New insights in the mechanism of amine catalyzed epoxidation: Dual role of protonated ammonium salts as both phase transfer catalysts and activators of Oxone

Article abstract of DOI:10.1021/ja0289088

Amines have previously been reported to catalyze the epoxidation of alkenes using Oxone (2KHSO₅+KHSO₄+K2SO₄), and significant levels of asymmetric induction were observed. From screening a series of amines based on 2-subst

Full text of DOI:10.1021/ja0289088

Published on Web 05/28/2003
New Insights in the Mechanism of Amine Catalyzed
Epoxidation: Dual Role of Protonated Ammonium Salts as
Both Phase Transfer Catalysts and Activators of Oxone
Varinder K. Aggarwal,* Chrystel Lopin, and Franck Sandrinelli
Contribution from the School of Chemistry, UniVersity of Bristol, Cantock’s Close,
Bristol BS8 1TS, UK
Received January 11, 2002; E-mail: v.aggarwal@bristol.ac.uk
Abstract: Amines have previously been reported to catalyze the epoxidation of alkenes using Oxone
(2KHSO5+KHSO4+K2SO4), and significant levels of asymmetric induction were observed. From screening
a series of amines based on 2-substituted pyrrolidines, it has now been found that more consistent and
reproducible results are achieved with the HCl salt of the amine compared to the amine itself. Up to 66%
ee was achieved in epoxidation of 1-phenylcyclohexene. The chiral amine could be reisolated in >90%
yield when reactions were conducted at -10 °C, indicating that the integrity of the amine was maintained
during the oxidation process. At -10 °C, (S)-2-(diphenylmethyl)pyrrolidine 1 reacted with Oxone to give a
mixture of ammonium salts containing the peroxymonosulfate salt 6b. The enantioselectivity obtained with
this salt was compared to the amine‚HCl salt catalyzed process and identical results were observed,
indicating that the true oxidant was the peroxymonosulfate salt 6b. The relative rates of oxidation of cis-
and trans-â-methylstyrenes together with the F value of a series of 1-arylcyclohexenes were determined.
These studies indicated that the amine catalyzed process involved electrophilic oxidation. On the basis of
these findings, a new mechanism is advanced in which the protonated amine not only acts as a PTC but
also activates Oxone, through hydrogen bonding, toward electrophilic attack.
Scheme 1. Epoxidation of Alkenes by Oxone/NaHCO3 Catalyzed
by Amines
Introduction
We recently reported a novel process for epoxidation of
alkenes, using Oxone (2KHSO5+KHSO4+K2SO4) as oxidant,
which is catalyzed by simple amines.1 This process is attractively
simple and requires readily available cheap reagents in envi-
ronmentally friendly solvents. This novel method, which does
not require the use of transition metal catalyst,2 utilizes Oxone
buffered with NaHCO3/pyridine in MeCN:H2O and amines as
catalyst (Scheme 1).
recently discovered that these competition experiments are not
reproducible, they make the radical cation mechanism both less
credible and less compelling.4 A further problem we encountered
was the variable enantiomeric excess observed (32-38% ee5)
in different runs of our standard test reaction (1-phenylcyclo-
hexene was reacted with Oxone in the presence of 10% (S)-2-
(diphenyl-methyl)pyrrolidine 16). We therefore set out to
improve the reproducibility of the process, to improve the
enantioselectivity, and most importantly, to establish the mech-
anism of the amine catalyzed epoxidation reaction. In this
contribution we describe our success in achieving all of these
goals.
When a chiral amine was used significant asymmetric
induction was observed which could only result from intimate
association of the amine/amine derivative with either the alkene
or the oxidant. In attempting to determine the mechanism of
this novel process, we carried out competition experiments
between structurally similar alkenes and compared our selectivi-
ties with both electrophilic epoxidation, using either m-CPBA
or H2O2/MTO, and radical cation3 mediated reactions using
Ar3N+• SbCl6-. The very close similarity in selectivities of a
range of alkenes between the aminium salt (Ar3N+• SbCl6
)
-
Results and Discussion
catalyzed reaction of Bauld and Mirafzal3 and our own results
persuaded us that a similar radical cation mediated process was
occurring in our amine catalyzed process. However, as we have
Improvements in Reproducibility and Enantioselectivities.
In our amine catalyzed epoxidations, the difficulty in obtaining
(1) Adamo, M. F. A.; Aggarwal, V. K.; Sage, M. A. J. Am. Chem. Soc. 2000,
122, 8317-8318.
(4) The corrected data for the competition experiments have now been reported.
See: Adamo, M. F. A.; Aggarwal, V. K.; Sage, M. A. J. Am. Chem. Soc.
2002, 124, 11223-11223. We have also been unable to reproduce Bauld
and Mirafzal radical cation mediated epoxidations described in ref 3.
(5) In the original work this was erroneously reported as 57% ee.
(6) Bailey, D. J.; O’Hagan, D.; Tavasli, M. Tetrahedron: Asymmetry 1997, 8,
149-153.
(2) Jacobsen, E. N.; Wu, M. H. ComprehensiVe Asymmetric Catalytsis II;
Jacobsen, E. N., Pfaltz, A., Yamamoto, H., Eds.; Springer-Verlag: Berlin,
1999; pp 649-677.
(3) (a) Bauld, N. L. Tetrahedron 1989, 45, 5307-5363. (b) Bauld, N. L.;
Mirafzal, G. A. J. Am. Chem. Soc. 1991, 113, 3613-3614.
9
7596
J. AM. CHEM. SOC. 2003, 125, 7596-7601
10.1021/ja0289088 CCC: $25.00 © 2003 American Chemical Society

Mechanism of Amine Catalyzed Epoxidation
A R T I C L E S
Table 1. Asymmetric Epoxidation of 1-Phenylcyclohexene Using
Chiral Amines as Catalyst
the presence of triethylamine, which resulted in an acceleration
of the reduction.10
Scheme 2. Synthesis of (S)-2-(Di-R-naphthylmethyl)pyrrolidine 2
The epoxidation was equally effective at 5% loading of
1‚HCl (Table 1, entry 3) or using fewer equivalents of Oxone
(1 equiv)11 and NaHCO3 (5 equiv) (see Supporting Information).
Further reduction in the amounts of bicarbonate was less
effective.
Fate of Amine and Pyridine. At the end of the reaction (for
example, from Table 1, entry 2), neither amine 1 nor pyridine
were present, only their oxidation products (nitrone 5 and
pyridine-N-oxide) were observed (Scheme 3).
entry
catalysta
time
yield (%)b
ee (%)e,f
1
2
3
4
6
7
8
9
10% amine 1
10% 1‚HCl
5% 1‚HCl
30 min
20 min
40 min
2 h
60 min
2 h
87-99
93 (5c)
91 (8c)
68d
32-38
46
46
52
59
66
54
49
10% 1‚HCl (-10 °C)
10% 2‚HCl
89 (7c)
49d
10% 2‚HCl (-10 °C)
10% 3‚HCl
2 h
2 h
79 (11c)
7d
10% 3‚HCl (-10 °C)
Scheme 3. Fate of Amine and Pyridine after Reaction
a Standard conditions: 1-phenylcyclohexene (0.424 mmol), Oxone (0.848
mmol), NaHCO3 (4.24 mmol), pyridine (0.212 mmol), MeCN:H2O (95:5)
(0.5 mL), amine or amine‚HCl (0.042 or 0.021 mmol). b Yields determined
by 1H NMR relative to an internal standard (p-dimethoxybenzene). c Yield
of diol. d Remainder is alkene. e Enantioselectivities determined by chiral
GC using a R-CD column. f The epoxide has the (S,S) configuration. The
absolute configuration was determined by comparison of the optical rotation
with literature data (see Supporting Information).
As an excess of oxidant was used and the amines were
employed in substoichiometric amount, it is reasonable to expect
oxidation of the amine to occur. Indeed, Oxone in aqueous acetic
acid12 is known to oxidize amines but no reaction had been
conducted under our conditions of higher pH. We therefore
subjected amine 1 to our standard conditions for alkene oxidation
and obtained nitrone 5 in 32% yield. No starting material
remained (Scheme 4).
consistent and reproducible results was solved by employing
the hydrochloride salt of amine 1. Not only did this modification
give completely reproducible results, it also provided higher
enantiomeric excesses and shorter reaction times (Table 1, entry
2).
Even though a large excess of bicarbonate is employed, it is
almost entirely present in the solid phase with very little entering
the solution phase. Thus, under these conditions it seems
reasonable to assume that the protonated amine remains as the
HCl salt in the heterogeneous reaction mixture. Pyridine is
present to reduce the amount of epoxide hydrolysis, and it is
believed that it buffers the system by acting as a proton carrier
between the inorganic hydrogen sulfate and inorganic NaHCO3.
In the absence of pyridine, substantial amounts of diol were
formed.
The use of the protonated ammonium salt also resulted in
greater reproducibility with a broader range of amine catalysts.
This is exemplified in the case of the pyrrolidine derivatives 2
and 3,7 both of which gave higher enantioselectivity than 1
(Table 1). These amines were chosen as they provide substan-
tially greater steric hindrance and an additional polar group with
potential binding capabilities, respectively.
Pyrrolidine 2 was prepared in an analogous way to that for
16 (Scheme 2) although some modifications were required. The
reaction of the R-naphthyl Grignard with the ester derived from
proline was carried out using the cerium-complexed ester8 which
was superior to the procedure using the organocerium reagent
(7% yield).9 Hydrogenolysis of carbamate 4 was carried out in
Scheme 4. Oxidation of the Amine 1
If amine 1 and pyridine are both oxidized, what then are their
roles in the epoxidation process? We knew from earlier
experiments that in the absence of amine 1, no epoxidation
occurred and that none of the possible oxidation products
(hydroxylamine, nitrone, or N-hydroxylactam) were able to
promote epoxidation either. To answer this question, we repeated
the reaction shown in Scheme 3 but quenched the reaction after
5 min (Scheme 5).
In this case, even though substantial epoxidation had occurred,
the amine remained intact, and only a small amount of pyridine-
N-oxide was observed. Thus, it is clear that the alkene is
oxidized at a much faster rate than pyridine so that pyridine is
able to fulfill its role in acting as a proton shuttle during the
oxidation process, thus limiting epoxide hydrolysis. Amine 1
(10) (a) Effenberger, F.; Ja¨ger, J. Chem. Eur. J. 1997, 3, 1370-1374. (b)
Zymalkowski, F.; Schuster, T.; Scherer, H. Arch. Pharm. (Weinhiem, Ger.)
1969, 302, 272-284.
(7) (S)-diphenylprolinol 3 was purchased from Aldrich.
(8) (a) Dimitrov, V.; Kostova, K.; Genov, M. Tetrahedron Lett. 1996, 37,
6787-6790. (b) Dimitrov, V.; Bratovanov, S.; Simova, S.; Kostova, K.
Tetrahedron Lett. 1994, 35, 6713-6716. (c) Aggarwal, V. K.; Sandrinelli,
F.; Charmant, J. P. H. Tetrahedron: Asymmetry 2002, 13, 87-93.
(9) (a) Imamoto, T.; Takiyama, N.; Nakamura, K. Tetrahedron Lett. 1985, 26,
4763-4766. (b) Imamoto, T.; Sugiura, Y.; Takiyama, N. Tetrahedron Lett.
1984, 25, 4233-4236.
(11) One equivalent of Oxone provides two equivalents of KHSO5. An excess
of KHSO5 is needed as Oxone is known to decompose to produce 1O2.
The rate of decomposition is maximum at pH 9: Evans, D. F.; Upton, M.
W. J. Chem. Soc., Dalton Trans. 1985, 6, 1151-1153.
(12) For examples of oxidation of amines with Oxone see: Kennedy, R. J.;
Stock, A. M. J. Org. Chem. 1960, 25, 1901-1906.
9
J. AM. CHEM. SOC. VOL. 125, NO. 25, 2003 7597

A R T I C L E S
Aggarwal et al.
Scheme 5. Quench of the Epoxidation Reaction after 5 Min
Scheme 7. Reaction of Oxone with
(S)-2-(Diphenylmethyl)pyrrolidine 1 and Pyridine at -10 °C
process. We now believe that after alkene oxidation at room
temperature with amine 3, the amine is oxidized to the N-oxide
of the hydroxylamine and then fragments.15
is protected from oxidation by protonation. However, once all
of the alkene is consumed, the very small amount of free base
present in equilibrium with the protonated salt is slowly oxidized
by the excess oxidant to give the nitrone (Scheme 6). These
experiments implicate an ammonium salt as being the active
species responsible for promoting epoxidation.
Further evidence that the basic amine is protonated and
protected from oxidation came from reacting a mixture of amine
and pyridine with Oxone at -10 °C in the presence of sodium
bicarbonate (Scheme 7). The reaction resulted in complete
conversion of pyridine to the N-oxide and, after basic workup,
>90% reisolation of the amine. These results can be rationalized
by assuming that the less basic amine (pyridine) remains largely
unprotonated and so is oxidized,16 whereas the more basic amine
is protonated and effectively protected from oxidation.
Note that the low yield obtained with the hydrochloride salt
of amine 3 at -10 °C (Table 1, entry 9) is probably due to the
low solubility of the salt at this temperature.
Scheme 6
Isolation of Actual Oxidizing Species. In attempting to gain
further insights into the mechanism of the process, we discov-
ered that a complex, 6, was formed between Oxone and the
amine at -10 °C (Scheme 8). Through a combination of IR,
1H NMR, 13C NMR and elemental analysis of the complex and
comparison with an authentic sample of 1‚H2SO4, we concluded
that complex 6 was a mixture of the potassium sulfate 6a17 and
peroxymonosulfate anions 6b (see Supporting Information for
data).18 Attempts to grow suitable crystals of the complex for
X-ray analysis were unsuccessful. The complex titrated to a
value of 57-62 wt % of active oxidizing agent 6b19 and is
relatively stable: a sample left for 6 weeks at 0 °C lost
approximately 10% of activity.
Conducting Reaction at Low Temperature (Reisolation
of Amine). The epoxidation reaction could also be conducted
at -10 °C which resulted in an additional increase in enantio-
selectivity (Table 1, entries 4, 7, 9). Furthermore, at the lower
temperature, the amine could be reisolated in >90% yield,
indicating that amine oxidation was essentially inhibited at -10
°C. This was remarkable as we had shown that at room
temperature pyrrolidine was oxidized by Oxone.12,13 However,
this observation can also be rationalized. At -10 °C, the
concentration of the free base 1 is significantly lower than that
at room temperature so that the rate of amine oxidation is
reduced further. Interestingly, even â-hydroxyamine 3 could be
reisolated in >90% yield. Following alkene oxidation at room
temperature, this amine subsequently decomposed, and benzo-
phenone was generated, a process which either occurs from the
amine radical cation14 or from the N-oxide of the hydroxyl-
amine.15 Our early observations on the formation of benzo-
phenone provided additional circumstantial evidence for the
intermediacy of radical cations. However, the fact that we can
reisolate even amine 3 at low temperature shows that the
irreversible oxidation of 3 to the corresponding amine radical
cation cannot be an intermediate in the catalytic epoxidation
Scheme 8. Reaction of Oxone (2KHSO5, KHSO4, K2SO4) with
(S)-2-(Diphenylmethyl)pyrrolidine at -10 °C
This isolated complex 6 was tested in the epoxidation of four
different alkenes (Table 2, method A) and compared with the
catalytic process using 10% of the corresponding hydrochloride
salt (Table 2, method B). In every case identical enantioselec-
tivities were obtained (Table 2). As it is highly unlikely that
the ammonium salt 6b reacts with pyridine to return the neutral
(13) For examples of oxidation of amines with Oxone/acetone system see: (a)
Brik, M. E. Tetrahedron Lett. 1995, 36, 5519-5522. (b) Webb, K. S.;
Seneviratne, V. Tetrahedron Lett. 1995, 36, 2377-2378. (c) Crandall, J.
K.; Reix, T. J. Org. Chem. 1992, 57, 6759-6764.
(16) In the absence of amine 1, pyridine is not oxidised at -10 °C with Oxone
in MeCN/H2O.
(17) (a) pKa of pyrrolidine ) 11.31: Crampton, M. R.; Robotham, I. A. J. Chem.
Res. (S) 1997, 22-23. (b) pKa of KHSO4 ) 1.9: Zhu, W.; Ford, W. T. J.
Org. Chem. 1991, 56, 7022-7026.
(18) For spectroscopic analysis of Oxone, see: (a) Flanagan, J.; Griffith, W.
P.; Skapski, A. C. J. Chem. Soc., Chem. Commun. 1984, 1574-1575. (b)
Arnau, J. L.; Giguere, P. A. Can. J. Chem. 1970, 48, 3903-3910.
(19) We used the method employed by Trost for the determination of active
oxidant in tetrabutylammonium Oxone (TBO) to determine the amount of
active oxidant in complex 6. See: Trost, B. M.; Braslau, R. J. Org. Chem.
1988, 53, 532-537. Complex 6 is a mixture of ∼40% 6a and ∼60% 6b
by weight; 100 mg of complex 6 provides ∼0.17 mmol of oxidant (HSO5-).
(14) Su, Z.; Mariano, P. K.; Falvey, D. E.; Yoon, U. C.; Oh, S. W. J. Am.
Chem. Soc. 1998, 120, 10676-10686.
(15) Murahashi, S.-I.; Imada, Y.; Ohtake, H. J. Org. Chem. 1994, 59, 6170-
6172. In the case of 3, the mechanism below could account for the formation
of benzophenone and nitrone.
9
7598 J. AM. CHEM. SOC. VOL. 125, NO. 25, 2003

Mechanism of Amine Catalyzed Epoxidation
A R T I C L E S
Table 2. Epoxidation Reaction with Peroxymonosulfate
Ammonium 6b (Method A) or Using the Corresponding
Hydrochloride in Catalytic Amount (Method B) and Epoxidation
Reaction at -10 °C (Method C)
Scheme 9. Catalytic Cycle for Alkene Epoxidation
In competition experiments, the relative rates of oxidation
of cis- and trans-â-methyl styrenes was 56:44. A similar ratio
(58:42) was observed with mCPBA, while in the radical cation
mediated cyclopropanation of stilbene reported by Bauld, (E)-
stilbene is at least 100 times more reactive than (Z)-stilbene.20
Amine radical cation catalyzed cyclopropanations are compared
rather than epoxidations because, in both cases, the rate-
determining step involves reaction of the triarylaminium cation
with the alkene to give the alkene radical cation. Furthermore,
the cyclopropanation reactions are more reproducible than the
epoxidation reactions. The F value for oxidizing a series of
1-arylcyclohexenes was found to be -0.71. This is much closer
to the value obtained for electrophilic epoxidation of styrenes
with m-CPBA (F ) -0.93),21 whereas the F value for radical
cation mediated cyclopropanation is -5.17.20,22
a Method A (rt): 1-phenylcyclohexene (0.212 mmol), complex 6 (223
mg, 0.38 mmol oxidant), pyridine (0.212 mmol), MeCN:H2O (95:5) (1 mL).
Method B (rt): 1-phenylcyclohexene (0.424 mmol), Oxone (0.848 mmol),
NaHCO3 (4.24 mmol), pyridine (0.212 mmol), MeCN:H2O (95:5) (0.5 mL),
1‚HCl (0.042 mmol). Method C (-10 °C): 1-phenylcyclohexene (0.424
mmol), Oxone (1.272 mmol, added in three portions over 4 h 30 min),
NaHCO3 (2.12 mmol), pyridine (0.212 mmol), MeCN:H2O (95:5) (0.5 mL),
1‚HCl (0.042 mmol). b Yields determined by 1H NMR relative to an internal
standard (p-dimethoxybenzene). c Yield of diol. d Enantioselectivities de-
termined by chiral GC using a R-CD column. e The absolute configuration
was determined by comparison of the optical rotation with literature data.
f The absolute configuration was determined by chiral GC (Chiraldex γ-TA
column). g The absolute configuration was determined by chiral HPLC using
an OD column. See Supporting Information.
This evidence, together with the results from the isolated
complex, lead us to propose a new catalytic cycle for the
epoxidation process (Scheme 9). The amine is initially converted
into the ammonium salt 6a (X ) KSO4-), which upon exchange
with peroxymonosulfate gives the active oxidant 6b. After
oxidation, the hydrogen sulfate counterion of the ammonium
salt can either be exchanged for one of the other counterions
present including persulfate or react with NaHCO3 to give the
amine17 (which could then catalyze the epoxidation through an
alternative pathway), this salt is believed to be the active oxidant
in both methods A and B. Thus, whether the amine, or
hydrochloride salt, or complex 6 is employed, it is belieVed that
the key oxidant in all of these cases is the same and is the
ammonium salt 6b. The application of the optimized condition
at -10 °C (method C) with the hydrochloride salt of amine 1
on the series of alkenes is also given in Table 2. Again, slightly
higher enantioselectivities were obtained, but more importantly,
the amine could be reisolated in >90% yield in every case.
Note, the low yield obtained with trans-stilbene under
catalytic conditions (Table 2, entry 11) is due to the very low
solubility of the alkene in MeCN/H2O in the presence of the
salts (Oxone/bicarbonate). Our previous communication1 de-
scribed the broad scope of this epoxidation process: tetra-, tri-,
and certain disubstituted alkenes were good substrates, but
terminal alkenes (except for styrene) were not.
-
NaSO4 salt.
A series of ammonium salts with different levels of alkylation
and therefore different numbers of N-H protons were prepared
(7, 8) and tested as catalysts for epoxidation. These amines were
chosen with minimal differences to amine 1‚HCl in steric and
therefore physical properties (solubility) so that the effect of
degree of alkylation could be probed. It was found that tertiary
amine 7 was quite effective although both the enantioselectivity
and conversion were markedly reduced (Scheme 10). However,
the quaternary ammonium salt 8 was essentially ineffective,
giving similar amounts of epoxide to tetrabutylammonium
Oxone. In the absence of amine, the level of background
epoxidation is less than 5%. In fact, we originally considered
the possibility of a phase transfer catalyzed process in operation,
Discussion of Mechanism. There is now overwhelming
evidence that the epoxidation process is the result of the reaction
of a nucleophilic alkene with an electrophilic oxidant rather than
a one-electron transfer to form an intermediate amine radical
cation-alkene complex. The following experiments provide
additional support for this mechanism.
(20) Bauld, N. L.; Yueh, W. J. Am. Chem. Soc. 1994, 116, 8845-8846.
(21) Brook, M. A.; Smith, J. R. L.; Higgins, R.; Lester, D. J. Chem. Soc., Perkin
Trans. 2 1985, 1049-1055.
(22) Yueh, W.; Bauld, N. L. Res. Chem. Intermed. 1997, 23, 1-16.
9
J. AM. CHEM. SOC. VOL. 125, NO. 25, 2003 7599

A R T I C L E S
Aggarwal et al.
Scheme 10. Epoxidation of 1-Phenylcyclohexene Using the
Corresponding Secondary, Tertiary, and Quaternary Ammonium
Salts as Catalyst
Table 3. Variation of the Amount of Water
CH CN/H O ratioa
yield (%)b
ee (%)d,e
but discounted this mechanism on the basis of the lack of activity
of quaternary ammonium salts. These results show that the
protonated ammonium salts are uniquely effective in the
epoxidation process and that the more N-H’s there are available
the more effective is the catalyst in terms of both conversion
and enantioselectivity.
entry
3
2
1
2
3
4
5
6
95/5
93
90
92
60c
20c
8c
46
42
38
33
40
31
90/10
75/25
50/50
25/75
5/95
It now seems that the protonated ammonium salts serve a
dual role. Not only do they act as PTCs and help bring the
oxidant into solution, but we now believe that they also activate
the peroxymonosulfate, through hydrogen bonding, generating
a more electrophilic species.23 There are a number of possible
peroxymonosulfate-ammonium salt complexes, which can be
formed with amine 1 and they all activate the oxidant either
directly (II, III) or indirectly (I) (Scheme 11). The different
complexes clearly place the peroxy group in a different steric
environment, and this is likely to be a factor responsible for
the moderate enantioselectivity observed. The key to enhancing
the selectivity is to reduce the number of reactive complexes
or to engineer complexes so that one complex is considerably
more reactive than the others. Nevertheless, the high levels of
enantioselectivity observed are remarkable, considering the fact
that oxone is complexed to the chiral amine without covalent
bonds, especially considering that such levels of selectivity have
never been achieved in chiral peracid mediated epoxidations
(<15%)24 where the oxidant is coValently attached to the chiral
controller.
a Standard conditions: 1-phenylcyclohexene (0.424 mmol), Oxone (0.848
mmol), NaHCO3 (4.24 mmol), pyridine (0.212 mmol), MeCN:H2O (0.5
mL), 1‚HCl (0.042 mmol). b Yields determined by 1H NMR relative to an
internal standard (p-dimethoxybenzene). c Remainder is alkene. d Enantio-
selectivities determined by chiral GC using a R-CD column. e The epoxides
have the (S,S) configuration. The absolute configuration was determined
by comparison of the optical rotation with literature data.
the inorganic ions in preference to the lipophilic complex 6b
reducing its ability to break up this ion pair.
The difficulty in obtaining consistent and reproducible results
starting with the amine can now be understood if the rate of
protonation to form the salt 6a is slow and variable. In this
case amine degradation can occur through oxidation, giving rise
to different potential catalyst species, which will have different
activities. Rate of protonation versus oxidation could vary for
different reactions as it will depend on the amount of Oxone
and bicarbonate in solution which in turn will depend on the
particle size of Oxone and NaHCO3 and the rate of stirring (the
reaction mixture is heterogeneous). Starting with the protonated
amine salt, the amine is effectively protected from oxidation
and leads to completely reproducible results.
Scheme 11. Possible Forms of Peroxymonosulfate Ammonium 6b
Conclusions
In summary, the following evidence supports the intermediacy
of pyrrolidinium peroxymonosulfate 6b as the active oxidizing
species in the amine catalyzed epoxidation process: (i) a
complex was isolated containing the peroxymonosulfate salt 6b
which showed identical results when used in stoichiometric
amounts compared to catalytic amounts of the hydrochloride
salt of amine 1 and Oxone, (ii) the amine could be reisolated in
essentially quantitative yield, showing that it was protonated
(and therefore protected) throughout the oxidation process when
the reaction was conducted at -10 °C and not irreversibly
oxidized to the corresponding amine radical cation, (iii) the
relative rates of oxidation of structurally similar alkenes were
closer to those observed in m-CPBA oxidations than radical
cation mediated processes, (iv) the F value for oxidizing a series
of arylcyclohexenes was much closer (and smaller) to m-CPBA
oxidations than radical cation mediated processes, indicating
that the process involved an electrophilic oxidant.
If this proposal is correct, increased solvation of the am-
monium peroxymonosulfate 6b should not only result in a
reduction in rate of epoxidation but also in the enantioselectivity.
To test this, reactions were conducted in MeCN with increasing
concentrations of water (Table 3) (for the effect of other solvents
on epoxidation, see Supporting Information). Indeed, as the
water content was increased, both conversion and enantio-
selectivity generally decreased.25 However, the decrease was
not as precipitous as we had expected perhaps because of the
high salt content of the reaction. Water would no doubt solvate
(23) Protonated Oxone (i.e. Oxone at low pH) is a powerful electrophilic oxidant.
See: Zhu, W.; Ford, W. T. J. Org. Chem. 1991, 56, 7022-7026.
(24) (a) Ewins, R. C.; Henbest, H. B.; McKervey, M. A. Chem. Commun. 1967,
1085-1086. (b) Bowman, R. M.; Collins, J. F.; Grundon, M. F. Chem.
Commun. 1967, 1131-1132. (c) Montarani, F.; Moretti, I.; Torre, G. Chem.
Commun. 1969, 135-136. (d) Bowman, R. M.; Collins, J. F.; Grundon,
M. F. J. Chem. Soc., Perkin Trans. I 1972, 626-632. (e) Pirkle, W. H.;
Rinaldi, P. L. J. Org. Chem. 1977, 42, 2080-2082. (f) Rebek, J., Jr.;
McCready, R. J. Am. Chem. Soc. 1980, 102, 5602-5605.
(26) There are examples of protonated ammonium salts directing epoxidation
through hydrogen bonding with the oxidant (m-CPBA or dimethyloxirane)
whilst quaternary ammonium salts gave opposite selectivity. See: (a)
Asensio, G.; Gonza´lez-Nu´n˜ez, M. E.; Boix-Bernardini, C.; Mello, R.; Adam,
W. J. Am. Chem. Soc. 1993, 115, 7250-7253. (b) Asensio, G.; Boix-
Bernardini, C.; Andreu, C.; Gonza´lez-Nu´n˜ez, M. E.; Mello, R.; Edwards,
J. O.; Carpenter, G. B. J. Org. Chem. 1999, 64, 4705-4711. These reactions
(25) We cannot account for the small increase in ee in the 25/75 mixture of
MeCN/H2O (Table 3, entry 5).
9
7600 J. AM. CHEM. SOC. VOL. 125, NO. 25, 2003

Mechanism of Amine Catalyzed Epoxidation
A R T I C L E S
The remarkable finding that the secondary ammonium salt
is considerably more active than the tertiary ammonium salt
which in turn is more active than the quaternary ammonium
salt shows that the protonated amine is not just acting as a PTC
but is also activating Oxone, through hydrogen bonding, toward
electrophilic attack. Phase transfer catalyzed processes have a
long and well-established history, but as far as we are aware,
this is the first example in which a protonated ammonium salt
is far superior to a quaternary ammonium salt.26 As we now
have a much better understanding of the mechanism of this
unique amine catalyzed epoxidation process, we can begin to
improve the levels of asymmetric induction through variations
and derivatizations of the bounty of chiral amines bequeathed
by nature. Studies in this endeavor are ongoing.
Acknowledgment. We thank GSK (Andy Walker), Pfizer
(Clive Mason), Degussa (Ian Grayson), and EPSRC for support
of this work. We acknowledge helpful discussions with Dr. Bill
Sanderson.
Supporting Information Available: Synthesis and spectro-
scopic data of all amines/ammonium salts, synthesis and
analytical data for complex 6, titration of 6, typical procedures
(methods A, B, and C), competition experiments and determi-
nation of F (PDF). This material is available free of charge via
the Internet at http://pubs.acs.org or from the author.
are analogous to the seminal work of Henbest27 on hydroxy-directed
epoxidations but do not relate to phase transfer catalysed epoxidations of
the type described in this paper. Nevertheless, we appreciate one of the
reviewers directing us to this work.
(27) (a) Henbest, H. B.; Wilson, R. A. L. J. Chem. Soc. 1959, 1958-1965. (b)
Henbest, H. B. Proc. Chem. Soc. 1963, 159-165.
JA0289088
9
J. AM. CHEM. SOC. VOL. 125, NO. 25, 2003 7601