Manganese porphyrin hosts as epoxidation catalysts - Activity and stability control by axial ligand effects

Article abstract of DOI:10.1002/ejoc.200600648

The catalytic performance of a cavity-containing [(porphyrin)Mn] in the epoxidation of alkenes is described. Inside the catalyst cavity, nitrogen-containing axial ligands can be bound to the porphyrin metal with high association constants, resulting in strong activation of the catalyst in the presence of only one equivalent of the ligand. Complexation of a sterically demanding ligand to the outside of the cavity-containing catalyst can prevent catalyst decomposition through the formation of ν-oxo dimers. Wiley-VCH Verlag GmbH & Co. KGaA, 2007.

Full text of DOI:10.1002/ejoc.200600648

FULL PAPER
DOI: 10.1002/ejoc.200600648
Manganese Porphyrin Hosts as Epoxidation Catalysts – Activity and Stability
Control by Axial Ligand Effects
Johannes A. A. W. Elemans,*[a] Edward J. A. Bijsterveld,[a] Alan E. Rowan,*[a] and
Roeland J. M. Nolte[a]
Keywords: Manganese porphyrins / Supramolecular chemistry / Host–guest chemistry / Catalysis / Epoxidation
The catalytic performance of
a
cavity-containing [(por-
of a sterically demanding ligand to the outside of the cavity-
containing catalyst can prevent catalyst decomposition
through the formation of µ-oxo dimers.
phyrin)Mn] in the epoxidation of alkenes is described. Inside
the catalyst cavity, nitrogen-containing axial ligands can be
bound to the porphyrin metal with high association con-
stants, resulting in strong activation of the catalyst in the
presence of only one equivalent of the ligand. Complexation
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2007)
preventing its dimerization into an unreactive and autodes-
tructive µ-oxo MnIV-porphyrin dimer. A drawback for the
optimal use of this system, however, is the need for a large
excess (ca. 500 equiv.) of axial ligand because of its rela-
tively weak binding to the porphyrin metal.
We have developed metal-porphyrins each appended with
a diphenylglycoluril-based cavity, which effectively shields
one side of the porphyrin plane.[16] The cavity is capable of
binding N-containing axial ligands very strongly as the re-
sult of a combination of metal–ligand coordination and
cavity-filling effects. Here the catalytic performance of the
cavity-containing porphyrin catalyst Mn1 in the epoxid-
ation of alkenes is described in terms of activity, selectivity
and stability, in the presence of several strongly binding ax-
ial ligands.[17]
Introduction
A wide variety of metalloporphyrins function as active
sites in enzymes, such as in cytochrome P-450.[1] This
monooxygenase oxidizes a range of alkanes and alkenes
and is of fundamental importance for detoxification in the
human body. The active site of the enzyme contains a FeIII
protoporphyrin IX to which a cysteinate ligand essential for
catalytic activity is axially coordinated.[2] During the past
decade, a myriad of synthetic model systems that mimic
the activity and selectivity of the natural enzyme have been
developed,[3] although because of the only moderate sta-
bility of FeIII-porphyrins in combination with cysteinate ax-
ial ligands, many of these model systems have been based
on more stable and versatile MnIII-porphyrins and N-con-
taining axial ligands.[4] The use of molecular oxygen as an
oxidant in model systems has appeared to be troublesome,
because both the catalyst and the oxidant have to be acti-
vated by a reducing agent,[5] and so a variety of single-oxy-
gen donor species – such as hypochlorites,[6] iodosyl-
arenes,[7] peracids,[8] monopersulfates,[9] pyridine N-ox-
ides,[10] alkyl hydroperoxides[11] and hydrogen peroxide –
have been used as alternatives.[12]
In our[13] and other groups[14] the use of manganese por-
phyrins as oxidation catalysts in combination with hypo-
chlorite as oxidant has been extensively studied. The pres-
ence of axial ligands such as pyridines or imidazoles ap-
peared to enhance the reaction rates and the stereoselecti-
vity of the reaction[15] by facilitating the formation and sta-
bilization of the proposed formally MnV=O intermediate,
Results and Discussion
[a] Institute for Molecules and Materials, Radboud University
Nijmegen,
Synthesis
Toernooiveld, 6525 ED Nijmegen, The Netherlands
E-mail: J.Elemans@science.ru.nl
The synthesis of H21, the free-base precursor of Mn1,
A.Rowan@science.ru.nl
has been described before,[16] but an alternative synthetic
Eur. J. Org. Chem. 2007, 751–757
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
751

J. A. A. W. Elemans, A. E. Rowan et al.
FULL PAPER
route that has dramatically increased the yield of H21 (see reference material.[20] 1H NMR and UV/Vis titrations were
Scheme 1) has been developed. Previously, H21 was synthe- carried out to establish the binding strength of Zn1 with a
sized in two steps from the tetratosylate derivative 2 by a number of pyridine and imidazole derivatives (Table 1), and
quadruple alkylation of this compound with salicylic alde- pyridine (Py) was found to bind in the cavity of Zn1 very
hyde and a subsequent condensation of the resulting tetra- strongly (Figure 1, a), an event accompanied by large up-
aldehyde with pyrrole to close the porphyin “roof”.[16] The field shifts in the crown ether proton signals of the host. In
overall yield of these two reactions amounted to 3.8%. In contrast with this very strong binding, the Ka value between
the new procedure, host 2 was treated directly with one the bulky ligand 4-tert-butylpyridine (Bupy) and Zn1 is low,
equivalent of the meso-tetrakis(2-hydroxyphenyl)porphyrin which is attributed to Bupy being able to bind only to the
3 under high-dilution conditions in acetonitrile in the pres- outside of Zn1 and hence not experiencing the stabilizing
ence of base to yield H21 in a yield of 30%. Mn1 was syn- effects of the cavity (Figure 1, b). This was also evident
thesized from H21 by treatment with Mn(OAc)2·4H2O in from the absence of any shifts of the receptor side-wall and
DMF at reflux and subsequent ion exchange by stirring a crown ether spacer proton signals in the NMR spectrum
solution of the product in chloroform with aqueous sodium upon the addition of this ligand. Substitution of the pyri-
chloride.
dine ring with a meta-hydroxy group (OHPy) causes an
enormous increase in binding strength relative to that of Py,
because the ligand, as well as coordinating to the metal,
simultaneously forms a strong hydrogen bond with one of
the carbonyl groups of Zn1. 3-(Acetylamino)pyridine
(AcAmPy) and 3-(4-toluoylamino)pyridine (TolAmPy) can
also hydrogen bond to the carbonyl groups, but these hy-
drogen bonds are apparently weaker, since Ka values similar
to that between Zn1 and Py are observed. Imidazole (Im)
forms a stronger complex than Py with Zn1, which is not
unexpected since the latter ligand is a weaker base. Molecu-
lar modelling studies revealed Im to be too small to coordi-
nate to the zinc ion and at the same time form a hydrogen
bond with one of the carbonyl groups of the host, but a
Scheme 1.
Binding of Axial Ligands
The glycoluril-based cavity of Mn1 has a diameter of ap-
proximately 9 Å and is capable of accommodating small ax-
ially binding ligands. Since their binding strength would be
expected to have a great influence on the catalytic proper-
ties of Mn1, it was desired to determine the association con-
stants (Ka values) between these ligands and the host. How-
ever, since the determination of association constants of
paramagnetic manganese porphyrins by NMR and UV/Vis
Figure 1. Approaches in which Mn1 is used as an epoxidation cata-
lyst in combination with Py (left, binding inside the cavity) or with
BuPy (right, binding outside the cavity) as the axial ligands. The
spectroscopy is very difficult, cavity Zn1[16] was used as a arrows indicate where the epoxidation reaction takes place.
Table 1. Association constants (Ka [–1]) of complexes formed between Zn1 and various axial ligands,[a] and epoxidation of olefins with
Mn1 and one equivalent of axial ligand.[b]
Ligand[c]
Ka
Epoxidation of α-pinene
k0
Epoxidation of cis-stilbene
k0
[d]
[d]
Yield[e]
Yield[e]
c:t[f]
[h]
[h]
[h]
[h]
[h]
Py
1.1·105 [g]
12
11
34
–
–
17
8
81
82
44
–
–
82
38
20
12
21
15
28
–
–
57
57
18
12
37
96:4
[i]
BuPy
OHPy
AcAmPy
TolAmPy
Im
625
90:10
3.0·107 [j]
9.0·104 [g]
1.2·105 [g]
3.2·105 [g]
7.0·104 [g]
–
[k]
[k]
–
96:4
[k]
[k]
[k]
–
–
–
[k]
[k]
[k]
MeIm
–
[a] In chloroform solution at 298 K. [b] Standard reaction conditions. [c] See text for abbreviations. [d] Initial rate·105 moldm–3 s–1. [e]
Yield (%) after 3 h. [f] Ratio cis/trans epoxide product after 3 h. [g] Estimated error 20%. [h] Rate of formation of the cis epoxide. [i]
Estimated error 10%. [j] Estimated error 30%. [k] Not determined.
752
www.eurjoc.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2007, 751–757

Manganese Porphyrin Hosts as Epoxidation Catalysts
FULL PAPER
clear drop in binding strength was observed when N- Catalysis Studies
methylimidazole (MeIm) was used instead of Im as the li-
gand.
The strong binding of axial ligands in the cavity of Zn1
is of great interest with regard to catalytic epoxidation reac-
tions with Mn1 as a catalyst, since as a result of the high
binding constant it might be expected that only one equiva-
lent of such a ligand might be needed to achieve Ն99%
binding to the porphyrin metal. To investigate this assump-
tion, the epoxidation of α-pinene was carried out with host
Mn1 as the catalyst and porphyrin MnTPP as the reference,
in the presence of one equivalent of Py (Figure 1, a). Mn1
appeared to display an initial rate ten times higher than that
of MnTPP (Figure 3, Table 2), as a result of which it re-
quires more than 10 h for the reaction to be complete in
the case of MnTPP, but only 2 h in that of Mn1. The initial
epoxidation rate is generally related to the binding constant
of the axial ligand [i.e., the rate increases when the Ka value
of the ligand (measured with the related host Zn1) becomes
higher (Table 1)].[24] It is notable that the most strongly
binding ligand OHPy gives very high initial rates, but also
the lowest yield after 3 h; the reason for this is probably the
gradual expulsion of the ligand out of the cavity of Mn1 as
a result of deprotonation by the strongly basic OCl– spe-
cies.[25] Such deprotonation was also observed for AcAmPy
and to a lesser extent for TolAmPy. Surprisingly, deproton-
ation did not occur when Im was used, and in the case of
this ligand the epoxidation reaction was complete within
1 h, giving an 82% yield of epoxide. In addition to a high
epoxidation rate, a further advantage of the strong binding
of Im is that its degradation (by oxidation), which in other
systems has already been observed to occur at relatively low
concentrations of the ligand,[26] is minimal. Apparently, its
binding in the cavity of Mn1 protects it from this undesired
reaction.
1
The H NMR spectra of the 1:1 host–guest complexes
formed between Zn1 and the two imidazole derivatives in
CDCl3 appeared to be strikingly different. In the spectra of
the complex formed between Zn1 and MeIm, all crown
ether proton signals and the cavity side-wall proton signals
of Zn1 were shifted strongly upfield in relation to these sig-
nals in the free host, pointing to a host–guest geometry in
which the aromatic surface of MeIm is oriented parallel to
the cavity side-walls of Zn1 (Figure 2, a), similarly to the
geometries of the host–guest complexes formed between
Zn1 and the pyridine ligands. In contrast with this, the
crown ether and side-wall proton signals in the complex
formed between Zn1 and Im were shifted only slightly up-
field, indicating that the aromatic surface of Im is oriented
perpendicular to the aromatic surface of the cavity side-
walls (Figure 2, b). A 2D NOESY spectrum revealed that
the ligand was still exclusively bound within the cavity and
not to the outside of the porphyrin roof. It is not unlikely
that the reason for the deviating binding geometry is the
presence of an N–H···π interaction[21] between one of the
cavity side-walls[22] and the Im NH proton.[23]
Figure 3. Product formation curves for the epoxidation of α-pinene,
in the presence of one equivalent of an axial ligand. Mn1-Im (᭡),
Mn1-Py (᭹), Mn1-OHPy (o), Mn1-MeIm (+), MnTPP-Py (᭿).
In order to provide better understanding of the catalytic
performance of Mn1 in terms of activity and stereoselecti-
vity, additional epoxidation experiments with manganese
porphyrin catalysts and one equivalent of pyridine were car-
ried out (Table 2). It is clear that the reference compound
MnTPP is a very poor catalyst under these conditions. To
rule out any electronic effects, MnTMPP was taken as a
Figure 2. a) Computer-modelled host–guest geometry of the com-
plex formed between Zn1 and MeIm. b) Idem, of the host–guest
complex formed between Zn1 and Im, illustrating the possible N–
H···π interaction between the guest and one of the side-walls of the
host. c) Schematic picture of the coordination of Im to the zinc ion
in the porphyrin roof of Zn1 and possible simultaneous interaction
between the Im 3-NH group and a cavity side-wall and the Im 4-
CH group with the other cavity side-wall.
Eur. J. Org. Chem. 2007, 751–757
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
753

J. A. A. W. Elemans, A. E. Rowan et al.
FULL PAPER
Table 2. Epoxidation of olefins by Mn1 and the reference catalysts MnTPP and MnTMPP in the presence of one equivalent of Py as the
axial ligand.[a]
Substrate
Mn1
MnTPP
MnTMPP
Yield[c]
c:t[d]
k0
Yield[c]
c:t[d]
k0
Yield[c]
c:t[d]
[b]
[b]
[b]
k0
[e]
[e]
[e]
α-Pinene
cis-Stilbene
trans-Stilbene
12
81
57
72
1
10
39
9
–
–
–
20[f]
19
96:4
–
4[f]
4
65:35
–
3[f]
Ͻ 1
33
8
63:37
[g]
[g]
[g]
–
[a] Standard reaction conditions. [b] Initial rate·105 moldm–3 s–1. [c] Yield (%) after 3 h. [d] Ratio cis-trans epoxide product after 3 h. [e]
Not determined. [f] Rate of formation of the cis epoxide. [g] No cis epoxide was detected.
second reference, but this catalyst showed even lower rates species, but the catalyst then rapidly decomposes once all
of epoxidation than MnTPP. These results demonstrate that the substrate is converted, as is visible in the bleaching of
the enhanced rate of epoxidation is produced primarily by the brown reaction mixture. This decomposition also oc-
the strong binding of pyridine inside the cavity of Mn1 and curred in the case of the systems of Mn1 with one equiva-
not just by electronic effects of the porphyrin. In the case lent of a strongly binding axial ligand present.
of the latter catalyst it can also be seen that the epoxidation
is considerably more stereoselective for the conversion of
cis-stilbene into cis-stilbene oxide, an effect that can be in-
terpreted in terms of the strong coordination of the axial
ligand, which pulls the metal centre into the plane of the
porphyrin and thereby hinders the rotation of the C–C
bond of the cis-stilbene substrate coordinated to the other
face of the porphyrin in the transition state (Figure 4).[15]
To prevent this decomposition the bulky axial ligand
Bupy was used in combination with Mn1, as it was rea-
soned that coordination of this ligand (which does not fit
within the cavity of the catalyst) would efficiently prevent
µ-oxo dimer formation, including after all of the substrate
had been converted, since the other face of the porphyrin is
protected by the receptor cavity (Figure 1, b). Indeed, when
epoxidation experiments with Mn1 were carried out in the
presence of 500 equiv. of Bupy (Table 3), catalyst destruc-
tion did not occur, as was already indicated by the fact that
the organic layer retained its brown colour even after the
reaction mixture had been stirred in the absence of sub-
strate for more than a week. More importantly, however,
was the observation that freshly added portions of substrate
were epoxidized with similar initial rates as the first portion
without any apparent deterioration of the catalyst, thus
affording high turnover numbers per Mn1 molecule
(Ͼ1000).[27] Under the same conditions, MnTPP and
MnTMPP were not stabilized by Bupy and decomposed
Figure 4. Proposed effect of an axial ligand L on the stereospecific-
ity in the epoxidation of cis-stilbene a) in the absence of an axial
ligand, and b) in the presence of an axial ligand, which pulls the
metal center into the plane of the porphyrin, thereby sterically hin- rapidly.
dering the rotation of the C–C bond of the substrate coordinated
Whereas the initial conversion rates for the epoxidation
of α-pinene and trans-stilbene do not significantly differ for
to the other face of the porphyrin in the transition state.[15]
the three catalysts, there is a clear drop in rate for the epox-
A major drawback of MnTPP and MnTMPP appeared idation of cis-stilbene, going from MnTPP Ͼ MnTMPP
to be their decomposition, attributed to the formation of Ͼ Mn1. This decrease coincides with an increase in steric
[(P)MnIV–O–MnIV(P)]2+ µ-oxo dimers, during the catalytic hindrance imposed by the substituents of the meso-phenyl
reaction. These dimers are unreactive in further catalysis rings of the catalysts, which apparently only affects the con-
and decompose through electrophilic attack on the elec- version of cis- and not of trans-stilbene. The very low rate
tron-rich meso positions in the porphyrin ring.[13c] When a observed when Mn1 is the catalyst can be explained by the
reactive substrate is present, it generally competes success- fact that the rather bulky cis-stilbene would be expected to
fully with a [(P)MnIIIX]+ molecule for the reactive MnV-oxo experience severe hindrance by the confinement of the host
Table 3. Epoxidation of olefins by Mn1 and the reference catalysts MnTPP and MnTMPP in the presence of 500 equiv. of BuPy as the
axial ligand.[a]
Substrate
Mn1
MnTPP
MnTMPP
[b]
[b]
[b]
k0
Yield[c] c:t[d]
k0
Yield[c] c:t[d]
k0
Yield[c] c:t[d]
[e]
[e]
[e]
α-Pinene
cis-Stilbene
trans-Stilbene
11
82
57
72
13
57f
21
80
70
63
–
–
–
12[f]
24
90:10
–
90:10
–
39[f]
20
52
65
92:8
[g]
[g]
[g]
–
[a] Standard reaction conditions. [b] Initial rate·105 moldm–3 s–1. [c] Yield (%) after 3 h. [d] Ratio cis-trans epoxide product after 3 h. [e]
Not determined. [f] Rate of formation of the cis epoxide. [g] No cis epoxide was detected.
754
www.eurjoc.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2007, 751–757

Manganese Porphyrin Hosts as Epoxidation Catalysts
FULL PAPER
Melting points were recorded on a Jeneval THMS 600 hot-stage
polarization microscope and are uncorrected. Molecular modelling
calculations were performed on a Silicon Graphics Indigo II work
station with use of the CHARMm force field.[18]
cavity, which it has to enter completely to reach the reactive
manganese oxo species, whereas molecular modelling stud-
ies show this steric hindrance for the much flatter trans iso-
mer to be negligible (Figure 5).
Syntheses
meso-Tetrakis(2-hydroxyphenyl)porphyrin (H2THPP): meso-Tet-
rakis(2-methoxyphenyl)porphyrin[19] (1.15 g, 1.6 mmol) was dis-
solved in CH2Cl2 (20 mL), the solution was cooled to –25 °C, and
BBr3 (3 mL) was added. This mixture was slowly warmed to room
temperature, stirred overnight under a nitrogen atmosphere and
subsequently poured into ice/water (150 mL), to which ethyl acetate
(350 mL) was added. NaHCO3 was added carefully until the or-
ganic layer became purple, the organic phase was washed with satu-
rated aqueous NaHCO3 and dried with MgSO4, and the solvents
were evaporated to dryness. The product was purified by column
chromatography [silica, 5% MeOH in CHCl3 (v/v)] to yield
H2THPP (0.98 g, 92%) as a purple solid. 1H NMR (CDCl3,
300 MHz): δ = 8.91 (s, 8 H, β-pyrrole), 8.02–7.91 (m, 4 H, ArH),
7.79–7.67 (m, 4 H, ArH), 7.42–7.28 (m, 8 H, ArH), 6.9–6.8 (br. s,
4 H, OH), –2.73 (br. s, 2 H, NH) ppm. FAB-MS: m/z = 679 [M +
H]+.
Figure 5. a) Computer-modelled space-filling structure of cis-stil-
bene (side view and view along the double bond). b) Idem, of trans-
stilbene. c) Computer-modelled space-filling structure of Mn1
showing the size of the cavity. All compounds are displayed on the
same scale.
Conclusions
Porphyrin Host H21: A solution of host 2 (750 mg, 0.55 mmol),
porphyrin 3 (375 mg, 0.55 mmol) and K2CO3 (750 mg, 5.4 mmol)
in degassed acetonitrile (700 mL) was heated at reflux under nitro-
gen for 16 h. After cooling, the mixture was filtered, the filtrate was
evaporated to dryness, and the product was purified by column
chromatography (alumina Act. III, CHCl3) to yield H21 (220 mg,
30%) as a purple powder. Physical properties were in agreement
with those previously reported.[16]
We have shown that very strong binding of axial ligands
can be achieved through their complexation to a manganese
porphyrin functionalized with a substrate binding cavity
based on diphenylglycoluril. As a result of this strong bind-
ing, only one equivalent of ligand is required to achieve a
Ͼ99+% occupation of the manganese metal, which is re-
flected in a high activity in the catalytic epoxidation of ole-
fins by this system. In a different approach, a bulky axial
ligand was complexed exclusively to the outside of the cav-
ity, which effectively protected the catalyst from µ-oxo di-
mer formation and thus assured its stability over several
rounds of catalysis. Currently, the latter system is being em-
ployed by our group as a very stable processive catalyst in
the epoxidation of polyolefin substrates.[28]
Porphyrin Host Mn1: A degassed solution of H21 (25 mg, 19 µmol)
in DMF (2 mL) was heated to reflux. Mn(OAc)2·4H2O (10 mg,
41 µmol) was added and the mixture was heated at reflux under
nitrogen for 2 h. After cooling, the solvent was evaporated to dry-
ness and the residue was dissolved in CHCl3 (10 mL). A saturated
aqueous NaCl solution (10 mL) was added, the mixture was stirred
vigorously for 16 h, the organic layer was extracted with water
(3ϫ50 mL), and the solvents were evaporated to dryness. The resi-
due was purified by column chromatography (silica, CHCl3/MeOH
9:1, v/v, Rf = 0.17), the product was dissolved in a minimal amount
of CHCl3, and this solution was added dropwise to stirred n-hex-
ane. After centrifugation, the product was dried under vacuum to
yield Mn1 (26 mg, 97%) as dark green needles. M.p. Ͼ 400 °C. IR
Experimental Section
General: All solvents were distilled under nitrogen prior to use.
Acetonitrile was distilled from CaH2, chloroform from CaCl2,
dichloromethane from CaH2, and DMF from BaO. K2CO3 was
dried in an oven (150 °C). Pyrrole was distilled at room temperature
under reduced pressure prior to use. All other solvents and chemi-
cals were commercial materials and were used without further puri-
fication. Merck silica gel (60 H) was used for column chromatog-
raphy and Merck silica gel plates (F254) for thin layer chromatog-
(KBr pellet): ν = 3048 (ArH), 2923, 2854 (CH ), 1696 (C=O), 1600,
˜
2
1580, 1513, 1462, 1449, 1427 (C=C), 1308, 1279, 1249, 1204 (CH2),
1107, 1064, 1010 (COC) cm–1. UV/Vis (CH2Cl2): λ/nm [log-
(ε/–1 cm–1)] = 346 (4.55), 375 (4.66), 409 (4.54), 450 (3.70), 480
(4.96), 527 (3.40), 586 (3.54), 617 (3.50). FAB-MS: m/z = 1397 [M–
Cl]+. C84H62N8O10MnCl (1433.85): calcd. C 70.44, H 4.29, N 7.82;
found: C 70.12, H 4.14, N, 8.12.
1
raphy. H NMR and 13C NMR spectra were recorded on Bruker
3-Acetyliminopyridine (AcAmPy): 3-Aminopyridine (1.00 g,
10.6 mmol) was dissolved in acetone (150 mL), and K2CO3 (7.4 g,
54 mmol) was added, followed by the dropwise addition of acetyl
chloride (2.5 g, 32 mmol) in acetone (5 mL). The reaction mixture
was stirred for 3 h, after which it was quenched by the addition of
water and the solvent was removed by evaporation. The residue
was taken up in CH2Cl2 and this solution was washed with water,
while the organic layer was dried with Na2SO4, and after filtration
was evaporated to dryness. The product was purified by column
chromatography (silica, 5% MeOH in CHCl3), and then recrys-
DPX 200, Bruker DMX 300 and Bruker DRX 500 instruments.
Chemical shifts are reported in ppm downfield from internal TMS
(δ = 0.00 ppm) in the case of 1H NMR spectra in CDCl3. In the
case of 13C NMR spectra, the solvent peak was used as a reference
(CDCl3: δ = 77.0 ppm). Symbols used are: s = singlet, d = doublet,
dd = doublet of doublets, m = multiplet, br = broad. EI and FAB
mass spectra were recorded on a VG 7070E and a Finnigan
MAT 900 S instrument. The matrix used for FAB was m-nitroben-
zyl alcohol. Elemental analyses were determined with a Carbo
Erba Ea 1108 instrument. UV/Vis spectra were measured on a Var-
tallized from toluene. Yield: 1.18 g (82%). M.p. 131.5 °C. 1H NMR
3
ian Cary 50 UV/Vis spectrophotometer. GC spectra were measured (CDCl3, 300 MHz): δ = 8.55 (s, 1 H, PyH-2), 8.35 (d, J = 4.1 Hz,
on a Varian GC3800 instrument with a 8200 CX autosampler.
1 H, PyH-6), 8.18 (d, 3J = 8.2 Hz, 1 H, PyH-4), 7.80 (brs, 1 H,
Eur. J. Org. Chem. 2007, 751–757
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
755

J. A. A. W. Elemans, A. E. Rowan et al.
FULL PAPER
[2] B. Meunier, S. P. de Visser, S. Shaik, Chem. Rev. 2004, 104,
3947.
NH), 7.40–7.20 (m, 1 H, PyH-3), 2.22 (s, 3 H, CH3) ppm;
13C{1H}NMR (CDCl3, 75 MHz): δ = 169.25, 144.83, 140.93,
135.25, 127.44, 123.82, 24.17 ppm. EI-MS: m/z = 136 [M]+.
C7H8N2O (136.15): calcd. C 61.75, H 5.92, N 20.58; found: C
62.10, H 5.91, N 19.81.
[3] a) M. C. Feiters, A. E. Rowan, R. J. M. Nolte, Chem. Soc. Rev.
2000, 29, 375; b) H. C. Zhou, J. T. Groves, J. Porphyrins Phthal-
ocyanines 2004, 8, 125; c) W. D. Woggon, Acc. Chem. Res. 2005,
38, 127; d) S. Kozuch, T. Leifels, D. Meyer, L. Sbaragli, S.
Shaik, W. D. Woggon, Synlett 2005, 674; e) E. Rose, B. Andri-
oletti, S. Zrig, M. Quelquejeu-Ethève, Chem. Soc. Rev. 2005,
34, 573.
3-(4-Toluoylamino)pyridine (TolAmPy): 3-Aminopyridine (5.00 g,
53.1 mmol) was dissolved in pyridine (60 mL), p-toluoyl chloride
(15 mL, 113 mmol) was added slowly to this solution, and the mix-
ture was stirred overnight at room temperature. Aqueous HCl (1 ,
100 mL) and ethyl acetate (100 mL) were added. After phase sepa-
ration, the aqueous phase was neutralized with aqueous NaOH
(1 ), the mixture was filtered, and the residue was recrystallized
from water, affording TolAmPy. Yield: 3.86 g (34%). M.p. 126.5 °C.
1H NMR ([D6]DMSO, 200 MHz): δ = 10.45 (brs, 1 H, NH), 8.98
[4] B. Meunier, Chem. Rev. 1992, 92, 1411.
[5]
[6]
I. Tabushi, Coord. Chem. Rev. 1988, 86, 1.
a) J. P. Collman, J. L. Brauman, B. Meunier, T. Hayashi, T.
Kodadek, J. Am. Chem. Soc. 1985, 107, 2000; b) F. Montanari,
M. Penso, S. Quici, P. Vigano, J. Org. Chem. 1985, 50, 4888.
a) J. T. Groves, W. J. Kruper, R. C. Haushalter, J. Am. Chem.
Soc. 1980, 102, 6375; b) C. L. Hill, B. C. Schardt, J. Am. Chem.
Soc. 1980, 102, 6374; c) A. J. Castellino, T. C. Bruice, J. Am.
Chem. Soc. 1988, 110, 158; d) J. P. Collman, X. Zhang, R. T.
Hembre, J. I. Brauman, J. Am. Chem. Soc. 1990, 112, 5356; e)
E. Baciocchi, T. Boschi, L. Cassioli, C. Galli, L. Jaquinod, A.
Lapi, R. Paolesse, K. M. Smith, P. Tagliatesta, Eur. J. Org.
Chem. 1999, 3281.
[7]
3
(s, 1 H, PyH-2), 8.34 (d, J = 4.1 Hz, 1 H, PyH-6), 8.24 (d, 3J =
3
8.5 Hz, 1 H, PyH-4), 7.95 (d, J = 8.1 Hz, 2 H, ArH), 7.5–7.3 (m,
3 H, PyH-5 and ArH), 2.43 (s, 3 H, ArCH3) ppm; 13C{1H}NMR
(CDCl3, 75 MHz): δ = 165.96, 145.40, 142.96, 141.39, 131.30,
129.58, 127.51, 127.08, 123.78, 21.53 ppm. EI-MS: m/z = 212 (M+).
C13H12N2O·0·4H2O (219.46): C 71.15, H 5.88, N 12.77; found: C
71.57, H 5.46, N 12.30.
[8]
[9]
L. C. Yuan, T. C. Bruice, J. Am. Chem. Soc. 1986, 108, 1643.
a) B. de Poorter, B. Meunier, J. Chem. Soc., Perkin Trans. 2
1985, 1735; b) A. Cagnina, S. Campestrini, F. di Furia, P. Ghi-
otto, J. Mol. Catal. A 1998, 130, 221.
MnTMPP: Compound H2TMP (100 mg, 0.136 mmol) was dis-
solved in DMF (10 mL) and the system was heated to reflux.
Mn(OAc)2·4H2O (168 mg, 0.686 mmol) was added and the reac-
tion was monitored by UV/Vis spectroscopy. After 30 min, the re-
action was complete and the solvent was evaporated. CH2Cl2
(10 mL) and saturated aqueous NaCl solution (10 mL) were added
and the resulting mixture was stirred at room temperature over-
night. The organic layer was separated, dried with MgSO4, evapo-
rated to dryness and subjected to column chromatography (silica,
2–10% MeOH in CHCl3). The product was dissolved in the small-
est possible amount of CHCl3 and precipitated from n-hexane.
Yield: 103 mg (92%) of MnTMPP as a dark green solid. UV/Vis
(CH2Cl2): λ/nm [log(ε/–1 cm–1)] = 348 (4.60), 374 (4.69), 400 (4.58),
449 (4.07), 479 (5.01), 527 (3.74), 582 (3.97), 618 (3.92). FAB-MS:
m/z = 787 [M – Cl]+.
[10]
[11]
[12]
[13]
W. H. Wong, D. Osovic, T. C. Bruice, J. Am. Chem. Soc. 1987,
109, 3428.
D. Mansuy, P. Battioni, J. P. Renaud, J. Chem. Soc., Chem.
Commun. 1984, 1255.
P. Battioni, J. P. Renaud, J. F. Bartoli, M. Reina-Artiles, M.
Fort, D. Mansuy, J. Am. Chem. Soc. 1988, 110, 8462.
a) J. A. S. J. Razenberg, R. J. M. Nolte, W. Drenth, Recl. Trav.
Chim. Pays-Bas 1986, 105, 103; b) R. J. M. Nolte, J. A. S. J. Ra-
zenberg, R. J. W. Schuurman, J. Am. Chem. Soc. 1986, 108,
2751; c) A. W. van der Made, R. J. M. Nolte, W. Drenth, Recl.
Trav. Chim. Pays-Bas 1990, 109, 537.
[14]
[15]
[16]
B. Meunier, E. Guilmet, M.-E. de Carvalho, R. Poilblanc, J.
Am. Chem. Soc. 1984, 106, 6668.
A. W. van der Made, R. J. M. Nolte, J. Mol. Catal. 1984, 26,
333.
J. A. A. W. Elemans, M. B. Claase, P. P. M. Aarts, A. E.
Rowan, A. P. H. J. Schenning, R. J. M. Nolte, J. Org. Chem.
1999, 64, 7009.
Catalysis Experiments: Catalysis experiments were carried out in a
Schlenk tube (2ϫ2ϫ12 cm) under nitrogen. Substrate (0.6 ),
phase-transfer catalyst (tetra-n-butylammonium chloride, 5.0 m),
the axial ligand (approach A: 2.5 m; approach B: 1.25 ) and the
internal standard (1,3,5-tri-tert-butylbenzene, 0.6 m) were added
[17] A preliminary communication of this work has appeared:
J. A. A. W. Elemans, E. J. A. Bijsterveld, A. E. Rowan, R. J. M.
Nolte, Chem. Commun. 2000, 2443.
to
a solution of distilled CH2Cl2 containing Mn-porphyrin
[18]
CHARMm Version 22.0, Revision 92.0911. Resident and Fel-
lows of Harvard College, 1984, 1992, with the use of template
charges.
(2.5 m, 650 µL), together with an aqueous solution of NaOCl
(0.6  , 2.0 mL). The resulting two-phase system was magnetically
stirred at 1100 rpm. At intervals, the stirrer and timer were stopped,
and after phase separation an aliquot (1 µL) was withdrawn from
the organic phase and injected into CH2Cl2 (0.5 mL). The resulting
CH2Cl2 solution was analysed by GC and compared with authentic
samples. After 3 h, the organic phase was evaporated to dryness
[19]
[20]
M. Monteneau, J. Mispelter, B. Loock, E. Bisagni, J. Chem.
Soc., Perkin Trans. 1 1983, 189.
In UV/Vis titrations of Mn1 with pyridine- and imidazole-con-
taining ligands the Soret band at 480 nm displayed only a mini-
mal increase in intensity after the addition of Ͼ10 equiv. of
ligand, but no shift in wavelength. The obtained titration
curves could not be fitted to give reliable association constants
and for this reason host Zn1 was used as a reference. Generally,
the Ka values of N-ligands with manganese porphyrins are
somewhat lower than with zinc porphyrins.
1
and analysed by H NMR spectroscopy.
Acknowledgments
[21]
[22]
[23]
a) H. Adams, F. J. Carver, C. A. Hunter, N. J. Osborne, Chem.
Commun. 1996, 2529; b) T. Nakanaga, F. Ito, J. Phys. Chem.
A 1999, 103, 5440.
In addition to an NH···π interaction, there is also a possible
CH···π interaction between the imidazole 4-proton and the
other side-wall of the host; see part c of Figure 2.
Unfortunately the Im proton signals in the spectrum of the
complex were very broad, which prohibited their use as a diag-
nostic tool for the determination of the binding geometry of
the ligand.
The Dutch National Research School for Combination Catalysis
(NRSC-C) and the Council for the Chemical Sciences of the
Netherlands Organization for Scientific Research (CW-NWO) are
acknowledged for financial support to J. A. A. W. E. (Veni grant),
A. E. R. (Vidi grant) and R. J. M. N. (Top grant).
[1] P. R. Ortiz de Montellano (Ed.), Cytochrome P450: Structure,
Mechanism and Biochemistry, 2nd ed., Plenum Press, New
York, 1995.
756
www.eurjoc.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2007, 751–757

Manganese Porphyrin Hosts as Epoxidation Catalysts
FULL PAPER
[24] It must be noted that the catalytic reactions were performed in
a two-phase solvent mixture of dichloromethane in water,
which would probably lower the Ka values somewhat.
[25] The presence of the conjugated base of OHPy in the aqueous
layer supported this conclusion.
[26] S. Campestrini, F. di Furia, P. Ghiotto, F. Novello, C. Travag-
lini, J. Mol. Catal. A J. Mol. Catal. 1996, 105, 17.
measure a reliable rate of conversion due to phase separation
between the solvent and the epoxidation products.
[28] P. Thordarson, E. J. A. Bijsterveld, A. E. Rowan, R. J. M.
Nolte, Nature 2003, 424, 915.
Received: July 26, 2006
Published Online: December 5, 2006
[27] Five portions of substrate were oxidized without any apparent
decomposition of the catalyst. After that it became difficult to
Eur. J. Org. Chem. 2007, 751–757
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
757