The Hendrickson reagent and the Mitsunobu reaction: A mechanistic study

Article abstract of DOI:10.1039/b305375j

The alkoxytriphenylphosphonium ion intermediate of the Mitsunobu reaction can be generated using the Hendrickson reagent, triphenylphosphonium anhydride trifluoromethanesulfonate, 1. Strangely, while the reagent 1 can be used in place of the Mitsunobu reagents (triphenylphosphine and a dialkylazodicarboxylate) for the esterification of primary alcohols, secondary alcohols such as menthol undergo elimination. Evidence is presented to show that this unexpected result is due to the presence of trialkylammonium triflate salts. Such salts lead to a dramatic decrease in the rate of esterification relative to competing elimination. The Mitsunobu esterification of menthol with p-nitrobenzoic acid was re-examined and the occurrence of elimination reported for the first time. The presence of traces of tetrabutylammonium triflate led to a dramatic reduction in the yield of inverted ester and a corresponding increase in the yield of anti elimination product 2-menthene. The mechanism of the Mitsunobu reaction is discussed in the light of the dramatic salt effect on both the rate and outcome of the reaction and the possible involvement of ion pair clustering. In contrast, use of the reagent 1 resulted in syn elimination to give a 1:2 mixture of 2- and 3-menthenes. Finally, 1 and sodium azide can be used to convert a primary alcohol into an azide in high yield. There was no reaction under Mitsunobu conditions.

Full text of DOI:10.1039/b305375j

View Article Online / Journal Homepage / Table of Contents for this issue
The Hendrickson reagent and the Mitsunobu reaction:
a mechanistic study†
Kathryn E. Elson,a Ian D. Jenkins*b and Wendy A. Loughlina
a School of Science, Griffith University, Nathan, Brisbane, QLD, 4111, Australia
b Natural Product Discovery, Griffith University, Nathan, QLD, 4111, Australia.
E-mail: I.Jenkins@griffith.edu.au
Received 14th May 2003, Accepted 25th June 2003
First published as an Advance Article on the web 15th July 2003
The alkoxytriphenylphosphonium ion intermediate of the Mitsunobu reaction can be generated using the
Hendrickson reagent, triphenylphosphonium anhydride trifluoromethanesulfonate, 1. Strangely, while the
reagent 1 can be used in place of the Mitsunobu reagents (triphenylphosphine and a dialkylazodicarboxylate)
for the esterification of primary alcohols, secondary alcohols such as menthol undergo elimination. Evidence is
presented to show that this unexpected result is due to the presence of trialkylammonium triflate salts. Such
salts lead to a dramatic decrease in the rate of esterification relative to competing elimination. The Mitsunobu
esterification of menthol with p-nitrobenzoic acid was re-examined and the occurrence of elimination reported for
the first time. The presence of traces of tetrabutylammonium triflate led to a dramatic reduction in the yield of
inverted ester and a corresponding increase in the yield of anti elimination product 2-menthene. The mechanism
of the Mitsunobu reaction is discussed in the light of the dramatic salt effect on both the rate and outcome of
the reaction and the possible involvement of ion pair clustering. In contrast, use of the reagent 1 resulted in syn
elimination to give a 1 : 2 mixture of 2- and 3-menthenes. Finally, 1 and sodium azide can be used to convert
a primary alcohol into an azide in high yield. There was no reaction under Mitsunobu conditions.
Introduction
The Mitsunobu reaction1–4 is one of the most useful and vers-
atile reactions in organic synthesis. The beauty of the reaction
lies in its generality and the mild reaction conditions employed.
It enables the synthetic organic chemist to convert a primary or
secondary alcohol into an excellent leaving group which can
subsequently be displaced by a wide range of nucleophiles,
either inter- or intra-molecularly. The one thousand or so refer-
ences to the Mitsunobu reaction testify to its generality and
usefulness.
In seeking an alternative protocol for the Mitsunobu
reaction that avoided the use of dialkyl azodicarboxylates, we
examined the Hendrickson “POP” reagent 1.5,6 This reagent,
triphenylphosphonium anhydride trifluoromethanesulfonate,
was discovered over 20 years ago, and appears to bring about
dehydrations (ester, amide formation etc.) in a manner analo-
gous to the Mitsunobu reaction. Indeed, the same key inter-
mediate (an alkoxyphosphonium salt) is almost certainly
involved in both reactions (Scheme 1).
Surprisingly, the similarity between the two reactions appears
to have gone largely unnoticed although it was mentioned in an
early paper by Ramos and Rosen7 and in more recent papers by
Hendrickson et al.5,8 In contrast to the Mitsunobu reaction that
is widely used, the Hendrickson reagent has been used only
occasionally and as far as we are aware, has not been investi-
gated as a means for inverting the configuration of an alcohol
(a common use for the Mitsunobu reaction).
Scheme 1
Some advantages of the Hendrickson reagent over the
Mitsunobu reagents are: (a) the recovered phosphine oxide can
be readily recycled by treatment with trifluoromethanesulfonic
anhydride (Tf2O); (b) azodicarboxylates are not required; (c)
competing Mitsunobu-type side reactions such as alkylation of
the hydrazinedicarboxylate are eliminated. One disadvantage
with using the reagent 1, is that two moles of triphenyl-
phosphine oxide are liberated per dehydration reaction [e.g.,
eqn. (1)] compared with only one mole in the Mitsunobu reac-
tion. This exacerbates the problem of separation of the product
from triphenylphosphine oxide by chromatography. However,
Hendrickson has reported the use of a POP reagent containing
a basic group to facilitate removal of the phosphine oxide.6a,b
A
similar strategy has been employed in the Mitsunobu reaction.9
The reagent 1 sometimes gives a reaction where the Mitsu-
nobu reaction fails. For example, alcohols can be reduced to
alkanes with NaBH4 (via alkoxyphosphonium salt formation
with 1) but not under Mitsunobu conditions.8 We also investi-
gated the latter reaction some years ago and observed only low
yields of toluene from benzyl alcohol after addition of tri-
phenylphosphine/diisopropyl azodicarboxylate (TPP/DIAD),
followed by Super Hydride-LiEt3AlH.10 This difference in
reactivity is possibly related to the presence of phosphoranes
that are formed under Mitsunobu conditions.11–13 In a more
recent example, Hendrickson has shown that heating epoxides
† Electronic supplementary information (ESI) available: (1) 1H and 13
C
NMR spectra of 4-chlorobenzyl thioacetate 3 and 4-chlorobenzyl
azide 4. (2) 1H NMR stack spectra of alkene region of menthenes
5a–5b formed by (a) thermolysis of menthol diphenylphosphate
ester;21 (b) Mitsunobu reagents (Organic Syntheses conditions);16 (c)
Hendrickson reagent (Table 1, entry 3) and (d) Hendrickson reagent
(Table 1, entry 8). (3) 31P spectra of Mitsunobu reactions performed in
the presence of (a) diisopropylethylammonium triflate; (b) diisopropyl-
ethylammonium triflate and diisopropylethylamine (2 eq.) and
1
1
(c) tetrabutylammonium triflate. (4) H and H{31P} NMR spectra of
(Ϫ)-menthoxytriphenylphosphonium triflate 11, H3. (5) Table 2 – raw
data for Fig. 3. See http://www.rsc.org/suppdata/ob/b3/b305375j/
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 9 5 8 – 2 9 6 5
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
2958

View Article Online
Table 1 Reaction of -(Ϫ)-menthol with the Hendrickson reagent and nucleophiles
Entrya, b
Solvent
Nucleophile
Products
Product ratiosc
1
2
3
4
5
6
7
8
9
DCM
THF
4-Nitrobenzoic acid
4-Nitrobenzoic acid
4-Nitrobenzoic acid
4-Nitrobenzoic acid
4-Nitrobenzoic acide
4-Nitrobenzoic acide
Thiol acetic acid
Thiol acetic acid
4-Methoxybenzoic acid
Sodium azide
5a ؉ 5b : menthol : 6
5a ؉ 5b : menthol : 6
5a ؉ 5b : menthol : 6
5a ؉ 5b : menthol : 6
5a ؉ 5b : menthol : 6f
5a ؉ 5b : menthol : 6
5a ؉ 5b : menthol : 7
5a ؉ 5b : menthol : 7
5a ؉ 5b : menthol : 8
5a ؉ 5b : menthol : 9
5a ؉ 5b : menthol : 9
97 : 1 : 2
84 : 12 : 4
68 : 10 : 22
52 : 48 : 0
27 : 0 : 72f
86 : 1 : 13
38 : 24 : 38
70 : 20 : 10
54 : 35 : 11
20 : 54 : 26
92 : 2 : 6
Toluene
Toluened
Toluene
DCM
Toluene
DCM
Toluene
DCM–DMF (2 : 1)
Toluene
10
11
Zn(N3)2ؒ2Py
a All reactions were carried out with 1 (1.0 eq.), l-(Ϫ)-menthol (1.0 eq.), nucleophile (1.0 eq.) and diisopropylethylamine (2.2 eq.) in 10 mL solvent.
b Reaction conditions were 40 ЊC for 24 h. c Product ratios determined by GC/MS analysis. d 100 mL solvent. e 4-Nitrobenzoic acid (4.5 eq.) and
diisopropylethylamine (5.5 eq.). f Ester displays retention and inversion of configuration in the ratio of 88 : 12 by 1H NMR.
(1)
with 1 results in the formation of dienes.14 There is no known
analogy for this using the Mitsunobu reagents. In this paper, we
compare and contrast the Hendrickson and Mitsunobu proto-
cols, particularly in relation to the esterification of menthol
with 4-nitrobenzoic acid.
Several other nucleophiles were also investigated briefly.
Thus, treatment of 1 with 4-chlorobenzyl alcohol, thiolacetic
acid, and diisopropylethylamine in DCM at room temperature
overnight gave a 97% yield of 4-chlorobenzyl thioacetate 3.
Similarly, treatment of 1 with 4-chlorobenzyl alcohol, sodium
azide, and diisopropylethylamine in 3 : 1 DCM–dimethyl-
formamide (DMF) gave a 93% yield of 4-chlorobenzyl azide
4. Interestingly, treatment of 4-chlorobenzyl alcohol under
Mitsunobu conditions with sodium azide in 3 : 1 DCM–DMF
at room temperature overnight gave no reaction, the alcohol
being recovered unchanged. It is clear from these examples and
the work of Hendrickson5 that the reagent 1 provides a mild,
useful procedure for the esterification of primary alcohols and
the synthesis of primary azides.
Esterification of secondary alcohols on the other hand was
surprisingly sluggish and resulted in very low yields. Menthol,
which has been used previously to investigate the stereo-
chemistry of esterification under Mitsunobu conditions,15,16
was chosen for this study. Treatment of 1 with the more
hindered secondary alcohol (Ϫ)-menthol, 4-nitrobenzoic acid
and diisopropylethylamine in DCM at room temperature over-
night gave no detectable ester formation. The reaction was
repeated at 40 ЊC for 24 h, then quenched and analysed by GC-
MS. It was found that the major products formed were the
products of elimination, 2- and 3-menthenes 5a and 5b (97%),
along with small amounts of menthol (1%), and the inverted
ester, neomenthyl 4-nitrobenzoate 6 (2%). The reaction was
examined under a range of conditions, and with different sol-
vents and different nucleophiles but in almost all cases, the
major reaction observed was elimination (Table 1). Comparing
entries 1, 2 and 3, toluene was found to give a higher relative
yield of inverted ester 6 (and less elimination) than tetrahydro-
furan (THF) and DCM, but even here the yield of inverted
ester 6 was only 22%. From Table 1, the ratio of elimination to
substitution is clearly both solvent dependent and concen-
tration dependent.
Results and discussion
Initially, we examined the use of 1 for the simple esterification
of a primary alcohol. The reagent 1 was generated by addition
of triflic anhydride to an ice-cold solution of triphenyl-
phosphine oxide in dry dichloromethane (DCM). 4-Nitro-
benzyl alcohol, 4-nitrobenzoic acid and diisopropylethylamine
were added and after stirring for 2 h at room temperature, work
up and chromatography, 4-nitrobenzyl 4-nitrobenzoate 2 was
obtained in high (95%) yield. A small excess of the triphenyl-
phosphine oxide was employed to ensure that all of the triflic
anhydride was consumed. Excess triflic anhydride could react
with the alcohol to form the triflate ester or the carboxylic acid
to form the corresponding anhydride. It was important to distil
the triflic anhydride from a small amount of P2O5 prior to use;
without distillation, conversion to ester 2 was 86%; with distil-
lation, conversion to ester 2 was 97.5% by 1H NMR. It was also
important to use only 1 equivalent of 1. Excess 1 led to reduced
yields of ester 2 (Fig. 1). The reason for this is not clear, however
it may be due to the excess reagent generating a more polar
reaction environment thereby slowing the reaction (cf. later).
Use of an excess of 4-nitrobenzoic acid when DCM was
employed (entry 6), did not improve the yield of inverted ester.
Use of 4-methoxybenzoic acid instead of 4-nitrobenzoic acid
led to a decrease in the yield of inverted ester 8 relative to
elimination (entry 9). The only reaction where esterification was
the major reaction was when a large excess of 4-nitrobenzoic
acid was employed (entry 5). Under these conditions however,
the ester formed, menthyl 4-nitrobenzoate,17 was mainly the
product of retention not inversion. Presumably with excess
acid, carboxyl group activation or anhydride formation become
kinetically competitive leading to menthol acylation with reten-
tion. Both the Hendrickson5 and Mitsunobu18 reagents are
Fig. 1 The effect of equivalents of Hendrickson reagent, 1, on yield
(by 1H NMR) of 4-nitrobenzyl 4-nitrobenzoate 2. Reagents and
conditions: 1 (0.8–2.0 eq.), 4-nitrobenzyl alcohol (1.0 eq., 3 mmol),
4-nitrobenzoic acid (1.0 eq., 3 mmol), diisopropylethylamine (5.3–
13.2 eq.) and DCM (10 mL); 2 h at room temperature.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 9 5 8 – 2 9 6 5
2959

View Article Online
known to convert carboxylic acids into their corresponding
anhydrides.
Use of the much more nucleophilic thiolacetic acid instead of
4-nitrobenzoic acid led to an increase in the ratio of SN2 substi-
tution product 7 to elimination (entries 7 and 8), but elimination
was still a significant pathway. Once again, the least polar solvent
toluene gave the highest proportion of SN2 substitution product 7
(38%). In contrast to the primary alcohol case, treatment of 1
with menthol, sodium azide, and diisopropylethylamine in 2 : 1
DCM–DMF at room temperature overnight gave only a low rel-
ative yield (26%) of the expected neomenthyl azide 9 (entry 10).
Use of toluene and the much more soluble zinc azide/
pyridine complex at 40 ЊC for 24 h gave a higher overall conver-
sion but a lower yield (6%) of the azide 9, the major pathway
(92%) being elimination (entry 11). Interestingly Maeda and
Ohmori,19 who generated analogous alkoxytriphenylphos-
phonium ions of secondary alcohols such as α-cholestanol by
an electrochemical procedure, found that these did not undergo
SN2 displacement by azide (Bu4NN3) in DCM but did undergo
SN2 displacement by bromide and chloride ions.
It is clear that with menthol, use of the reagent 1 leads
predominantly to elimination, irrespective of the nucleophilic
species present. Interestingly, the elimination mechanism
cannot be anti (E2), as both 2-menthene 5a and 3-menthene 5b
were formed (ratio ∼ 1 : 2, respectively). Hendrickson and
Schwartzman20 have previously reported that 1 reacts with
menthol in the presence of potassium carbonate to produce a
mixture of menthenes and suggested5 that cis elimination may
be favoured. An E1 mechanism would also be consistent with
the formation of a mixture of 2- and 3-menthenes 5a and 5b.
However, this seems less feasible since an intermediate carbo-
cation should be trapped by strong nucleophiles such as azide
and thioacetate to give a mixture of menthyl and neomenthyl
products, and this was not observed. It is also interesting to
note that a similar ratio (1 : 2) of 2- and 3-menthenes 5a–5b was
formed by thermolysis of menthyldiphenyl phosphate.21
A mechanism for the cis elimination is suggested in Scheme
2, by analogy with that for the thermolysis of menthyl-
diphenyl phosphate (10a, mechanism shown for formation
of 3-menthene 5b).
Scheme 2
benzoate at 0 ЊC is approximately 100 times slower in aceto-
nitrile (MeCN) than in THF]. We considered that the presence
of salts (diisopropylethylammonium triflate) in reactions
involving the Hendrickson reagent would result in a more polar
reaction environment, and that this effective increase in solvent
polarity might be tipping the balance in favour of elimination
over substitution. In order to test this hypothesis, we examined
the Mitsunobu esterification of menthol with 4-nitrobenzoic
acid in the presence of diisopropylethylammonium triflate by
31P NMR and GC/MS. Initially, several control esterification
reactions were carried out in the absence of added salt.
Reaction of menthol with 1 would be expected to result in the
initial formation of the mixed phosphorane/phosphonium salt
10. This could undergo an electrocyclic fragmentation to give
the two possible products of cis elimination (2- and 3-menthene
5a–5b), or it could eliminate triphenylphosphine oxide to give
the normal Mitsunobu intermediate 11, which could undergo
SN2 displacement by a nucleophile XϪ to give the neomenthyl
products 6–9. Presumably the latter process is kinetically slow
relative to elimination. We were unable to find any 31P NMR
evidence for the presence of 10. Upon mixing menthol with 1 at
room temperature, the oxyphosphonium salt 11 is formed
1
Organic syntheses conditions16
1
immediately as evidenced by both 31P (δ 59.5 ppm) and H
Under these conditions (menthol, 4-nitrobenzoic acid, TPP,
DIAD, THF, rt, 24 h) GC/MS showed the formation of neo-
menthyl 4-nitrobenzoate 6 (88%) and 2-menthene 5a (12%).
The ester 6 was isolated in 82% yield after chromatography
(lit.16 yield, 85%). It is interesting to note that although alkene
formation has previously been observed in the Mitsunobu reac-
tion during the total synthesis of doliculide22 and with sterically
hindered alcohols such as homopropargylic alcohol 13,23
cyanoalcohol 1424 and lactone 15,25 formation of menthene
from menthol under Mitsunobu conditions has not been
reported previously. Presumably being quite volatile, it is lost
during normal work-up and solvent evaporation.
NMR data (δ 4.25 ppm, dddd which collapsed to a ddd upon
phosphorus decoupling).
We also examined a less hindered secondary alcohol. Treat-
ment of cyclohexanol with 1, 4-nitrobenzoic acid and diiso-
propylethylamine in DCM at 40 ЊC for 24 h (same conditions as
entry 1, Table 1) gave cyclohexene (66%) as the major product,
cyclohexyl 4-nitrobenzoate 12 (28%) and cyclohexanol (6%).
It is clear from the contrasting results between primary and
secondary alcohols, that steric factors are very important
for Hendrickson reagent-mediated esterification and that the
greater the degree of steric crowding, the greater the amount of
competing elimination. The question is why hasn’t elimination
been observed previously with the Mitsunobu reaction of men-
thol and 4-nitrobenzoic acid, a well studied reaction15 described
in “Organic Syntheses”?16
2
Reaction in DCM (preferred solvent for reactions employing
the Hendrickson reagent5 and solvent used for Mitsunobu
reactions in the presence of salts)
Previous work from this laboratory10 had shown that the
rate of Mitsunobu esterification is markedly dependant on the
solvent polarity [for example, the rate of formation of ethyl
When the reaction was carried out in CD2Cl2 in an NMR tube,
the 31P NMR spectrum showed two main signals (ratio ∼ 9 : 1)
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 9 5 8 – 2 9 6 5
2960

View Article Online
corresponding26 to triphenylphosphine oxide (δ 29.1 ppm) and
the oxyphosphonium salt 18 (δ 59.5 ppm) (Scheme 4). Analysis
by GC/MS gave a yield of neomenthyl 4-nitrobenzoate 6 of
approximately 76% along with small amounts of menthol
(10%) and 2-menthene 5a (14%). The reaction was repeated but
in the presence of diisopropylethylammonium triflate. This
time, the 31P NMR spectrum showed slow formation of the
oxyphosphonium salt 18 (δ 59.5 ppm) from the protonated
betaine 17 (δ 52.6 ppm) (Scheme 3).26 After 24 h at room tem-
perature, the ratio of 17 : 18 was approximately 4 : 1. There was
very little increase (<5%) in the triphenylphosphine oxide peak
over this time indicating that the presence of diisopropylethyl-
ammonium triflate was effectively blocking the Mitsunobu
reaction. This was presumably due to protonation of the
betaine by the diisopropylethylammonium ion, leaving no base
to form the alkoxide anion, which, as noted by Hughes et al.,27
is required for oxyphosphonium salt formation. This was con-
firmed by carrying out the reaction in the presence of an extra
2 equivalents of diisopropylethylamine, which resulted in an
increase in the proportion of oxyphosphonium salt 18 (ratio
17 : 18 = 7 : 8). The presence of proton sources however, is
also known27 to slow the SN2 step of the Mitsunobu reaction
through hydrogen bonding with the carboxylate anion, hence
the lack of progression of the reaction beyond the oxy-
phosphonium intermediate stage. In order to avoid the com-
plications of protonation and hydrogen bonding effects, the
diisopropylethylammonium triflate was replaced by tetrabutyl-
ammonium triflate (n-Bu4NOTf ).
A solution of TPP, (Ϫ)-menthol and 4-nitrobenzoic acid was
prepared in dry DCM and n-Bu4NOTf in DCM was added
followed by the slow addition of DIAD to the reaction mixture
cooled in ice. After 24 h at room temperature, the reaction mix-
ture was analysed by GC/MS. The results, showing the effect
of the concentration of n-Bu4NOTf on the outcome of the
Mitsunobu reaction, are summarized in Fig. 2. It can be seen
that in the presence of one molar equivalent of n-Bu4NOTf
(relative to TPP/DIAD), the normal Mitsunobu esterification
reaction is slowed dramatically (4% yield of ester 6, 18% yield
of 2-menthene 5a, 78% recovered menthol) and elimination has
become the major reaction pathway (Fig. 2(a)). These results
are in sharp contrast to the results of Walker and coworkers28
who found that addition of sodium benzoate dramatically
accelerated the formation of trifluoroacetate esters. However, as
noted by Hughes,27 a catalytic amount of dissolved benzoate
can act as a base and increase the rate of the alcohol activation
step. This is not possible in the present case as n-Bu4NOTf is
not basic. Interestingly, the reaction is sensitive to even trace
amounts of n-Bu4NOTf salt (Fig. 2(b)). Thus, in going from
no salt to just 0.01 molar equivalents of salt (0.036 mmol salt,
3.6 mmol TPP/DIAD), the ratio of elimination to substitution
changed from 0.2 to 1.4. At higher proportions of salt, the ratio
of elimination to substitution increased steadily reached 10.4 at
2 molar equivalents of n-Bu4NOTf. At this level of salt, >80%
of the menthol was recovered unreacted after 24 h at room
temperature.
Scheme 3
Fig. 2 The effect of added tetrabutylammonium triflate on the
outcome of the Mitsunobu reaction between -(Ϫ)-menthol and 4-
nitrobenzoic acid. 2-Menthene 5a (ᮀ), -(–)-menthol (᭹), (1S,2S,5R)-5-
methyl-2-(1-methylethyl)cyclohexyl-4-nitrobenzoate 6 (᭺). Reagents
and conditions: TPP (1.0 eq.), DIAD (1.0 eq.), -(–)-menthol (0.85 eq.),
4-nitrobenzoic acid (0.85 eq.), n-Bu4NOTf (0.0–2.0 eq., Fig. 2(a);
0.0–0.2 eq., Fig. 2(b)) and DCM (8 mL); 24 h at room temperature.
It is clear from these results that salts have a dramatic effect
both on the rate of Mitsunobu esterification and on the out-
come. Just traces of salts lead to elimination becoming the
major reaction pathway rather than SN2 substitution. Previous
work by Hughes27 has shown that the SN2 substitution rate is
decreased 10-fold by addition of a large amount of a swamping
salt (0.5 M Bu4NBF4), while previous work from this labor-
atory10 has shown that the rate of Mitsunobu esterification is
very much slower in polar solvents. Both of these effects are
Scheme 4
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 9 5 8 – 2 9 6 5
2961

View Article Online
readily understood in terms of a late transition state (TS) for
the SN2 substitution reaction, a reaction in which the reactants
are oppositely charged and the products are neutral (Scheme 4).
The formation of such a TS would be favoured by a non-polar
reaction environment in accordance with the Hughes–Ingold
rules.29 By the Hammond Postulate30 the TS would resemble
the products.
It is interesting to note that increasing the concentration of
n-Bu4NOTf causes a decrease in both the rate of formation of
ester (SN2 product) and in the rate of formation of 2-menthene
5a (E2 product), but the effect on the SN2 process is greater than
on the E2 process. This suggests the TS for the E2 process 19 is
earlier (less product like) than for the corresponding SN2 pro-
cess 20. With a late TS, steric factors would be expected to be
more important, so it is perhaps not surprising that in the case
of menthol where the SN2 transition state 20 is quite crowded
that small changes in the reaction environment might tip the
balance in favour of the elimination pathway. While the trends
can be rationalised in terms of a later TS for the SN2 process
than the E2 process, it is still difficult to explain why traces of
n-Bu4NOTf cause such a dramatic turnaround in reaction
pathway from substitution to elimination.
One possible explanation is that the mechanism of the
Mistunobu esterification reaction involves the formation of ion
pair aggregates.7,28,31–33 In the normal SN2 process, the alkoxy-
phosphonium and carboxylate ions must be separated in order
to form the SN2 TS 20 required for backside attack. As noted by
Jones et al.,32b such charge separation can be avoided by ion pair
clustering wherein a positive phosphorus ion in one ion pair
is in part electrically neutralised by the negative carboxylate
moiety in another ion pair (as illustrated in Scheme 5). Traces
of salts (or polar solvents) could easily interfere with the form-
ation of such ion pair clusters, thereby inhibiting the SN2 pro-
cess. The corresponding E2 process (19) could be brought
about by any Lewis base present in the reaction medium, not
just the carboxylate ion shown. Thus, ion pair clustering is
not necessary for the E2 elimination process.
stable, undergoing slow elimination to give 2-menthene 5a,
and a minor amount of SN2 displacement to form neomenthyl
4-nitrobenzoate 6. When both the E2 and SN2 processes are
sufficiently slow, side reactions such as acylation of the hydra-
zine can become competitive,17,35 so that most of the menthol is
recovered at the end of the reaction (Fig. 2(a)).
One other intriguing difference between the Mitsunobu- and
Hendrickson-induced elimination reactions with menthol is
that the Mistunobu reaction resulted in stereospecific anti
elimination to give only 2-menthene (5a), whereas the
Hendrickson reagent resulted in the formation of 3-menthene
(5b) as the major product (ratio 3-menthene 5b : 2-menthene 5a
= 2 : 1). It is difficult to reconcile such differences if both the
Hendrickson and Mitsunobu reactions take place via a com-
mon intermediate (18). However, if the Hendrickson reagent
results in the initial formation of intermediate 10, then cis
elimination becomes feasible (Scheme 2). We thought that it
might be possible to convert the Mitsunobu intermediate 18
into an intermediate analogous to 10 by carrying out a
Mitsunobu reaction in the presence of a large excess of tri-
phenylphosphine oxide. However, even in the presence of a
5-fold excess of triphenylphosphine oxide (relative to TPP,
DIAD and menthol – no 4-nitrobenzoic acid was added in this
experiment) only 2-menthene 5a was formed.
Conclusion
31P NMR evidence suggests that both the Hendrickson reagent
and Mitsunobu reagent generate the same menthyloxy-
phosphonium ion intermediate, but the reaction outcome is
different in each case. The major reaction pathway under
Mitsunobu conditions is inverted ester formation (86%) with
only a minor amount (14%) of competing elimination. Elimin-
ation is anti to give exclusively 2-menthene 5a. When the
Hendrickson reagent is used, however, the major reaction
pathway is elimination (97%), and elimination appears to be
syn, to give a 2 : 1 mixture of 3- and 2-menthenes 5a : 5b,
respectively. The ratio of esterification to elimination can be
increased significantly by employing a non-polar solvent (tolu-
ene instead of DCM).
The difference between the two reactions appears to be
related to the more ‘ionic’ conditions generated when the
Hendrickson reagent is employed. When the Mitsunobu esteri-
fication reaction of menthol is carried out under analogous
‘ionic’ conditions (in the presence of salts), elimination also
becomes the dominant reaction pathway. With primary alco-
hols, substitution is the dominant pathway with both the
Hendrickson and Mitsunobu reagents. As a general rule, 1 can
be used as an alternative to the Mitsunobu reaction for
replacement of a primary hydroxyl by various nucleophiles.
With some nucleophiles (such as sodium azide) the more
‘ionic’ nature of 1 brings about reactions where the Mitsunobu
reaction fails.
Scheme 5
One other factor that affects the rate and overall outcome of
the Mitsunobu reaction is that whereas the oxyphosphonium
ion intermediate 11, is formed immediately upon mixing men-
thol and 1 at room temperature, reaction of menthol with the
Mitsunobu reagents to give the same oxyphosphonium ion 11
(or 18) occurs much more slowly. Thus, a mixture of menthol,
4-nitrobenzoic acid, and n-Bu4NOTf was treated with a slight
excess of DIAD and TPP in cold, dry CD2Cl2 under nitrogen.
The 31P NMR spectrum showed 3 major peaks, correspond-
ing to the oxyphosphonium salt 11 (δ 59.5 ppm, 22%), the
betaine 16 (δ 44.2 ppm,34 58%)) and triphenylphosphine oxide
(δ 29.1 ppm, 20%). After 24 h at room temperature, triphenyl-
phosphine oxide had become the major peak (71%) presumably
due to competing acylation of the hydrazine.17,35 The oxy-
phosphonium salt 11 was still present (16%) and there was a
small amount of protonated betaine 17 (13%). The much faster
rate of reaction of menthol with 1 is not surprising as the phos-
phonium center in 1 would be expected to be more electrophilic
than the phosphonium center in 16 or 17, and triphenyl-
phosphine oxide is a much better leaving group than the
hydrazide anion. Thus, the oxyphosphonium ion is being
generated, but in the presence of n-Bu4NOTf, it is relatively
Experimental
General methods
All reactions were carried out in flame-dried glassware under an
inert atmosphere. THF was freshly distilled from sodium benz-
ophenone ketal, toluene from sodium and DCM from calcium
hydride. Triflic anhydride was distilled from a small amount of
phosphorus pentoxide before use. Diisopropylethylamine was
distilled from potassium hydroxide and stored over molecular
sieves. All other reagents were purchased from commercial sup-
pliers and used without further purification. Flash chromato-
graphy was carried out using a Biotage system with pre-packed
silica columns. Melting points were determined by the capillary
method and are uncorrected. IR spectra were recorded on a
Thermo Nicolet-Nexus FTIR apparatus. Elemental analyses
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 9 5 8 – 2 9 6 5
2962

View Article Online
1
were performed at the University of Queensland. H NMR
washed with aqueous saturated NaHCO3 (2 × 20 mL) and
spectra were obtained at either 200 or 400 MHz and chemical
shifts are reported in parts per million, using the appropriate
signal for solvent protons as a reference. 13C NMR spectra were
recorded at 100 MHz and 31P NMR spectra at 162 MHz with
an internal reference of phosphoric acid (H3PO4). GC/MS
analyses were performed on a gas chromatograph (20 m db-5
ms capillary column) equipped with a mass-selective detector.
The identity of products was confirmed by H NMR, which
agreed with those reported in the literature, and by means of
co-injection of authentic samples.
water (2 × 20 mL), then dried (MgSO4). The combined organic
layers were concentrated in vacuo and the resulting residue
submitted to flash chromatography to remove the phosphine
oxide. Elution with DCM–hexane (1 : 1) afforded 4-chloro-
benzyl azide 4 as a colourless oil (0.31 g, 93%); nD 151.5556
(lit.,40 1.5600); νmax (Nujol)/cmϪ1 2100; δH (400 MHz, CDCl3)
4.33 (s, 2H), 7.26 (d, J = 8.5 Hz, 2H), 7.37 (d, J = 8.5 Hz, 2H);
δC (100 MHz, CDCl3) 54, 129, 129.5, 133.9, 134.2; HRMS
(Found: m/z 167.02495. C7H6ClN3 requires 167.02502).
1
The mixture of 2- and 3-menthenes 5a–b was prepared
by thermolysis of the diphenylphosphate ester of menthol.21
Authentic samples of (1S,2S,5R)-5-methyl-2-(1-methylethyl)-
cyclohexyl-4-nitrobenzoate 6,16 (ϩ)-neomenthyl thioacetate 7,36
(1S,2S,5R)-5-methyl-2-(1-methylethyl)cyclohexyl-4-methoxy-
benzoate 8,15 (1S,2S,5R)-5-methyl-2-(1-methylethyl)cyclohexyl-
azide 937 and cyclohexyl 4-nitrobenzoate 12 (mp 49–51 ЊC,
lit.,38 47–48 ЊC) were generated by standard Mitsunobu proto-
cols. The zinc azide/bis-pyridine complex was prepared from
zinc nitrate and sodium azide.37
Attempted synthesis of 4-chlorobenzyl azide using triphenyl-
phosphine, DIAD and NaN3
To a solution of TPP (1 g, 3.81 mmol) and 4-chlorobenzyl alco-
hol (0.46 g, 3.2 mmol) in DCM (10 mL) was added sodium
azide (0.27 g, 4 mmol) in DMF (4 mL). The heterogenous mix-
ture was cooled on ice and DIAD (0.75 mL, 3.81 mmol) added
slowly. After 30 min, the solution was left to warm to room
temperature overnight. The reaction mixture was washed with
aqueous saturated NaHCO3 (2 × 25 mL) and water (2 ×
25 mL), then dried (MgSO4). Concentration of the organic
layers in vacuo yielded a pale yellow residue that was shown to
contain unreacted 4-chlorobenzyl alcohol by 1H NMR analysis.
Typical example for the synthesis of 4-nitrobenzyl
4-nitrobenzoate 2 using 1.0 equivalent of Hendrickson reagent, 1
Representative example for the reaction of the Hendrickson
reagent with a secondary alcohol and nucleophile: attempted
synthesis of (1S,2S,5R)-5-methyl-2-(1-methylethyl)cyclohexyl-
4-nitrobenzoate 6
Addition of triflic anhydride (0.5 mL, 3 mmol) to an ice-cooled
solution of triphenylphosphine oxide (2 g, 7.2 mmol) in dry
DCM (10 mL) generated a white precipiate which was stirred
at 0 ЊC for 30 min. 4-Nitrobenzyl alcohol (0.46 g, 3 mmol),
4-nitrobenzoic acid (0.5 g, 3 mmol) and diisopropylethylamine
(1.15 mL, 6.6 mmol) were added in succession to form a yellow
solution. After stirring for 2 h at room temperature, the mixture
was washed with saturated aqueous NaHCO3 and water. The
organic layers were combined, dried (MgSO4) and concentrated
Triflic anhydride (0.5 mL, 3 mmol) was added slowly to an ice-
cooled solution of triphenylphosphine oxide (2 g, 7.2 mmol) in
dry DCM (10 mL). A white solid precipitated immediately and
the solution was stirred at 0 ЊC for 30 min. The DCM was
removed in vacuo and the remaining solid re-suspended in dry
toluene (10 mL). Addition of (1R,2S,5R)-(Ϫ)-menthol (0.46 g,
3 mmol) to the slurry generated a clear solution over 5 min, to
which the remaining 4-nitrobenzoic acid (0.5 g, 3 mmol)
and diisopropylethylamine (1.15 mL, 6.6 mmol) were added.
The resulting yellow solution was warmed at 40 ЊC for 24 h.
The reaction mixture was then quenched using aqueous satur-
ated NaHCO3 and an aliquot of the organic phase analysed
directly by GC/MS. The identity of the products generated was
confirmed by 1H NMR.
1
in vacuo. Analysis by H NMR revealed 97.5% conversion of
the benzyl alcohol to 4-nitrobenzyl 4-nitrobenzoate 2. Purifi-
cation by flash chromatography (eluent DCM–hexane, 1 : 1)
afforded 4-nitrobenzyl 4-nitrobenzoate 2 as a yellow solid (0.86
g, 95%). Mp 165–167 ЊC (lit.,39 168 ЊC).
Synthesis of 4-chlorobenzyl thioacetate 3 using 1 and thiolacetic
acid
Triflic anhydride (0.34 mL, 2 mmol) was added to triphenyl-
phosphine oxide (1.34 g, 4.8 mmol) in dry DCM (10 mL) as
above. 4-Chlorobenzyl alcohol (0.29 g, 2 mmol), thiol acetic
acid (0.143 mL, 2 mmol) and diisopropylethylamine (0.77 mL,
4.4 mmol) were added to generate a homogeneous solution that
was allowed to warm to room temperature overnight. The
resulting mixture was washed with aqueous saturated NaHCO3
(2 × 20 mL) and water (2 × 20 mL) and dried (MgSO4). The
organic layers were combined and reduced in vacuo to afford a
yellow residue that was purified by flash chromatography, using
DCM–hexane (1 : 1) as eluent. 4-Chlorobenzyl thioacetate 3
was obtained as a yellow oil (0.39 g, 97%) (Found: C, 53.62; H,
4.53. Calc. for C9H9ClOS: C, 53.86; H, 4.52%); νmax (Nujol)/
cmϪ1 1703, 1133, 1094; δH (400 MHz, CDCl3) 2.36 (s, 3H), 4.08
(s, 2H), 7.17 (d, J = 8.7 Hz, 2H), 7.27 (d, J = 8.7 Hz, 2H); δC (100
MHz, CDCl3) 30.3, 32.7, 128.7, 130.1, 133.0, 136.2, 194.8;
(ESMS ϩ) m/z 199 (35Cl MHϩ, 10%), 201 (37Cl MHϩ, 3%).
NMR study of (؊)-menthoxytriphenylphosphonium triflate 11
formation
A solution of triphenylphosphine oxide (0.1 g, 0.36 mmol) was
dissolved in dry CD2Cl2 (0.75 mL) in a 5 mm NMR tube. The
tube was cooled on ice and triflic anhydride (29 µL, 0.17 mmol)
added. A white solid formed immediately and after 10 min,
(1R,2S,5R)-(Ϫ)-menthol (0.027 g, 0.17 mmol) was added.
Brief mixing (vortex) generated a clear solution which showed
(by 1H NMR) complete conversion of (Ϫ)-menthol to
(Ϫ)-menthoxytriphenylphosphonium triflate 11. The 31P NMR
spectrum revealed a 1 : 1 mixture of (Ϫ)-menthoxytriphenyl-
phosphonium triflate 11 and protonated triphenylphosphine
oxide. δP (162 MHz, CD2Cl2) 51.4 (br s); 59.5 (s). 1H NMR (400
MHz, CD2Cl2) δ 4.25 (dddd, J = 10.7, 10.7, 6.6, 4.6 Hz, 1H,
H3). The key 31P and 1H NMR peaks for (Ϫ)-menthoxy-
triphenylphosphonium triflate 11 were in agreement with the
literature.19,33
Synthesis of 4-chlorobenzyl azide 4 using 1 and NaN3
Treatment of an ice-cooled solution of triphenylphosphine
oxide (1.34 g, 4.8 mmol) in DCM (10 mL) with triflic anhydride
(0.34 mL, 2 mmol) generated a white precipitate which was
stirred at 0 ЊC for 30 min. 4-Chlorobenzyl alcohol (0.29 g,
2 mmol), a solution of sodium azide (0.2 g, 3 mmol) in DMF
(3 mL) and diisopropylethylamine (0.77 mL, 4.4 mmol)
were added consecutively and the solution left to warm to
room temperature overnight. The cloudy yellow mixture was
Formation of 2-menthene 5a and neomenthyl 4-nitrobenzoate 6
(organic syntheses Mitsunobu procedure)16
A solution of TPP (4.02 g, 15.32 mmol), (1R,2S,5R)-(Ϫ)-
menthol (0.6 g, 3.84 mmol) and 4-nitrobenzoic acid (2.58 g,
15.44 mmol) in dry THF (30 mL) was cooled on ice. DIAD
(3.03 mL, 15.4 mmol) was added slowly and the reaction left to
warm to room temperature overnight. A sample of the reaction
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 9 5 8 – 2 9 6 5
2963

View Article Online
mixture was removed for GC/MS analysis which showed men-
thene 5a:neomenthyl 4-nitrobenzoate 6 = 12 : 88. The remaining
solution was diluted with ether (30 mL) and washed with satur-
ated aqueous sodium bicarbonate (2 × 20 mL). The aqueous
layers were combined and extracted with ether (20 mL). The
combined etheral layers were dried (MgSO4) and concentrated
in vacuo. The resulting solid was suspended in ether (10 mL)
and allowed to stand overnight. Hexane (5 mL) was added
slowly to the stirring mixture and the precipitated white solid
removed by filtration. The solvent was removed from the filtrate
and the resulting residue submitted to column chromatography
(8% ether–hexanes as eluent) to yield (1S,2S,5R)-5-methyl-2-
(1-methylethyl)cyclohexyl-4-nitrobenzoate 6 as a pale yellow
crystalline solid (0.96 g, 82%). Mp 92–94 ЊC, (lit.,16 93–95 ЊC).
(0.75 mL, 3.8 mmol) to the ice-cooled solution generated a
yellow solution which was left to warm to room temperature
overnight. A sample was then removed for GC/MS analysis and
the remaining solution concentrated in vacuo. Analysis by
both GC/MS and 1H NMR confirmed complete conversion of
menthol to 2-menthene 5a.
Acknowledgements
We gratefully acknowledge financial support for this work from
the Australian Research Council and Griffith University and
the support of Natural Product Discovery, Griffith University,
formerly AZGU.
31P NMR study of oxyphosphonium salt formation in the
presence of Pr2iNH؉EtOTf؊
References
1 O. Mitsunobu, Synthesis, 1981, 1–28.
2 D. L. Hughes, Org. React., 1992, 42, 335–656.
3 I. D. Jenkins and O. Mitsunobu, in Encyclopaedia of Reagents for
Organic Synthesis, ed. L. A. Paquette, Wiley, 1995, vol. 8, pp. 5379–
5390.
4 D. L. Hughes, Org. Prep. Proced. Int., 1996, 28(2), 127–164.
5 J. B. Hendrickson, Encyclopaedia of Reagents for Organic Synthesis,
ed. L. A. Paquette, Wiley, 1995, vol. 8, pp. 5405–5407.
6 (a) J. B. Hendrickson and Md. S. Hussoin, J. Org. Chem., 1987, 52,
4137–4139; (b) J. B. Hendrickson and Md. S. Hussoin, J. Org.
Chem., 1989, 54, 1144–1149; (c) J. B. Hendrickson and Md. S.
Hussoin, Synlett, 1990, 423–424.
7 S. Ramos and W. Rosen, Tetrahedron Lett., 1981, 22, 35–38.
8 J. B. Hendrickson, M. Singer and Md. S. Hussoin, J. Org. Chem.,
1993, 58, 6913–6914.
9 (a) D. Camp and I. D. Jenkins, Aust. J. Chem., 1988, 41, 1835–1839;
(b) M. von Itzstein and M. Mocerino, Synth. Commun., 1990, 20,
2049–2057; (c) M. von Itzstein, M. J. Jenkins and M. Mocerino,
Carbohydr. Res., 1990, 208, 287–288.
10 P. Harvey, PhD thesis, Griffith University, 1996.
11 M. von Itzstein and I. D. Jenkins, Aust. J. Chem., 1983, 36, 557–563;
M. von Itzstein and I. D. Jenkins, Aust. J. Chem., 1983, 37, 2447–
2479.
12 D. Camp and I. D. Jenkins, Aust. J. Chem., 1992, 45, 47–55.
13 M. von Itzstein and I. D. Jenkins, J. Chem. Soc., Perkin Trans., 1,
1986, 437–445; M. von Itzstein and I. D. Jenkins, J. Chem. Soc.,
Perkin Trans., 1, 1987, 2057–2060.
A solution of triflic acid (28 µL, 0.32 mmol) and diisopropyl-
ethylamine (56 µL, 0.32 mmol) in CD2Cl2 (0.75 mL) was
prepared in a 5 mm NMR tube. TPP (100 mg, 0.38 mmol),
(1R,2S,5R)-(Ϫ)-menthol (50 mg, 0.32 mmol) and 4-nitro-
benzoic acid (55 mg, 0.32 mmol) were added. A chilled solution
of DIAD (75 µL, 0.38 mmol) was then added to the tube at
0 ЊC. The tube was stoppered and sealed with parafilm. After
brief mixing by vortex, a 31P NMR spectrum was obtained at
25 ЊC and again after standing at room temperature 24 h. The
above procedure was repeated using an excess of diisopropyl-
ethylamine (168 µL, 0.96 mmol) and in the absence of
i
Pr2 NHϩEtOTfϪ.
At 25 ЊC: δP (162 MHz, CD2Cl2) 29.1 (s), 52.6 (s), 2 : 3,
respectively; at 25 ЊC after 24 h: δP 29.1 (s), 52.6 (s), 59.5 (s),
47 : 42 : 11, respectively; with an excess of diisopropylethyl-
amine at 25 ЊC: δP 29.1 (s), 52.6 (s), 53 : 47, respectively; with an
excess of diisopropylethylamine at 25 ЊC after 24 h: δP 29.1 (s),
52.6 (s), 59.5 (s), 55 : 21 : 33, respectively; in the absence of
i
Pr2 NHϩEtOTfϪ at 25 ЊC after 24 h: δP 29.1 (s), 59.5 (s), 9 : 1,
respectively.
General procedure for the addition of tetrabutylammonium
triflate to the standard Mitsunobu reaction
14 J. B. Hendrickson, M. A. Walker, A. Varvak and Md. S. Hussoin,
Synlett, 1996, 661–662.
A solution of TPP (1 g, 3.8 mmol), (1R,2S,5R)-(Ϫ)-menthol
(0.5 g, 3.2 mmol) and 4-nitrobenzoic acid (0.55 g, 3.2 mmol)
was prepared in dry DCM (8 mL). Tetrabutylammonium tri-
flate (0.01–2 eq., 0.038–7.6 mmol) in DCM (2 mL) was added.
Slow addition of DIAD (0.75 mL, 3.8 mmol) to the ice-cooled
liquid generated a yellow solution that was left to warm to
room temperature over 24 h. An aliquot of the reaction mixture
was removed for GC/MS analysis.
15 J. A. Dodge, J. I. Trujillo and M. Presnell, J. Org. Chem., 1994, 59,
234–236.
16 J. A. Dodge, J. S. Nissen and M. Presnell, Org. Synth., 1996, 73, 110–
115.
17 J. A. Gómez-Vidal, M. T. Forrester and R. B. Silverman, Org. Lett.,
2001, 3, 2477–2479.
18 P. J. Harvey, M. von Itzstein and I. D. Jenkins, Tetrahedron, 1997, 53,
3933–3942.
19 H. Maeda and H. Ohmori, Acc. Chem. Res., 1999, 32, 72–80.
20 J. B. Hendrickson and S. M. Schwartzman, Tetrahedron Lett., 1975,
277–280.
21 H. Quast and T. Dietz, Synthesis, 1995, 1300–1304.
22 H. Ishiwata, H. Sone, H. Kigoshi and K. Yamada, Tetrahedron,
1994, 50, 12853–12882.
23 R. W. Friesen and S. Bissada, Tetrahedron Lett., 1994, 35, 5615–
5618.
24 P. Castejon, A. Moyano, M. A. Pericas and A. Riera, Tetrahedron,
1996, 52, 7063–7086.
25 H. Oikawa, T. Kagawa, T. Kobayashi and A. Ichihara, Tetrahedron
Lett., 1996, 37, 6169–6172.
31P NMR study of oxyphosphonium salt formation in the
presence of Bu4N؉OTf؊
A solution of TPP (0.1 g, 0.38 mmol) in CD2Cl2 (0.75 mL)
was prepared in a 5 mm NMR tube. Tetrabutylammonium
triflate (0.15 g, 0.38 mmol), (1R,2S,5R)-(Ϫ)-menthol (50 mg,
0.32 mmol) and 4-nitrobenzoic acid (55 mg, 0.32 mmol) were
added. A chilled solution of DIAD (75 µL, 0.38 mmol) was
then added to the tube at 0 ЊC. The tube was stoppered, sealed
with parafilm and briefly vortexed. A 31P NMR spectrum was
obtained at 25 ЊC and again after standing at room temperature
24 h.
26 D. Camp and I. D. Jenkins, J. Org. Chem., 1989, 54, 3045–3049.
27 D. L. Hughes, R. A. Reamer, J. J. Bergan and J. J. Grabowski, J. Am.
Chem. Soc., 1988, 110, 6487–6491.
28 M. Varasi, K. A. M. Walker and M. L. Maddox, J. Org. Chem.,
1987, 52, 4235–4238.
29 C. Reichardt, Solvent and Solvent Effects in Organic Chemistry,
2nd edn., 1988, vol. 5, pp. 137––147.
At 25 ЊC: δP (162 MHz, CD2Cl2) 29.1 (s), 44.2 (s), 59.5(s)
∼1:3 : 1, respectively; at 25 ЊC after 24 h: δP 29.1 (s), 52.6 (s), 59.5
(s), ∼5 : 1 : 1, respectively.
30 G. S. Hammond, J. Am. Chem. Soc., 1955, 77, 334–338.
31 D. Camp and I. D. Jenkins, J. Org. Chem., 1989, 54, 3049–3054.
32 (a) R. G. Weiss and E. I. Snyder, J. Org. Chem., 1970, 35, 1627–1632;
(b) L. A. Jones, C. E. Sumner, Jr., B. Franzus, T-S. Huang and
E. I. Snyder, J. Org. Chem., 1978, 43, 2821–2827; (c) J. D.
Slagle, T-S. Huang and B. Franzus, J. Org. Chem., 1981, 46, 3526–
3530.
Addition of excess triphenylphospine oxide to a standard
Mitsunobu reaction (in the absence of 4-nitrobenzoic acid)
A solution of TPP (1 g, 3.8 mmol), triphenylphosphine oxide
(4.45 g, 16 mmol) and (1R,2S,5R)-(Ϫ)-menthol (0.5 g, 3.2
mmol) was prepared in dry THF (25 mL). Addition of DIAD
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 9 5 8 – 2 9 6 5
2964

View Article Online
33 S. Ramos and W. Rosen, J. Org. Chem., 1981, 46, 3530–3533.
34 R. D. Guthrie and I. D. Jenkins, Aust. J. Chem., 1982, 35, 767–774.
35 D. L. Hughes and R. A. Reamer, J. Org. Chem., 1996, 61, 2967–
2971.
36 B. Strijtveen and R. M. Kellogg, J. Org. Chem., 1986, 51, 3664–3671.
37 M. C. Viaud and P. Rollin, Synthesis, 1990, 130–132.
38 C. J. O’Connor, A. S. H. Mitha and P. Walde, Aust. J. Chem., 1986,
39, 249–257.
39 E. Lyons and E. E. Reid, J. Am. Chem. Soc., 1917, 39, 1727–
1735.
40 H. Hayashi, A. Ohno and S. Oka, Bull. Inst. Chem. Res., Kyoto
Univ., 1975, 53, 489–493.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 9 5 8 – 2 9 6 5
2965