74. Vogel, A. I. A Textbook of Practical Organic Chemistry, 3rd ed.; Longman: London,
UK, 1956; pp 780e790.
Please cite this article in press as: Camp, D.; et al., Tetrahedron (2015), http://dx.doi.org/10.1016/j.tet.2015.04.035