Effect of additives on chemoselectivity and diastereoselectivity in the catalytic epoxidation of chiral allylic alcohols with hydrogen peroxide and binuclear manganese complexes

Article abstract of DOI:10.1021/jo801974e

The catalytic oxidations of chiral allylic alcohols 2 by manganese complexes of the cyclic triamine 1,4,7-trimethyl-1,4,7-triazacyclononane (tmtacn) 1 and hydrogen peroxide as oxygen donor in the presence of co-catalyst are investigated to understand the

Full text of DOI:10.1021/jo801974e

Effect of Additives on Chemoselectivity and Diastereoselectivity in
the Catalytic Epoxidation of Chiral Allylic Alcohols with Hydrogen
Peroxide and Binuclear Manganese Complexes
Hamdullah Kilic,*,† Waldemar Adam,‡ and Paul L. Alsters§
Faculty of Science, Department of Chemistry, Ataturk UniVersity, 25240 Erzurum, Turkey, Institut fu¨r
Organische Chemie, UniVersita¨t Wu¨rzburg, Am Hubland, D-97074 Wu¨rzburg, Germany and Department of
Chemistry, FB-110, UniVersity of Puerto Rico, RIO PIEDRAS, Puerto Rico 00931, and DSM Pharma
Products, InnoVatiVe Synthesis and Catalysis, and DeVelopment, P.O. Box 18,
6160 MD Geleen, The Netherlands
hkilic@atauni.edu.tr
ReceiVed September 17, 2008
The catalytic oxidations of chiral allylic alcohols 2 by manganese complexes of the cyclic triamine 1,4,7-
trimethyl-1,4,7-triazacyclononane (tmtacn) 1 and hydrogen peroxide as oxygen donor in the presence of
co-catalyst are investigated to understand the factors that affect the catalyst selectivity. Chemoselectivity
and diastereoselectivity of catalyst 1 are significantly affected by the structure of the allylic alcohol and
the nature and amount of co-catalyst. More pronounced is the influence of the amount of added molar
equivalents of H2O2 (20-110 mol % with respect to the substrate). Our present results reflect the complex
redox chemistry of the Mn catalyst 1/H2O2/co-catalyst system in the early phase of the alkene oxidation.
Introduction
transition metal catalysts such as titanium, vanadium, rhenium,
and manganese.3 The epoxides, which are widely used in the
fine chemicals industry, may be readily prepared by the
stoichiometric oxidation of alkenes with peracids such as
peracetic and m-chlorobenzoic acids.4 The use of peracids is,
however, not an environmentally friendly method because an
equivalent amount of acid waste is produced. As an alternative
approach, much effort in this field has been directed toward
the efficient use of hydrogen peroxide as the oxygen source,
since it is relatively cheap, environmentally benign, and readily
available.5 The only byproduct is water, which makes this
reagent undoubtedly appealing for synthetic applications; un-
The oxidation of olefins to the corresponding epoxides is of
great synthetic value, from both an academic and an industrial
research point of view.1 Of special interest are catalytic
procedures, because they allow economic use of the employed
resources.2 Numerous catalytic systems are known to utilize
† Ataturk University.
‡ Universita¨t Wu¨rzburg and (present address) University of Puerto Rico.
§ DSM Pharma Products.
(1) (a) Sheldon, R. A.; Kochi, J. K. Metal Catalyzed Oxidations of Organic
Compounds; Academic Press: New York, 1981. (b) Lane, B. S.; Burgess, K
Chem ReV 2003, 103, 2457–2474.
(2) (a) Sheldon, R. A Top. Curr. Chem. 1993, 164, 23–43. (b) Sobkowiak,
A.; Tung, H.; Sawyer, D. T. Prog. Inorg. Chem. 1992, 40, 291–352.
(3) (a) Adam, W.; Corma, A.; Reddy, T. I.; Renz, M. J. Org. Chem. 1997,
62, 3631–3637. (b) Rossiter, B. E.; Verhoeven, T. R.; Sharpless, K. B.
Tetrahedron Lett. 1979, 19, 4733–4736. (c) Adam, W.; Mitchell, C. Angew.
Chem., Int. Ed. Engl. 1996, 35, 533–535. (d) Jacobsen, E. N. In Catalytic
Asymmetric Synthesis; Ojima, I., Ed.; VCH: New York, 1993; Chapter 4.2.
(4) Shapless, K. B.; Verhoeven, T. R. Aldrichimica Acta 1979, 12, 63–74.
(5) (a) Sato, K.; M.; Ogawa, M.; Hashimoto, T.; Noyori, R J. Org. Chem.
1996, 61, 8310–8311. (b) Hermann, W. A.; Fischer, R. W.; Marz, D. Angew.
Chem., Int. Ed. Engl. 1991, 30, 1683–1641. (c) Larrow, J. F.; Jacobsen, E. N.;
Gao, Y.; Hong, Y.; Nie, X.; Zepp, C. M. J. Org. Chem. 1994, 59, 1939–1942.
(d) Lane, B. S.; Burgess, K. J. Am. Chem. Soc. 2001, 123, 2933–2934.
10.1021/jo801974e CCC: $40.75
Published on Web 01/05/2009
2009 American Chemical Society
J. Org. Chem. 2009, 74, 1135–1140 1135

Kilic et al.
fortunately, hydrogen peroxide requires efficient activation for
oxygen transfer, advantageously by catalysis.
FIGURE 1. H2O2-activated bis(µ-carboxylato)-MnIII complex A.
2
Manganese complexes of cyclic amines such as 1,4,7-
trimethyl-1,4,7-triazacyclononane, abbreviated as Mn-tmtacn 1,
were reported to catalyze styrene epoxidation with hydrogen
peroxide in methanol-carbonate buffers.6 Because of the
extensive oxidant decomposition that such catalysts cause,
oxidant-to-substrate ratios of at least 100 need to be applied,
which cause highly dilute product mixtures. In the presence of
the alkene substrate, the balance between oxidant decomposition
and epoxidation may be shifted in favor of the latter, when
appropriate reaction conditions are chosen, e.g., acetone as
solvent. Although the peroxide decomposition is suppressed,
undesirable radical reactions take place.7
FIGURE 2. Proposed transition-state structure for the epoxidation of
Z-2c by catalyst 1.
constitutes a valuable mechanistic probe for the elucidation of
oxygen-transfer processes but also offers the opportunity to test
the chemoselectivity of the oxidation in terms of epoxide versus
enone formation, that is, the insertion of an oxygen atom into
the CC double bond or into the allylic CH bond.
We report herein the oxidation of allylic alcohols 2 with the
Mn-tmtacn catalyst 1 and hydrogen peroxide as the oxygen
source. The oxidative activity of catalyst 1, its chemoselectivity,
and stereoselectivity have been investigated in the presence of
various co-catalysts. The present results display that the
chemoselectivity and stereoselectivity of the oxidation of the
chiral allylic alcohols 2 depend not only on the nature and
amount of the carboxylic acid co-catalyst but also on the molar
equivalents of added H2O2 (20-110 mol % with respect to the
substrate).
Recent studies have shown that the catalytic activity of
manganese complexes 1 is strongly enhanced when carboxylic
acids such as oxalic acid, ascorbic acid, and their salts are
employed as co-ligands.8 Besides the readily oxidized phenols
and sulfides, even alkanes have been oxyfunctionalized with
the catalyst 1.9 A spectroscopic examination has shown that
complex 1 undergoes complex redox chemistry in contact with
H2O2.10,8d,e A very recent detailed study of the alkene oxidations
catalyzed by 1 plus carboxylic acid co-catalysts has revealed
that in the presence of H2O2, 1 is transformed into catalytically
active bis(µ-carboxylato)-MnIII complex A (Figure 1) during
2
Results
an initial lag period before the onset of alkene oxidation.8d
Manganese complexes Mn-tmtacn 1 have so far not been
employed for the epoxidation of chiral allylic alcohols 2. The
set of stereochemically labeled alcohols 2 in Table 2 not only
The Mn-tmtacn catalyst11 1 and the chiral allylic alcohols
2a-f were prepared according to literature procedures.12 The
oxidations were conducted with a catalytic amount of Mn-tmtacn
1 and 1.1 equiv of hydrogen peroxide in the presence of a co-
catalyst. A general procedure is given in the Experimental
Section.
(6) Hage, R.; Iburg, J. E.; Kerschner, J.; Koek, J. E.; Lempers, E. L. M.;
Martens, R. J.; Racheria, U. S.; Russell, S. W.; Swathoff, T.; van Vliet, M. R. P.;
Warnaar, J. B.; van der Wolf, L.; Krijnen, B. Nature 1994, 369, 637–639.
(7) (a) De Vos, D.; Bein, T. J. Chem. Soc., Chem. Commun. 1996, 91, 7–
918. (b) De Vos, D.; Bein, T. J. Organomet. Chem. 1996, 520, 195–200.
(8) (a) De Vos, D. E.; Sels, B. F.; Reynaers, M.; Subba Rao, Y. V.; Jacobs,
P. A. Tetrahedron Lett. 1998, 39, 3221–3224. (b) Berkessel, A.; Sklorz, C. A.
Tetrahedron Lett. 1999, 40, 7965–7968. (c) De Boer, J. W.; Brinksma, J.; Browne,
W. R.; Meetsma, A.; Alsters, P. L.; Hage, R.; Feringa, B. L. J. Am. Chem. Soc.
2005, 127, 7990–7991. (d) De Boer, J. W.; Browne, W. R.; Brinksma, J.; Alsters,
P. L.; Hage, R.; Feringa, B. L. Inorg. Chem. 2007, 46, 6353–6372. (e) De Boer,
J. W.; Alsters, P. L.; Meetsma, A.; Hage, R.; Browne, W. R.; Feringa, B. L.
Dalton Trans., accepted for publication.
(9) (a) Smith, J. R. L.; Shul’pin, G. B Tetrahedron Lett. 1998, 39, 4909–
4912. (b) Barton, D. H. R.; Choi, S.-Y.; Hu, B.; Smith, J. A. Tetrahedron 1998,
54, 3367–3378. (c) Barton, D. H. R.; Li, W.; Smith, J. A. Tetrahedron Lett.
1998, 39, 7055–7058. (d) Shul’pin, G. B.; Su¨ss-Fink, G.; Lindsay Smith, J. R.
Tetrahedron 1999, 55, 5345–5358. (e) Nizova, G. V.; Bolm, C.; Ceccarelli, S.;
Pavan, C.; Shul’pin, G. B. AdV. Synth. Catal. 2002, 344, 899–905. (f) Nizova,
G. V.; Shul’pin, G. B. Tetrahedron 2007, 63, 7997–8001. (g) dos Santos, V. A.;
Shul’pina, L. S.; Veghini, D.; Mandelli, D.; Shul’pin, G. B. React. Kinet. Catal.
Lett. 2006, 88, 339–348. (h) Shul’pin, G. B.; Matthes, M. G.; Romakh, V. B.;
Barbosa, M. I. F.; Aoyagi, J. L. T.; Mandelli, D. Tetrahedron 2008, 64, 2143–
2152. (i) Shul’pin, G. B.; Su¨ss-Fink, G.; Shul’pina, L. S. J. Mol. Catal., A: Chem.
2001, 170, 17–34. (k) Woitiski, C. B.; Kozlov, Y. N.; Mandelli, D.; Nizova,
G. V.; Schuchardt, U.; Shul’pin, G. B. J. Mol. Catal. A: Chem. 2004, 222, 103–
119. (l) Shul’pin, G. B.; Nizova, G. V.; Kozlov, Y. N.; Arutyunov, V. S.; dos
Santos, A. C. M.; Ferreira, A. C. T.; Mandelli, D. J. Organomet. Chem. 2005,
690, 4498–4504. (m) Mandelli, D.; Steffen, R. A.; Shul’pin, G. B. React. Kinet.
Catal. Lett. 2006, 88, 165–173. (n) Romakh, V. B.; Therrien, B.; Su¨ss-Fink, G.;
Shul’pin, G. B. Inorg. Chem. 2007, 46, 1315–1331. (o) Sibbons, K. F.; Shastri,
K.; Watkinson, M. J. Chem. Soc., Dalton Trans. 2006, 645–661. (p) Smith,
J. R. L.; Murray, J.; Walton, P. H.; Lowdon, T. R. Tetrahedron Lett. 2006, 47,
2005–2008. (r) Tanese, S.; Bouwman, E. AdV. Inorg. Chem. 2006, 58, 29–75.
(10) Gilbert, B. C.; Kamp, N; W, J.; Smith, J. R. L.; Oakes, J. J. Chem.
Soc., Perkin Trans. 2 1997, 2161–2165.
The allylic alcohol 2e with A1,3 strain was used as the model
substrate to test the reactivity (% conversion of the allylic
alcohol 2e), chemoselectivity (allylic CH oxidation versus
epoxidation), and stereoselectivity (threo versus erythro dia-
stereomers) of the catalyst 1 in the presence of a variety of co-
catalysts. In the absence of either the catalyst Mn-tmtacn 1 or
the co-catalyst, the allylic alcohol 2e remained unchanged even
after 9 h of exposure to the oxidation conditions. The results
for the co-catalysts oxalic acid and ascorbic acid and oxalate
and ascorbate buffers are shown in Table 1. The allylic alcohols
2 without allylic strain, with A1,2 or A1,3 strain, and with both
are given in Table 2, for the best set of oxidation conditions
determined in Table 1. We shall now briefly focus on the
important findings in Tables 1 and 2, by separately considering
the reactivity, chemoselectivity, and diastereoselectivity of this
oxidation system.
(11) Wieghardt, K.; Bossek, U.; Nuber, B.; Weiss, J.; Bonvoisin, J.; Corbella,
M.; Vitols, S. E.; Girerd, J. J. Am. Chem. Soc. 1988, 110, 7398–7411.
(12) (a) Morgan, B.; Oehlshla¨ger, A. C.; Stokes, T. M. J. Org. Chem. 1992,
57, 3231–3236. (b) Renz, M. Ph.D. Thesis, University of Wu¨rzburg, 1996. (c)
Adam, W.; Mitchell, C. M.; Paredes, R.; Smerz, A. K.; Veloza, L. A. Liebigs
Ann./Recl. 1997, 1365–1369. (d) Fatiadi, A. J. Synthesis 1976, 65–92. (e) House,
H. O.; Wilkins, J. M. J. Org. Chem. 1978, 43, 2443–2454. (f) Chamberlain, P.;
Roberts, M. L.; Witham, G. H. J. Chem. Soc. B 1970, 1374–1381. (g) Ho, N. H.;
le Noble, W. L. J. Org. Chem. 1989, 54, 2018–2021. (h) Schalley, C. A.;
Schro¨der, D.; Schwarz, H. J. Am. Chem. Soc. 1994, 116, 11089–11097.
1136 J. Org. Chem. Vol. 74, No. 3, 2009

Catalytic Epoxidation of Chiral Allylic Alcohols
TABLE 1. Catalytic Oxidation of the Allylic Alcohol 2e by the Mn-tmtacn Catalyst 1 with Hydrogen Peroxide As Oxygen Source in the
Presence of Various Co-catalystsa
selectivityb
entry
co-catalyst
ascorbic acid
oxalic acid
Na ascorbate
ascorbic acid/Na ascorbate
oxalic acid/ Na oxalate
Na ascorbate
Na ascorbate
Na ascorbate
H2O2 [mol %]
convnb [%]
mbc [%]
chemo 3: 4
diastereo threo:erythro
1
2
3
4
5
6d
7
8
110
110
110
110
110
110
50
66
48
45
69
65
75
25
10
60
66
89
60
70
81
90
90
88:12
91:09
91:09
92:08
91:09
92:08
54:46
40:60
85:15
75:25
83:17
85:15
64:35
70:30
57:43
40:60
20
a Stoichiometric amounts: allylic alcohol 2e (0.67 mmol); catalyst 1 (0.67 µmol, 67 µL of a 0.01 M stock solution in CH3CN); 85% H2O2 (mol %
specified in the third column), diluted with 0.4 mL of CH3CN; co-catalyst (0.225 mol %) in 25 µL of water and 0.5 mL of CH3CN. b Conversion of the
allylic alcohol 2e after complete consumption of the H2O2 (determined by the peroxide test). An aliquot was taken of the crude reaction mixture and the
1
conversions and 3:4 product ratios were assessed by H NMR analysis with 1,3-dichlorobenzene as internal standard; error ≈ (5% of the stated values.
c Mass balance refers to the characterized products 3 and 4 and recovered allylic alcohol 2e. d A 3-fold amount (0.675 mol %) of the co-catalyst was
employed in entry 6, compared to 0.225 mol % used in the other entries.
Reactivity. The catalytic reactivity of the 1-H2O2 system
for the oxidation of the allylic alcohol 2e was studied in the
presence of oxalic acid, ascorbic acid, and oxalate and ascorbate
buffers as co-catalysts. The results in Table 1 show that the
Mn-tmtacn 1/co-catalyst/hydrogen peroxide combination affords
the epoxy alcohol 3e as a major product. The reactivity of the
catalytic system, measured in terms of allylic alcohol conversion,
gave at complete H2O2 consumption the following order for the
various co-catalysts: Na ascorbate (entry 3) e oxalic acid (entry
2) < ascorbic acid (entry 1) ≈ oxalic acid/Na oxalate (entry 5)
e ascorbic acid/Na ascorbate (entry 4). Under these conditions,
the highest reactivity is displayed by the ascorbic acid/Na
ascorbate buffer (69% convn, entry 4), the lowest one by Na
ascorbate alone (45% convn, entry 3). An increased catalytic
activity was obtained when a 3-fold amount of sodium ascorbate
was employed as the co-catalyst (75% convn, entry 6). When,
however, the amount of H2O2 was reduced (entries 3, 7, and
8), expectedly the % convn dropped drastically.
The optimized co-catalyst conditions (Table 1, entry 3) were
then applied to the set of allylic alcohols 2a-f (Table 2) to
assess the structural effects on the reactivity of the present
catalytic system. Depending on the degree and pattern of
substitution of allylic alcohols 2a-f, the conversions vary from
40% to 80%. The reactivity of the catalytic system decreases
with the degree of substitution of the double bond, e.g., 2f ≈
2e < 2d ≈ 2c < 2b ≈ 2a. Thus, the higher its degree of methyl
substitution, the less reactive the allylic alcohol 2, with the
trimethyl derivative 2f the least (40% convn, entry 8) and the
unsubstituted substrate 2a the most reactive (80% convn, entry
1). The most pronounced difference in the substitution pattern
is exhibited for the monosubstituted pair 2b (85% convn, entry
2) and 2c (60% convn, entries 3 and 4).
pointed out, however, that the 3:4 product ratio does not remain
constant during the progress of the reaction. This is best
exemplified by the results for Na ascorbate as co-catalyst (Table
1, entries 3, 7, and 8), in which the initial amount (mol %) of
H2O2 was varied. Whereas for 100 mol % H2O2 (entry 3) the
chemoselectivity is 91:9, for 50 mol % (entry 7) it is only 54:
46, and for 20 mol % H2O2 (entry 8) the ratio even inverts to
40:60. Clearly, the chemoselectivity is a function of the initial
amount of H2O2 used and, thus, on the reaction progress.
The dependence of the chemoselectivity on the structural
variation of the allylic alcohols 2 (Table 2) shows that the
corresponding epoxy alcohols 3a-f are formed preferentially
in this Mn-catalyzed oxidation of the allylic alcohols 2a-f with
H2O2; the exception is substrate E-2c (Table 2, entry 3), for
which enone 4 is favored. In this context, mechanistically quite
significant is the monosubstituted pair of double-bond diaster-
eomers E-2c and Z-2c, for which the E isomer selects the E-4c
enone product (entry 3), whereas the Z isomer favors in much
higher predominance the threo-3c epoxy alcohol (entry 4). This
trend of dependence on the double-bond geometry is also
manifested for the diastereomeric pair of disubstituted substrates
E-2d (entry 5) and Z-2d (entry 6) but is much less pronounced.
In fact, the highest chemoselectivity (95:5) is observed for the
Z-2d diastereomer (entry 6).
Diastereoselectivity. The type and amount of co-catalyst
(Table 1) affects substantially the diastereoselectivity. The
highest threo:erythro ratio (85:15) in favor of the threo
diastereomer was obtained for ascorbic acid (entry 1) and the
ascorbic acid/Na ascorbate buffer (entry 4) Thus, if threo
selectivity is desired, of all the co-catalysts examined herein,
ascorbic acid or the ascorbic acid/Na ascorbate buffer are the
most effective. The threo selectivity drops substantially (70:
30) when a 3-fold amount of sodium ascorbate (entry 6) is
employed. Exceedingly disadvantageous for achieving a high
threo selectivity is to conduct this oxidation at lower than
stoichiometric initial amounts of H2O2, (Table 1) as shown in
entry 7 (50 mol % H2O2) and entry 8 (20 mol % H2O2); in fact,
the erythro isomer is selected in modest preference (40:60) in
entry 8.
Chemoselectivity. The data reveal a relatively small effect
on the chemoselectivity, measured in terms of the 3:4 product
ratio, in regard to the co-catalyst choice (Table 1) and the
structural variation of the allylic alcohols 2 (Table 2). For the
model substrate 2e, the ratio extends from 88:12 for ascorbic
acid (Table 1, entry 1) to 92:8 for the ascorbic acid/Na ascorbate
buffer (Table 1, entry 4), which is hardly significant since it
lies essentially within the experimental error. It should be
J. Org. Chem. Vol. 74, No. 3, 2009 1137

Kilic et al.
TABLE 2. Catalytic Oxidation of the Allylic Alcohol 2 by the Mn-tmtacn Catalyst 1 with Hydrogen Peroxide As Oxygen Source in the
Presence of Sodium Ascorbate As Co-catalysta
a Stoichiometric amounts: allylic alcohol 2 (0.67 mmol); catalyst 1 (0.67 µmol, 67 µL of a 0.01 M stock solution in CH3CN); 85% H2O2 (0.740
mmol), diluted with 0.4 mL of CH3CN; co-catalyst (0.225 mol %) in 25 µL of water and 0.5 mL of CH3CN. b Conversion of the allylic alcohol 2 after
complete consumption of the H2O2 (determined by the peroxide test). An aliquot was taken of the crude reaction mixture and the conversions and 3:4
product ratios were assessed by 1H NMR analysis with 1,3-dichlorobenzene as internal standard; error ≈ (5% of the stated values. c Mass balance
refers to the characterized products 3 and 4 plus recovered allylic alcohol 2. d Isomerization (9%) to the epoxy alcohols threo-Z-3c and erythro-Z-3c was
observed; the epoxy alcohols 3c and 3d are not isomerized under the reaction conditions but are oxidized to the corresponding epoxy enones.
e Isomerization (3%) to the epoxy alcohols threo-E-3c and erythro-E-3c was observed. f Isomerization (11%) to the epoxy alcohols threo-Z-3d and
erythro-Z-3d was observed and their subsequent CH oxidation to the ketone Z-4d. g Isomerization (16%) to the epoxy alcohols threo-E-3d and
erythro-E-3d was observed.
The structural variation, especially the substitution pattern,
discloses a pronounced effect on the diastereoselectivity in
the oxidation of the allylic alcohols 2a-f (Table 2). The
threo:erythro ratio varies from 50:50 (no diastereoselection)
for the unsubstituted substrate 2a (entry 1) to a high of 83:
17 for the most substituted derivatives 2e (entry 7) and 2f
(entry 8), when the oxidation was conducted with sodium
ascorbate as co-catalyst. In the latter two cases, the substrates
2e and 2f possess allylic strain, namely, A1,3 strain for 2e
and both A1,2 and A1,3 strain for 2f. Since for both substrates
the threo:erythro ratio is the same (83:17), it is evident that
A1,2 strain is ineffective in the control of diastereoselection.
This stereochemical trend is also displayed by the derivatives
2b (entry 2, ratio 50: 50) and E-2d (entry 5, ratio 55: 45)
with A1,2 but no A1,3 strain, since their diastereoselectivity
is negligible. Also the diastereomeric pair E-2d (entry 5, 55:
45) and Z-2d (entry 6, 65: 35) confirms that A1,2 strain is
ineffective. The pair of diastereomers E-2c (entry 3, 45: 55)
1138 J. Org. Chem. Vol. 74, No. 3, 2009

Catalytic Epoxidation of Chiral Allylic Alcohols
SCHEME 1. Proposed Mechanism for the Isomerization of the E/Z-2d Allylic Alcohols and Formation of the erythro/threo-E/
Z-3d Epoxy Alcohols and E/Z-4d Enones
and Z-2c (entry 4, 65: 35) corroborate these findings, but
in the initial phase of the reaction, during which catalyst 1 is
transformed into the bis(µ-carboxylato)-MnIII2 complex A by a
poorly understood reaction, in which presumably oxyl (RO•)
radicals are generated from hydrogen peroxide as active oxygen
species. At this point, we have no direct experimental evidence
(e.g., EPR spectroscopy) as to the nature of the RO• radicals.
We speculate that metal-free (e.g., HOO•) species might
intervene as active oxygen entities. Alternatively, high-valent
manganese oxo complexes may also be involved in the initial
the control by A1,3 strain is modest.
Discussion
The prominent trend displayed in Tables 1 and 2 is the
fact that the conversion of the allylic alcohols decreases while
the diastereoselectivity rises with the degree of substitution
of electron-donating methyl groups at the double bond.
Clearly, the allylic alcohols 2e and 2f, the most electron-
rich substrates, show a lower reactivity. Bein7 has reported
the epoxidation of a series of alkenes with catalyst 1 and
hydrogen peroxide and showed that the alkene structure plays
an important role in the oxidation rate. For example,
1-methylcyclohexene is known to be more reactive than
cyclohexene in the peracid epoxidation, but with the Mn-
tmtacn complex 1 and hydrogen peroxide, the reactivity order
is reversed. These results show that the more accessible but
less electron-rich double bond is more readily oxidized.
Presumably, steric repulsion between the catalytically active
Mn complexes and the double-bond substituents hinders the
transfer of the oxygen atom. Although it is often assumed
that the nature of the catalytically active Mn species is of
the high-valent oxo-Mn type,13 recent work on the present
system, namely, the combination Mn-tmtacn 1/carboxylic acid
co-catalyst/hydrogen peroxide, has convincingly demonstrated
stage of the reaction, en route to the bis(µ-carboxylato)-MnIII
2
formation. As mechanistically implicated in Scheme 1, the
intervening substrate-derived C-centered radicals possess suf-
ficient lifetime for substantial C-C bond rotation prior to
collapse to the epoxide ring.14
We propose that the lack of chemo- and diastereoselectivity,
observed initially at low conversion of substrate 2e (Table 1,
entries 7 and 8), may derive from unselective reactions mediated
by RO• radicals in this phase of the allylic alcohol oxidation.
Once catalyst 1 has been efficiently transformed into bis(µ-
carboxylato)-MnIII complex A, a selective epoxidation domi-
2
nates, catalyzed by the latter species. This accounts for the much
higher selectivities observed at high conversion of substrate 2e
(see Table 1, entry 3 versus entries 7 and 8). The fact that the
diastereoselectivity for the substrate 2e varies with different
carboxylate co-catalysts (Table 1) is in line with the role of
bis(µ-carboxylato)-MnIII complex A as catalytically active
2
the role of the catalytically active bis(µ-carboxylato)-MnIII
2
species.
complex A (see Figure 1).8c-e This also holds for the
ascorbate and oxalate co-catalysts, both of which not only
allow in situ generation of bis(µ-carboxylato)-MnIII2 complex
A but also are likely to facilitate reduction of catalyst 1.
Thereby, the catalyst is activated toward ligand exchange to
form the catalytically active state.8e
Hard to rationalize are the widely varying chemo- and
diastereoselectivities as a function of the allylic alcohol structure.
The fact, nevertheless, that the threo diastereoselectivity domi-
nates in the epoxidation of the allylic alcohols Z-2c and Z-2d
indicates that hydrogen bonding operates between the allylic
hydroxy group and the manganese complex in the oxygen-
transfer process. Therefore, in the absence of any further
experimental evidence, we suggest the transition-state geometry
B (Figure 2) for the oxygen transfer from complex A (Figure
1) to the allylic alcohol Z-2c, to account for the observed threo
selectivity.
The fact that cis/trans isomerization is observed in the
epoxidation of the allylic alcohols E-2c, Z-2c, E-2d, and Z-2d
indicates that at least part of the oxygen-atom transfer proceeds
through radical intermediates. The oxygen-atom transfer from
H2O2, catalyzed by the aforementioned bis(µ-carboxylato)-MnIII
2
complex A, is a concerted process without cis/trans isomeriza-
tion.8c-e It is, thus, likely that the cis/trans isomerization occurs
Conclusion
(13) (a) Smith, J. R. L.; Gilbert, B. C.; Mairata i Payeras, A.; Murray, J.;
Lowdon, T. R.; Oakes, J.; Pons i Prats, R.; Walton, P. H. J. Mol. Catal. A:
Chem. 2006, 251, 114–122. (b) Groves, J. T.; Lee, J.; Marla, S. S. J. Am. Chem.
Soc. 1997, 119, 6269–6273. (c) Bennur, T. H.; Srinivas, D.; Sivasanker, S.;
Puranik, V. G. J. Mol. Catal. A: Chem. 2004, 219, 209–216. (d) Bennur, T. H.;
Sabne, S.; Deshpande, S. S.; Srinivas, D.; Sivasanker, S. J. Mol. Catal. A: Chem.
2002, 185, 71–80. (e) Gilbert, B. C.; Smith, J. R. L.; Mairata i Payeras, A.;
Oakes, J. Org. Biomol. Chem. 2004, 2, 1176–1180.
The current experimental results in Tables 1 and 2 reveal
that the chemo- and diastereoselectivities in the epoxidation of
(14) (a) Jacobsen, E. N.; Deng, L.; Furukawa, Y.; Martinez, L. E. Tetrahedron
1994, 50, 4323–4334. (b) McGarrigle, E. M.; Gilheany, D. C. Chem. ReV. 2005,
105, 1563–1602.
J. Org. Chem. Vol. 74, No. 3, 2009 1139

Kilic et al.
in Tables 1 and 2 in 25 µL of water. The reaction was initiated by
addition of 0.4 mL of CH3CN containing hydrogen peroxide (85%,
20 µL, 0.737 mmol) with the aid of a syringe within the time
specified in Tables 1 and 2 at 0-2 °C, while the reaction progress
was monitored by peroxide test (KI/AcOH). After completion of
the reaction, the reaction mixture was diluted with 3 mL of CH2Cl2,
dried over Na2SO4, and passed through a short silica gel column
(150 mg). The solvent was removed under reduced pressure (25
°C, 450 Torr), and the conversion, mass balance, and chemo- and
diastereoselectivities were determined by 1H NMR analysis directly
on the crude mixture (Tables 1 and 2). The epoxy alcohols 3a-f
and ketones 4a-f are literature known and have been identified on
the basis of their NMR data.
chiral allylic alcohols 2 with hydrogen peroxide as oxygen
donor, catalyzed by the Mn-tmtacn catalyst 1 plus carboxylic
acid co-catalysts, vary during the course of the reaction.
Moreover, the observed diastereoselectivity for the substrates
with 1,3-allylic strain depends on the amount and nature of the
co-catalyst. We postulate that the enhanced selectivities with
progressing conversion arises from the transformation of catalyst
1 into the more selective bis(µ-carboxylato)-MnIII2 complex A.
In the initial phase, this process is accompanied by the
generation of oxygen-centered radical species, which are
responsible for the less selective oxidations. Additional experi-
ments should be helpful in clarifying our mechanistic specula-
tions, especially the involvement of radicals. Further mechanistic
understanding of the various selectivities would enable optimiz-
ing the synthetic value of this attractive catalytic oxidation.
Acknowledgment. This paper is dedicated to Professor Metin
Balci, Middle East Technical University, on the occasion of
his 60th birthday. We thank the European Commission (SUS-
TOX, G1RD-CT-2000-00347) for generous financing. Addition-
ally, W.A. appreciates the continuous support from the Deutsche
Forschungsgemeinschaft, Alexander von Humboldt-Stiftung, and
the Fonds der Chemischen Industrie.
Experimental Section
Typical Procedure for the Catalytic Epoxidation of the
Allylic Alcohols Manganese Complexes 1 and Hydrogen
Peroxide. Into a 5-mL, two-necked, round-bottomed flask equipped
with a thermostat and magnetic stirrer were placed 0.5 mL of
CH3CN, 0.67 mmol allylic alcohol, 76.3 µL (0.67 mmol) of 1,3-
dichlorobenzene (as internal standard), 67 µL of stock solution of
catalyst 1 (1 µmol/100 µL in CH3CN), and co-catalyst as specified
Supporting Information Available: Characterization data
for the epoxy alcohols 3a-f and structure matrix. This material
is available free of charge via the Internet at http://pubs.acs.org.
JO801974E
1140 J. Org. Chem. Vol. 74, No. 3, 2009