A 'waterproof' catalyst for the oxidation of secondary amines to nitrones with alkyl hydroperoxides

Article abstract of DOI:10.1016/S0040-4039(02)02490-5

Catalytic oxidation of secondary amines to nitrones using alkyl hydroperoxides as primary oxidant has been demonstrated for the first time. The titanium alkoxide catalyst is protected from co-product water by the combined use of a tightly binding trialkanolamine ligand and molecular sieves. Nitrones can be obtained in high yield (up to 98%) under homogeneous, anhydrous conditions and even in the absence of solvent. The reactions are fast (2-7 h) and good selectivity can be achieved with as little as 1% catalyst.

Full text of DOI:10.1016/S0040-4039(02)02490-5

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 49–52
A ‘waterproof’ catalyst for the oxidation of secondary amines to
nitrones with alkyl hydroperoxides
a
b
a,
Massimiliano Forcato, William A. Nugent and Giulia Licini *
a
Universit a` di Padova, Dipartimento di Chimica Organica, ITM del CNR, Sezione di Padova, via Marzolo 1, 35131 Padova,
Italy
b
Bristol-Myers Squibb Co., Process Research and Development Dept., PO Box 269, Deepwater, NJ 08023, USA
Accepted 6 November 2002
Abstract—Catalytic oxidation of secondary amines to nitrones using alkyl hydroperoxides as primary oxidant has been
demonstrated for the first time. The titanium alkoxide catalyst is protected from co-product water by the combined use of a tightly
binding trialkanolamine ligand and molecular sieves. Nitrones can be obtained in high yield (up to 98%) under homogeneous,
anhydrous conditions and even in the absence of solvent. The reactions are fast (2–7 h) and good selectivity can be achieved with
as little as 1% catalyst. © 2002 Elsevier Science Ltd. All rights reserved.
9
Nitrones are versatile and valuable synthetic intermedi-
oxide (2–7 equiv.) is used as primary oxidant in the
presence of a catalyst (1–8% mol) such as Na MoO or
1
ates for the synthesis of heterocycles and natural
2
4
2
5,6,10,11
12
10,13
products and they are also effective spin trap
Na WO4
MeReO3 or SeO2.
Excellent yields
2
3
reagents. Usually, they are synthesized either via con-
have been obtained for the catalytic oxidation of cer-
tain amines. However, other cases suffer from limited
chemoselectivity such that a significant amount of
hydroxylamine is recovered. With highly reactive
nitrones, over-oxidation and hydrolysis can be signifi-
densation of carbonyl compounds with N-monosubsti-
4
tuted hydroxylamines (Scheme 1, path a) or by
1
2
oxidation of secondary amines or the corresponding
5
hydroxylamines (Scheme 1, path b).
6,14
cant problems.
It occurred to us that these problems
In principle, the oxidative approach provides the most
might be addressed by developing a catalytic system
that uses an alkyl hydroperoxide rather than hydrogen
peroxide as primary oxidant.
6
direct and general method for preparing nitrones. Sto-
ichiometric oxidants used in these reactions include
7
8
oxaziridines or dioxirane. Alternatively hydrogen per-
A literature report indicated that reaction of dibutyl-
amine with tert-butyl hydroperoxide in the presence of
4
%
titanium(IV) isopropoxide produced ‘trace
15
amounts’ of the corresponding nitrone. One interpre-
tation of this result is that the active titanium catalyst
species is being destroyed by hydrolysis as the reaction
proceeds. (Note that a molar equivalent of water is
necessarily formed in the conversion of 2 to 3.) We
hoped to address this problem in two ways: (1) addition
of molecular sieves to remove water from the system as
it is generated, and (2) replacement of some of the
isopropoxide ligands on the titanium catalyst with a
chelating polydentate ligand. In particular, the C -sym-
3
Scheme 1. Synthesis of nitrones via (a) condensation of car-
bonyl compounds with N-monosubstituted hydroxylamine or
metric trialkanolamines 4a–c seemed attractive candi-
dates as ligands for this application. We have
previously utilized a catalyst generated from titanium-
(
b) the oxidation of secondary amines.
1
6
(
IV) isopropoxide and 4c in the enantioselective
*
Corresponding author. Tel.: +39 049 8275289; fax: +39 049 827
239; e-mail: giulia.licini@unipd.it
oxidation of alkyl aryl sulfides using cumyl hydroperox-
ide as primary oxidant. The catalyst proved highly
17
5
0
040-4039/03/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)02490-5

5
0
M. Forcato et al. / Tetrahedron Letters 44 (2003) 49–52
efficient, reaching enantiomeric excesses up to 84% and
up to 1000 turnovers. Significantly, these catalysts were
highly robust, maintaining their solution integrity and
catalytic performance even in the presence of a large
h in presence of molecular sieves. Thus increasing the
reaction temperature accelerates the reaction without a
loss of selectivity while molecular sieves apparently
extend catalyst life, presumably by removal of the water
20
1
8
produced during the reaction.
excess of alkyl hydroperoxide or alcohol.
The nature of the ligand and the hydroperoxide both
significantly impact the rate and selectivity of the reac-
tion as shown in Table 1. These reactions were run at
6
0°C in the presence of 10% catalyst and 5 A
,
molecular
sieves.
Interestingly, at this rather high catalyst loading and in
the presence of molecular sieves, complete conversion
of the amine could be achieved using inexpensive
In an initial study, dibenzylamine 1a was used as model
substrate and the optimized conditions from our sulfox-
idation studies were utilized [10% catalyst generated
i
Ti(O Pr)4 in combination with cumyl hydroperoxide.
i
17,18
from Ti(O Pr) and 4c].
An excess (4 equiv.) of
However, the chemoselectivity was somewhat low
(91%) in this case. Reactions using the achiral titana-
trane derived from 4a (commercially available as
4
cumyl hydroperoxide was used as oxidant and reactions
1
19
were run in CDCl and monitored by H NMR.
3
®
TYZOR TE) proceeded only slowly; even after 21 h,
21
conversions were <15%.
As shown in Figure 1, complete conversion of the
amine into the nitrone was observed in all cases and no
hydroxylamine 2a was observed. At 25°C (Fig. 1A) the
reaction was slow and required 143 h for completion;
,
however, an identical run in the presence of 5 A molec-
ular sieves was complete in 45 h. An identical pair of
runs at 60°C (Fig. 1B) showed a further reduction in
reaction time to 4 h in the absence of additive and 2.5
i
In contrast, chiral catalysts prepared from Ti(O Pr)
4
and (S,S,S)-4b or (R,R,R)-4c afforded nitrone 3a in
high yield, with complete conversion of the reagent and
excellent chemoselectivity. Reaction rates for ligand 4c
were consistently faster than those for 4b. For both
ligands, cumyl hydroperoxide was superior to t-butyl
hydroperoxide in terms of both rate and chemoselectiv-
ity. Under the best conditions (ligand 4c/cumyl
hydroperoxide) complete conversion of the amine to 3a
could be achieved in 2.5 h.
It was subsequently determined that reactions pro-
ceeded somewhat faster in the presence of 4 A
,
versus 5
A molecular sieves and consequently 4 A sieves were
,
,
used in the remainder of our studies.Table 2 shows the
effect of increasing the substrate to catalyst ratio on the
oxidation of dibenzylamine with cumyl hydroperoxide.
Using a catalyst prepared from Ti(O Pr) and 4c, com-
i
4
plete conversion of the dibenzylamine with 100% selec-
tivity for the nitrone was observed in 1 h, even at a
substrate to catalyst ration s/c=100. It is noteworthy
that, in order to maintain the catalyst concentration at
0.01M, this reaction was performed neat. However,
further increasing the ratio to s/c=1000 (which neces-
sarily dropped the catalyst concentration to 0.001 M)
resulted in a severe deterioration of chemoselectivity.
Figure 1. Dependence of the yields of nitrone 3a on the time
(
h) for the oxidation of dibenzyl amine 1a [0.1 M] with CHP
0.4 M] catalyzed by Ti(IV)/(R,R,R)-4c [0.01 M] in CDCl3.
A) T=25°C (ꢀ, curve a) and with 5 A molecular sieves (250
mg/mmol) (ꢁ, curve b). (B): at 60°C (ꢂ, curve c) and with 5
molecular sieves (250 mg/mmol) (", curve d).
[
(
,
A
,
a
Table 1. Effect of the ligand and oxidant on the Ti-catalyzed oxidation of dibenzylamine
Conv. (%)a
2a (%)
a
3a (%)
a
PhCHO (%)b
Catalyst/oxidant
Time (h)
i
Ti( PrO) /TBHP
8.0
8.0
21.0
21.0
105.0
6.0
66
100
13
3
–
–
–
3
–
–
–
60
91
13
14
91
\99
98
\99
3
9
–
–
6
–
2
–
4
i
Ti( PrO) /CHP
4
Ti(IV)/4a/TBHP
Ti(IV)/4a/CHP
Ti(IV)/4b/TBHP
Ti(IV)/4b/CHP
Ti(IV)/4c/TBHP
Ti(IV)/4c/CHP
14
\99
\99
\99
\99
8.0
2.5
a
Conditions: hydroperoxide (0.4 M), Ti(IV) (0.01 M) and dibenzylamine (0.10 M) in CDCl3 at 60°C with 5 A molecular sieves (250 mg/mmol).
,
b
1
Yields of dibenzylhydroxylamine 2a, N-benzylidene benzylamine N-oxide 3a and benzaldehyde determined via H NMR using DCE as internal
standard.

M. Forcato et al. / Tetrahedron Letters 44 (2003) 49–52
51
Table 2. Effect of substrate/catalyst ratio on oxidation of
reports isolated yields for a series of pure nitrones
prepared under these conditions. The nitrones
were readily isolated via removal of the solvent
under vacuum followed by chromatography over sil-
a
dibenzylamine
b
[
Bn NH] , M
Substrate/
Time (h)
Yield (3a) (%)
2
0
catalyst
22
ica gel.
0
0
1
1
.1
.2
.0
.0
10:1
20:1
100:1
1000:1
3.0
1.5
1.0
100
100
100
Linear secondary amines (1a–d) and the cyclic N-ben-
zyl derivative 1e afforded the corresponding nitrones
in synthetically useful yields (70–98%) in 2–6 h. In
the case of 1,2,3,4-tetrahydroisoquinoline 1e, exclusive
formation of the regioisomer originating from ben-
zylic oxidation was observed.
c
100.0
68
a
i
Conditions: Ti(O Pr) =4c=0.01 M, dibenzylamine 1a, cumyl
4
hydroperoxide (4 equiv.) and 4 A
,
sieves in CDCl3 at 60°C.
b
c
Determined via 1H NMR using DCE as internal standard.
1
00% conversion; other products: N-benzylidene benzylamine
(10%), PhCHO (22%).
For the more reactive cyclic secondary amines 1f and
1
g, it proved advantageous to modify the reaction
conditions. When we attempted to oxidize cis-2,6-
dimethylpiperidine 1f, using 4 equiv. of hydroperox-
ide, the over-oxidation product 6-(hydroxyimino)-
heptan-2-one was found in the reaction mixture (25%,
The conditions for the highest catalyst loading (10%)
from Table 2 were adopted to explore the scope of
the reaction with other secondary amines. Table 3
1
determined via H NMR), together with higher
amounts of nitrone 3f (39%). When the amount of
Table 3. Catalytic oxidation of secondary amines by
hydroperoxide was decreased to 2 equiv., the yield of
a
CHP
3
f increased dramatically to 85%. It should be noted
that nitrone 3f is a chiral molecule, and in principle
could be formed enantioselectively using our enan-
1
tiopure catalyst. However, H NMR analysis in the
presence of a chiral solvating agent indicated that 3f
23
was generated in racemic form.
The use of homochiral trialkanolamines in a reaction
where enantioselectivity is not an issue deserves some
comment. Synthesis of the trialkanolamine by reac-
tion of ammonia with an enantiopure epoxide in the
ligand
synthesis
avoids
the
formation
of
diastereomeric (R,S,S) and (R,R,S) co-products. We
initially assumed that use of the pure homochiral lig-
and would be important to avoid destabilizing inter-
actions between the ligand arms in the case of the
undesired diastereomers. Interestingly, we have
recently tested this premise using the crude mixture of
trialkanolamines generated from racemic styrene oxide
and ammonia, which contains a mixture of both
diastereomers and regioisomers derived from attack of
ammonia at the benzylic position. We were somewhat
surprised to discover that this catalyst shows com-
parable activity to that prepared from fully purified
(
R,R,R)-4c. This observation will be followed up in
our future studies.
We feel that the prospects for further improvements
in this reaction are excellent and plan to pursue this
research. However, the current study has broader
implications beyond the issue of nitrone synthesis.
Hydrolytic instability is a general problem when using
early transition metal catalysts. Among the early tran-
sition metals, titanium is especially attractive because
of its low cost and non-toxic nature. In this paper we
have demonstrated that, with the proper choice of
ligands and reaction conditions, a catalytic quantity
of titanium can function in a reaction which produces
a stoichiometric amount of water.

5
2
M. Forcato et al. / Tetrahedron Letters 44 (2003) 49–52
Acknowledgements
16. Nugent, W. A.; Harlow, R. L. J. Am. Chem. Soc. 1994,
16, 6142.
1
This work was supported by the University of Padova,
Progetti di Ateneo 1999. It was carried out in the frame
of AC3S RTN Project (HPRN-CT-2001-00187) and
COST, Action D12 (D12/0018/98) and Action D24
17. (a) Di Furia, F.; Licini, G.; Modena, G.; Motterle, R.;
Nugent, W. A. J. Org. Chem. 1996, 61, 5175; (b) Licini,
G.; Bonchio, M.; Modena, G.; Nugent, W. A. Pure Appl.
Chem. 1999, 71, 463.
(
D24/0005/02).
18. Bonchio, M.; Licini, G.; Modena, G.; Bortolini, O.;
Moro, S.; Nugent, W. A. J. Am. Chem. Soc. 1999, 121,
6
258.
9. Typical oxidation procedure with Ti(IV) catalysis: in a
.0 ml volumetric flask Ti(IV)-based catalyst (0.01 mmol)
References
1
1
1
. Gothelf, K. V.; Jørgensen, K. A. Chem. Rev. 1998, 98,
was dissolved in CDCl . Activated molecular sieves (25
3
8
63.
mg), the hydroperoxide (0.4 mmol) and, after 30 min, the
substrate were added. The solution was transferred in a
screw cap NMR tube, warmed to the reaction tempera-
2
3
. Bloch, R. Chem. Rev. 1998, 98, 1407.
. Tanner, D. D. Advances in Free Radical Chemistry; JAI
Press: Greenwich, CT, 1990.
. Robi, J. A.; Hwu, J. R. J. Org. Chem. 1985, 50, 5913 and
references cited therein.
1
ture and the course of the reaction was monitored by H
4
5
6
NMR. For a detailed procedure for the synthesis of
Ti(IV)/4a–c complexes see Refs. 17 and 18.
0. A reaction containing molecular sieves but no titanium
catalyst resulted in no significant conversion of the amine
into products.
. Murahashi, S.-I. Angew. Chem., Int. Ed. Engl. 1995, 34,
2
2
443.
. Murahashi, S.-I.; Mitsui, H.; Shiota, T.; Tsuda, T.;
Watanabe, S. J. Org. Chem. 1990, 55, 1736 and references
cited therein.
21. The oxidation of diethylamine with the achiral dimer
2
7
. Cicchi, S.; Ponzuoli, P.; Goti, A.; Brandi, A. Tetrahedron
Lett. 2000, 41, 1583 and references cited therein.
. Murray, R. W.; Singh, M. J. Org. Chem. 1990, 55, 2954.
. Both H O in water or its adduct with urea (UHP) have
((h -t-butylperoxo)titanatrane) as stoichiometric oxidant
2
was reported by Boche et al. Et NOH was formed and
2
8
9
recovered as the insoluble hydroxyldiethylamine titana-
trane complex: Boche, G.; M o¨ bus, K.; Harms, K; Marsh,
A. J. Am. Chem. Soc. 1996, 118, 2770.
2
2
been used.
1
0. Marcantoni, E.; Petrini, M.; Polimanti, O. Tetrahedron
2
2. Typical oxidation procedure with Ti(IV)/4c catalysis: in a
5 ml screw-cap vial, under nitrogen, the preformed
catalyst (0.1 mmol) was dissolved in 10 ml of CHCl . 250
Lett. 1995, 36, 3561.
2
1
1. See also: Ballistreri, F. P.; Bianchini, R.; Pinzino, C.;
Tomaselli, G. A.; Toscano, R. M. J. Phys. Chem. A 2000,
04, 2710; Ballistreri, F. P.; Chiacchio, U.; Rescifina, A.;
Tomaselli, G. A.; Toscano, R. M. Tetrahedron 1992, 40,
677.
2. (a) Murray, R. W.; Iyanar, K. J. Org. Chem. 1996, 61,
3
mg of 4 A activated molecular sieves were added, fol-
,
1
lowed by CHP (4.0 mmol, 80% in cumene) and, after 30
min, by the substrate (1.0 mmol). The solution was
heated at 60°C and the reaction course has been moni-
tored via TLC. After the disappearance of the reagent,
the solvent was removed under vacuum and the reaction
mixture was purified directly via radial chromatography
over silica gel (gradient: ethyl ether, petroleum ether).
For a detailed procedure for the synthesis of Ti(IV)/4c
complex see Refs. 17 and 18.
8
1
8
3
8
099; (b) Goti, A.; Nannelli, L. Tetrahedron Lett. 1996,
7, 6025; (c) Yamazaki, S. Bull. Chem. Soc. Jpn. 1997, 70,
77.
1
1
1
3. Ballini, R.; Marcantoini, E.; Petrini, M. J. Org. Chem.
992, 57, 1316.
4. Zauche, T. H.; Espenson, J. H. Inorg. Chem. 1997, 36,
257.
5. In contrast, MoO (acac) , Mo(CO) and VO(acac)2 do
1
2
3. The enantiomeric excess of the product was determined
5
1
via H NMR in CDCl in the presence of (R)-(−)-1-(9-
3
2
2
6
not afford any product (see Ref. 6).
anthryl)-2,2,2-trifluoroethanol.