Effects of pyridine and its derivatives on the equilibria and kinetics pertaining to epoxidation reactions catalyzed by methyltrioxorhenium

Article abstract of DOI:10.1021/ja9813414

The coordination of substituted pyridines to MTO (methyltrioxorhenium) is governed by both electronic and steric effects. For example, the binding constant of pyridine to MTO is 200 L mol^-1, whereas that of the better donor 4-picoline is 730 L mol^-1 and that of the sterically encumbered 2,6- di-tert-butyl-4-methylpyridine is <1 L mol^-1. A Hammett reaction constant ρ = -2.6, derived from meta- and para-substituted pyridine, applies to this equilibrium. Pyridine stabilizes the MTO/H₂O₂ system and accelerates the epoxidation of α-methylstyrene. The steady-state concentration of MTO is decreased during the catalytic epoxidation reaction by coordinating a pyridine derivative, thus stabilizing the MTO/H₂O₂ system against irreversible decomposition. Pyridine as a Lewis base accelerates the generation of the peroxorhehium catalysts, whereas coordination of pyridine to the diperoxorhenium complex appears responsible for the acceleration of epoxidation. Ultimately, however, it is the Bronsted basicity of pyridine that lowers the activity of hydronium ion, reducing the rate of epoxide ring opening.

Full text of DOI:10.1021/ja9813414

J. Am. Chem. Soc. 1998, 120, 11335-11341
11335
Effects of Pyridine and Its Derivatives on the Equilibria and Kinetics
Pertaining to Epoxidation Reactions Catalyzed by
Methyltrioxorhenium
Wei-Dong Wang and James H. Espenson*
Contribution from the Ames Laboratory and Department of Chemistry, Iowa State UniVersity,
Ames, Iowa 50011
ReceiVed April 20, 1998
Abstract: The coordination of substituted pyridines to MTO (methyltrioxorhenium) is governed by both
-
1
electronic and steric effects. For example, the binding constant of pyridine to MTO is 200 L mol , whereas
-
1
that of the better donor 4-picoline is 730 L mol and that of the sterically encumbered 2,6-di-tert-butyl-4-
-1
methylpyridine is <1 L mol . A Hammett reaction constant F ) -2.6, derived from meta- and para-substituted
pyridines, applies to this equilibrium. Pyridine stabilizes the MTO/H2O2 system and accelerates the epoxidation
of R-methylstyrene. The steady-state concentration of MTO is decreased during the catalytic epoxidation
reaction by coordinating a pyridine derivative, thus stabilizing the MTO/H2O2 system against irreversible
decomposition. Pyridine as a Lewis base accelerates the generation of the peroxorhenium catalysts, whereas
coordination of pyridine to the diperoxorhenium complex appears responsible for the acceleration of epoxidation.
Ultimately, however, it is the Brønsted basicity of pyridine that lowers the activity of hydronium ion, reducing
the rate of epoxide ring opening.
1
Introduction
a H NMR spectrometer. Generally a CD
3
NO
2
solution containing ∼10
mM MTO was titrated with a concentrated (∼10 M) solution of a
pyridine derivative. No extra H O was added in these measurements.
The observed chemical shift for the CH protons of rhenium complexes
a concentration-weighted average) was recorded after each addition
of the pyridine. For solid pyridine derivatives, such as pyridine-N-
oxide and 3-cyanopyridine, a stock solution (>3 M) was prepared in
CD
0%[Re]
the solution varied within 2-5% and thus [Re]
during the titration proceses, particularly for the pyridines with large
binding constants. The concentration of uncomplexed ligand, as given
This study was prompted by the recent discovery that
pyridines greatly enhance the selectivity, and mildly enhance
the rate, of epoxide formation from the reaction of alkenes with
2
3
(
1
-3
hydrogen peroxide, catalyzed by MTO (methyltrioxorhenium).
There are several distinct issues, the ones addressed here being
the following: (1) the kinetics and equilibrium for the two stages
of complexation of H2O2 and MTO; (2) the complexation of
pyridines to MTO and other matters relating to the Lewis acidity
of MTO; (3) the ternary system, peroxide/pyridine/MTO; and
3
NO
2
. The titration was generally carried out until [MTO‚L] >
. Since concentrated pyridine was used, the total volume of
changed only slightly
9
T
T
(4) the kinetics of epoxidation reactions under various condi-
in eq 5, was calculated by difference, assuming that [Re]
T
is constant.
tions, inevitably accompanied by the cooxidation of the pyridine.
The binding constants for BF and B(OMe) with pyridine were
3
3
1
determined in the presence of MTO by the use of the H NMR
technique. The weighted-average chemical shift of the rhenium species
was observed directly, and the concentration of free pyridine was
calculated from the known values for KPy, δMTO, and δMTO‚Py. The
Experimental Section
Materials and Techniques. Methyltrioxorhenium, CH
drich), H (30%, Fisher, standardized by iodometric titration),
pyridine and its derivatives (Aldrich), R-methylstyrene (Aldrich), boron
trifluoride dimethyl etherate ((CH O‚BF , Aldrich), trimethyl borate
B(OCH , Aldrich), CD NO (Aldrich), and CD CN (CIL) were used
as received. High-purity H O was obtained by passing laboratory
3 3
ReO (Al-
2 2
O
concentration of BX
3
‚Py was obtained from the calculated concentration
was taken as the difference
‚Py]. The equilibrium constants could then be calculated.
Kinetic Studies. The epoxidation of R-methylstyrene by the MTO/
2 2
O system in the presence of pyridine derivatives was followed by
)
3 2
3
of MTO‚Py. The concentration of BX
3
(
)
3 3
3
2
3
[
BX
]
3 0
- [BX
3
2
distilled water through a Millipore-Q water purification system. The
concentrations of pyridine, its derivatives, boron trifluoride dimethyl
etherate, and trimethyl borate were measured volumerically.
H
1
H NMR. A solution with desired concentrations of R-methylstyrene
CN
right before the kinetic data were collected.
The ortho-protons of pyridine (8.58 ppm) and pyridine-N-oxide (8.26
ppm), and the CH protons of R-methylstyrene (2.16 ppm) and its
and pyridine derivatives was added into a preequilibrated CD
solution of MTO and H
3
1
A Bruker DRX-400 spectrometer was used to record the H NMR
2 2
O
1
spectra at 400.13 MHz. The H chemical shifts were measured relative
1
to residual H resonance in the deuterated solvents, CD
3
NO
2
(δ ) 4.33
3
ppm) and CD
3
CN (δ ) 1.94 ppm). UV-vis measurements were carried
epoxide (1.65 ppm) were integrated. The intensities were then
converted to concentrations based on the known initial concentration
and mass conservation law. The concentration-time profile for the
first 200 s thus obtained is linear, and the slope is the initial rate (mol
out with a Shimadzu UV-3101PC spectrophotometer.
Determination of Equilibrium Constants. The binding constants
between MTO and substituted pyridines were measured by the use of
-
1
-1
(
1) Rudolph, J.; Reddy, K. L.; Chiang, J. P.; Sharpless, K. B. J. Am.
Chem. Soc. 1997, 119, 6189.
2) Coperet, C.; Adolfsson, H.; Sharpless, K. B. J. Chem. Soc., Chem.
Commun. 1997, 1565.
3) Herrmann, W. A.; Ding, H.; Kratzer, R. M.; K u¨ hn, F. E.; Haider, J.
J.; Fischer, R. W. J. Organomet. Chem. 1997, 549, 319.
L
s
).
Kinetic studies in the absence of pyridine derivatives were carried
out spectrophotometrically under these conditions: 0.070 mM R-
methylstyrene, 0.50 M H , and 3.4-28 mM Re . The absorbance
loss at 240 nm, corresponding to the disapperance of R-methylstyrene,
(
O
2 2
T
(
1
0.1021/ja9813414 CCC: $15.00 © 1998 American Chemical Society
Published on Web 10/20/1998

11336 J. Am. Chem. Soc., Vol. 120, No. 44, 1998
Wang and Espenson
was followed. Quartz cuvettes with an optical path length of 0.010
cm were used to accommodate the use of a large excess of diper-
Table 1. Comparisons of the Thermodynamic and Kinetic Data of
the MTO/H
and Aqueous Media
2 2 2
O System (Equation 1) in Organic ([H O] ) 0.6 M)
ψ
oxorhenium species. The slope of the pseudo-first-order plot (k vs
-
1
[
s
B]) yields the second-order rate constant of 0.086 ( 0.002 L mol
k/(10-2 L mol-1 s-1
dielectric
-
1
1
. The control experiment was also carried out by the use of a H
solvent
constanta
K
1
/M-1
K
2
/M-1
k
1
k
2
NMR spectrometer under the following conditions: [R-methylstyrene]
1.3 mM, [Re ] ) 26 mM, and [H ] ) 0.83 M. The second-order
rate constant of 0.060 ( 0.007 L mol
CD
CH
CH
3
3
3
NO
CN
OH
2
38.6
37.5
32.6
78.5
NA
NA
1300b
7.4
b
1.4
1.0
b
c
b
)
T
2 2
O
210c
700c
10
c
-
1
-1
4
s
was obtained.
d
d
b
e
f
261e
814e
68
8000
3250
1.8
Kinetic data for the formation of A and B in methanol were collected
by following the absorbance change at 320 (the maximum for A) or
e
H
2
O
7.7
145
136
520
f
f
105f
1
:1 H
2
O/CH
3
CN
13
3
60 nm (the maximum for B) at 25 °C and [H O] ) 0.60 M. The
2
5
a
biphasic kinetic traces were evaluated as described previously.
Angelici, R. J. Synthesis and Technique in Inorganic Chemistry;
2
nd ed.; University Science Books: Mill Valley, CA, 1986; p 215.
b
This work. c Reference 5. d Reference 7. e Reference 9. Reference 8.
f
Results and Interpretation
MTO-H2O2 Interactions in the Absence of Pyridine. Two
reactions occur, representing the stepwise complexation reac-
tions, giving rise to A, the monoperoxo rhenium complex, and
to B, the diperoxo complex. The latter species has been well-
2
characterized as (CH3)Re(O)(η -O2)2(H2O), even to the point
6
of a single-crystal diffraction study. A, on the other hand, has
1
been characterized only by its H NMR and optical spectra,
being more elusive owing to the pattern of equilibrium constants,
K2 > K1.7 To date, these reactions have been studied in
5
8
methanol, acetonitrile, 1:1 aqueous acetonitrile, and aqueous
solution;9 aqueous and semiaqueous solutions were necessarily
,10
11
acidified to minimize MTO decomposition. The rate constants
are defined in eq 1, and the equilibrium constants follow from
them: K1 ) k1/k-1, K2 ) k2/k-2. The concentration of water is
incorporated neither in the values of the reverse rate constants
nor in the definitions of the equilibrium constants, because water
coordination to A, although unlikely,11 cannot be ruled out
absolutely.
Figure 1. Kinetic curves depicting the buildup of A and B and the
loss of MTO. Also shown is the buildup of methanol, which occurs
only during the stage of MTO loss. Conditions: 13.9 mM MTO; 144
mM H
order kinetics, k ) 9.57 × 10
kinetics, fixing the one k at 9.57 × 10
2.38 × 10 . The curve for B was also fit to biphasic kinetics,
and the second rate constant was refined to k ) 2.42 × 10
2 2 2
O ; 0.6 M H O at 296 K. The decrease in MTO followed first-
-
3 -1
s
. The curve for A was fit to biphasic
-
3
-1
s
and floating the other, k
-
3
-1
)
s
-
3
-1
s
.
ity. Although we have no data to defend this premise, it is
tempting to suggest that the gain of multiple bond character in
a single RedO unit poses one thermodynamic barrier and that
it occurs to the greatest extent in the second stage.
The equilibrium constants are always considerably larger in
organic solvents than in aqueous or semiaqueous ones. That
trend might arise from the release of water in the coordination
reactions, but it is evident in K2 values as well, where there is
no release of water if the extent of hydration of A is correctly
described in eq 1. The polarity of the medium provides the
more likely explanation. The most notable effect in these
equilibrium constants, however, is the finding that K1 < K2 in
all of solvents examined. This is an example of a cooperativity
effect, and it must signal that the structural rearrangement
accompanying conversion of MTO to A provides a structure
than can coordinate peroxide to yield B with a greater spontane-
We find we must retract the values of the rate constants for
peroxorhenium formation (k1, k2) and dissociation (k-1, k-2)
7
reported for these reactions in methanol. Only later did we
2
3
realize these values were unrealistically large (×10 -10 ) as
compared to other solvents. An examination of the original
notebooks brought out the fact that these kinetic determinations
were carried out in solutions containing N,N-dimethylaniline,
the substrate for the subsequent catalysis. As we show here
for pyridine, peroxorhenium formation is accelerated by bases;
this effect is the reason that the values must now be revised.
The new values in methanol (Table 1) are in line with rate
constants in other organic solvents.
(4) The quality of the NMR kinetic data with 1.3 mM R-methylstyrene
was not as good as those with 0.32 M R-methylstyrene owing to the error
associated with the integration of small signals. As pointed out by one
reviewer, the initial rate method using the NMR spectrometer could cause
a problem, particularly when the signals of reactants and products are not
fully resolved. Reasonable kinetic data could be obtained if one integrates
the signals of products rather than those of the reactants and uses an internal
standard with an intensity similar to that of the products.
Since the effect of pyridine on the epoxidation reactions was
1
examined in nitromethane, we evaluated the equilibrium and
1
kinetic constants in that solvent by H NMR. At this particular
set of concentrations, 13.9 mM MTO and 144 mM H2O2, the
resonances of all three rhenium species were observable: MTO,
δ 2.79 ppm; A, δ 2.82 ppm; and B, δ 2.87 ppm. These chemical
shifts move slightly upfield with increasing [H2O2], during which
(
5) Wang, W.-D.; Espenson, J. H. Inorg. Chem. 1997, 36, 5069-5075.
(6) Herrmann, W. A.; Fischer, R. W.; Scherer, W.; Rauch, M. U. Angew.
Chem., Int. Ed. Engl. 1993, 32, 1157.
7) Zhu, Z.; Espenson, J. H. J. Org. Chem. 1995, 60, 1326-1332.
[H2O] necessarily increases as well. The kinetic profiles of these
species are displayed in Figure 1.
2
72.
(
9) Yamazaki, S.; Espenson, J. H.; Huston, P. Inorg. Chem. 1993, 32,
Most of the methanol was formed in the first 500 s, when
MTO was present in the solution. Although a significant
amount of A remained at that point, practically no more
methanol was produced. This result agrees with the deactivation
4
683.
(10) Hansen, P. J.; Espenson, J. H. Inorg. Chem. 1995, 34, 5839.
11) Abu-Omar, M.; Hansen, P. J.; Espenson, J. H. J. Am. Chem. Soc.
(
1
996, 118, 4966-4974.

Epoxidation Reactions Catalyzed by MTO
J. Am. Chem. Soc., Vol. 120, No. 44, 1998 11337
pathway dominated by a reaction in which MTO, but not A or
B, is the reactant:11
CH ReO + HO f ReO - + MeOH
-
(2)
3
3
2
4
Also, the H+ that results from eq 2 with the net reaction
written for H2O2 will increase the acidity of the solution. The
-
increase in acidity would prevent the reaction of A with OH ,
that being the kinetically indistinguishable alternative to reaction
1
2
.
H NMR studies in CD3NO2 gave rate constants similar to
those obtained in CD3CN. A summary of the rate and
equilibrium constants is given in Table 1.
Adduct Formation between MTO and Pyridines. Ligands,
Figure 2. Chemical shift varying with the concentration of pyridine.
especially those with oxygen or nitrogen donor atoms, form
Conditions: [MTO]
0
3 2
) 14 mM; 296 K in CD NO . The line through
Lewis acid-base adducts with MTO.1
2-14
Although both sites
the data is the fit according to eq 4.
of bidentate ligands can bind to the rhenium center to form a
distorted octahedral complex, usually only a single monodentate
ligand coordinated to the rhenium atom when a monodentate
ligand was used. What is the effect of ligand binding on
reactivity and selectivity? Selectivity enhancement by the
addition of Lewis bases has been observed previously, but
catalytic reactivity may be reduced.15 The effect of pyridine is
different: an excess of pyridine over MTO not only increases
Table 2. Equilibrium Constants for the Coordination of
Substituted Pyridines and Related Species to MTO in CD
3
2
NO at
2
96 K
L
σm or σp K/(L mol-1) δMTO‚L/ppm
0.56 7.21 ( 0.14 2.04 ( 0.007
3-cyanopyridine
3-bromopyridine
-chloropyridine
4-nicotinaldehyde
-phenylpyridine
pyridine
3-methylpyridine (3-picoline)
4-methylpyridine (4-picoline)
0.39 23.3 ( 0.2
0.37 21.8 ( 0.3
0.22 36.8 ( 0.9
0.06 156 ( 4
1.96 ( 0.002
1.95 ( 0.003
1.96 ( 0.006
1.91 ( 0.005
1.86 ( 0.005
1.82 ( 0.007
1.84
3
1
the selectivity toward epoxide but also accelerates its formation.
3
A concurrent effect of pyridine arises from its Brønsted
basicity. As cited earlier, MTO is much less stable the more
neutral or basic the solution becomes. Possible mechanisms
for the base-mediated deactivation of the MTO/H2O2 system
have been proposed on the basis of kinetic studies in aqueous
0
200 ( 6
-0.07 254 ( 11
-0.17 732 ( 20
2
2
-chloropyridine
,6-di-tert-butyl-4-methylpyridine
<1
<1
pyridine-N-oxide
,6-dimethylpyridine (2,6-lutidine)
2,6-dimethylpyrazine
210 ( 4
<1
1.87 ( 0.003
1.95 ( 0.03
11
systems. Thus complex formation between pyridine and MTO
needs to be explored.
2
20 ( 1.5
The interaction between MTO and several pyridines was
1
studied by H NMR techniques. Only one sharp singlet from
of a solution of MTO in pyridine-d5 is 1.87 ppm, consistent
with the fitting result and indicating that only one pyridine is
likely coordinated to MTO.
CH3ReO3 was observed with pyridine/rhenium e 1; this
resonance remained sharp and shifted upfield with more
pyridine. A fully averaged signal was observed even at -40
Derivatives of pyridine, 2-Cl, 3-Cl, 3-CN, 3-Me, 4-Me, 2,6-
°
C in CD2Cl2, which implies that the suggested equilibrium
t
di-Bu -4-Me, also interact with MTO to a greater or lesser ex-
K
tent. These data were also treated by the same procedure
used for pyridine. The resulting parameters are summarized in
Table 2.
MTO + L y
\
z MTO‚L
(3)
remains rapid relative to the NMR time scale even at the low
temperature. Because of that, the observed chemical shift is a
concentration-weighted average of the two
As the data show, the equilibrium constants are sensitive to
both steric and electronic effects. Those with steric hindrance,
t
such as 2-Cl and 2,6-di-Bu -4-Me, interact only slightly with
δMTO +K[L]δMTO‚L
MTO. On the other hand, electron-rich pyridines, such as 3-Me
and 4-Me, are strongly favored over electron-withdrawing ones,
such as 3-Cl and 3-CN. An attempted linear free energy
relationship (LFER) correlation according to the Hammett
equation is displayed in Figure 3. The slope of the line of log
δobs
)
(4)
1
+ K[L]
where [L] is the free (uncomplexed) concentration of pyridine
or other ligand, and is related, with ê ) 1+ K[Re]T - K[L]T,
to the total (starting) concentrations by
(
K/K0) vs σ is -2.6 ( 0.1, which is the reaction constant for
this equilibrium.
2
ê + 4K[L ] - ê
x
T
A related correlation is an attempt to correlate the chemical
shift of the adduct, δMTO•L, with Hammett’s σ. Figure 3 also
shows a decent but not excellent correlation with the slope of
the line being 0.26 ( 0.04. The quality of the correlation and
the small value of the slope no doubt reflect the small spread
in chemical shifts along the series of substituents. The
significant point, however, is this: the trends in K and in δ
vary sensibly with the substituent constants and in the direction
consistent with the interaction depicted in eq 3.
L )
(5)
2K
The data fit well to eq 4, affording these values of the
-
1
parameters: K ) 200 ( 6 L mol and δMTO‚L ) 1.86 ppm in
CD3NO2 at 296 K. The correlation between its chemical shift
and [Py]T in CD3NO2 is shown in Figure 2. The chemical shift
(
12) Herrmann, W. A.; K u¨ hn, F. E. Acc. Chem. Res. 1997, 30, 169-
80.
13) Herrmann, W. A.; Kuchler, J. G.; Weichselbaumer, G.; Herdtweck,
E.; Kiprof, P. J Organomet. Chem. 1989, 372, 351.
14) Herrmann, W. A.; Weichselbaumer, G.; Heerdtweck, E. J. Orga-
nomet. Chem. 1989, 372, 371.
15) Herrmann, W. A.; Fischer, R. W.; Rauch, M. U.; Scherer, W. J.
Mol. Catal. 1994, 86, 243.
1
(
Other Ligands. The data show that pyridine-N-oxide, PyO,
binds with MTO as strongly as pyridine does. The Lewis acid-
base interactions of these two with MTO are clearly different
(
(
+
than those with H3O , where there is a four-order-of-magnitude

11338 J. Am. Chem. Soc., Vol. 120, No. 44, 1998
Wang and Espenson
This approach is still lacking, however, in that a reaction’s
sensitivity to changes in substituent is not synonymous with its
thermodynamic driving force. For that reason, we directly
studied the binding constants of pyridine with BF3 and B(OMe)3
1
with H NMR. The same treatment presented in eq 4 gives
-
1
4
these values for Py of K/L mol : BF3 3.3 × 10 ; B(OMe)3,
2
1
.4 × 10; MTO, 2.00 × 10 ; OsO4, 3.4 × 10. Although the
binding constant for MTO was determined in nitromethane, it
can be compared with other K’s in acetonitrile, because other
equilibrium and kinetic parameters are similar, as given in Table
2
1
1
.
Clearly, MTO is a stronger Lewis acid toward pyridine
derivatives than OsO4.
Figure 3. Correlation of the equilibrium constants with the Hammett
substituent parameter σ. The reaction constant is -2.6 ( 0.1. The
variation of the chemical shift of the MTO‚L complex with the Hammett
substituent parameter σ is also shown.
Effect of H2O on the Formation of A and B. With 14 mM
Py′ (Py′ ) 2,6-di-tert-butyl-4-methylpyridine) and 1 mM MTO,
the observed rate constant for the formation of B at 87 mM
H2O2 varied inversely with [H2O] in the range 0.3-1.1 M H2O.
This trend is opposite to that in the absence of bases, where
added water accelerates the formation of B and more so that of
A. The effect of water with Py′ is not obvious. It is reminiscent
of the retarding effect of water on the formation of A and B in
methanol and acetonitrile, found during studies of the oxidation
difference (pKb ) 8.75 and 13.2 for Py and PyO, respectively).16
The oxophilic character of Re(VII) and less steric interference
of PyO possibly account for the similar binding constants with
1
7
MTO. The O NMR data for the MTO adducts with quinu-
clidine and quinuclidine-N-oxide also suggest a comparable
7
of anilines with MTO/H2O2.
equilibrium constant between quinuclidine and quinuclidine-
N-oxide with MTO.17
MTO/H2O2 System in the Presence of Pyridine. If one
regards the binding of pyridine and of hydrogen peroxide as
separate and sometimes competing reactions, the latter is
distinctly more favorable, as the equilibrium constants in Tables
The association between MTO and 2,6-dimethylpyrazine is
much stronger than with its pyridine analogue, Table 2. It is
worth noting that the H-D exchange between CH3ReO3 and
CD3NO2 becomes faster as a stronger Lewis base is used.
Significant incorporation of deuterium into MTO occurred in
an hour when 2,6-dimethylpyridine, the strongest base so far
used, was employed.
Lewis Acidity of MTO. MTO is believed to be a hard Lewis
acid on the basis of its strong interactions with N- and O-donor
ligands and its weak interaction with S-donor ligands. The
negative reaction constant for the pyridine equilibria, -2.6,
indicates that a positive charge develops on the pyridine nitrogen
in the adduct as compared with the free molecule. As expected,
the Re(VII) center acts as an electron acceptor, attracting
electrons upon coordination.
1
and 2 clearly show. Indeed, the yellow color characteristic
of B rapidly developed when H2O2 was added into the colorless
solution of MTO‚Py in nitromethane or acetonitrile. Interest-
ingly, and important to the catalytic cycle, that color developed
more rapidly from MTO‚Py than from MTO alone. This
phenomenon was more pronounced at low [H2O2].
The binding constants for MTO are nearly the same for
-
1
pyridine (K ) 200 L mol ) and pyridine-N-oxide (K ) 210 L
-
1
7,22-24
25
mol ). Pyridine, like amines
and hydroxylamines, is
oxidized by the MTO/H2O2 system. Thus when pyridine is
present, a considerable concentration of PyO develops. In such
cases, the interaction of MTO with both Lewis bases must be
considered. Though the 4-tert-butylpyridine-N-oxide adduct of
To put the Lewis acidity of MTO on an external scale, it is
necessary to make comparisons with another well-known acid;
BF3 was chosen for this purpose. Against that standard,
complexation with pyridines, eq 6, was also chosen. Although
1
7
1
B has been isolated, the H spectrum of B at high [H2O2] was
unaffected by the addition of a large excess of PyO. Either the
interaction of PyO with B is negligible or the chemical shifts
of the PyO and H2O adducts of B are the same. We have
obtained evidence that the interaction of PyO with A forms a
KB
BF + L y
\
z BF ‚L
(6)
1
3
3
complex of appreciable stability. The H spectrum of A at lower
H2O2] shifts upfield appreciably. Like the reaction between
[
values of KB have not been reported, the values of ∆H are
MTO and PyO, the equilibrium between A and PyO is
established rapidly on the NMR time scale, such that only one
MeRe resonance could be observed at room temperature.
Addition of hydrogen peroxide to a solution of MTO‚Py and
1
8
known for three substituted pyridines in nitrobenzene. With
the reasonable assumption that the ∆S values are the same, the
quantity (K/K0) could be obtained. The Hammett correlation
gave F ) -2.31 ( 0.25. The same procedure for BMe3, with
the use of the calorimetric values of ∆H,19 also gave F ) -2.3.
The same values of F appear to arise from the extensive and
offsetting π bonding in trifluoroborane, which reduces its acidity
significantly, the π component being lost in adduct formation.
A value of F ) -1.8 has been reported for the coordination of
pyridine derivatives to OsO4 in acetonitrile.20 Toward the
proton, pyridines have a significantly more negative reaction
constant, F ) -5.2 ( 0.3.
1
free pyridine gave rise to a H spectrum with two observable
CH3-Re resonances. One arises from B at 2.87 ppm; the other
appears at 2.11 ppm. Reverse additionsPy added to a solution
of B and excess H2O2sgave the same spectrum. The resonance
(
20) Nelson, D. W.; Gypser, A.; Ho, P. T.; Kolb, H. C.; Kondo, T.;
Kwong, H.-L.; McGrath, D. V.; Rubin, A. E.; Norrby, P.-O.; Gable, K. P.;
Sharpless, K. B. J. Am. Chem. Soc. 1997, 119, 1840.
(21) The binding constant with MTO was also measured in acetonitrile,
-
1
K ) 173 ( 6 L mol , to facilitate comparison with OsO4. The amount of
H2O (up to 3.1 M) does not affect the binding constant between MTO and
pyridine, and the binding constant of H2O with MTO is small in acetonitrile,
(
16) Shaw, E. N. In Heterocyclic CompoundsPyridine and DeriVatiVes;
Klingsberg, E., Ed.; Interscience Publishers: New York, 1961; Part II, p
17.
17) Herrmann, W. A.; Correia, J. D. G.; Rauch, M. U.; Artus, G. R. J.;
K u¨ hn, F. E. J. Mol. Catal. 1997, 118, 33.
-
1
0.4 L mol
.
1
(22) Yamazaki, S. Bull. Chem. Soc. Jpn. 1997, 70, 877.
(23) Murray, R. W.; Iyanar, K.; Chen, J.; Wearing, J. T. Tetrahedron
Lett. 1996, 37, 805-808.
(24) Goti, A.; Nannelli, L. Tetrahedron Lett. 1996, 37, 6025-6028.
(25) Zauche, T. R.; Espenson, J. H. Inorg. Chem. 1997, 22, 55257.
(
(
(
18) Brown, H. C.; Horowitz, R. H. J. Am. Chem. Soc. 1955, 77, 1733.
19) Brown, H. C. J. Chem. Soc. 1956, 1248.

Epoxidation Reactions Catalyzed by MTO
J. Am. Chem. Soc., Vol. 120, No. 44, 1998 11339
Scheme 1
Scheme 3
Scheme 2
at 2.11 ppm is tentatively assigned to A‚L (L ) Py, PyO), since
this signal disappeared and MTO‚L appeared at 1.91 ppm when
the limiting H2O2 had been consumed. Thus it seems that B
becomes A by transferring an O-atom to Py:
pKb’s,26 but Py′ is noncoordinating and nonoxidizable, its kinetic
effect was explored. With 1 mM MTO and 87 mM H2O2, the
observed rate constant for the buildup of B increased with 2-14
mM Py′. The yield of B was 40% of [MTO]0, independent of
[Py′]. There was significant decomposition of MTO, however,
B + Py f A + PyNO (or A‚OPy) + H O
(7)
2
-
11
no doubt as a result of its earlier-described reaction with HO2 .
From the description that follows, it is clear that it is not easy
to distinguish the adducts of A with Py and PyO. The
interaction of B with pyridine cannot easily be detected owing
to the oxidation of Py by B, eq 7.
On the basis of these observations, the MTO/H2O2/Py system
can be described as shown in Scheme 1. The dashed arrows
indicate reactions that are plausible but not directly detected.
The alternative routes (top and bottom) to B cannot be
distinguished.
The oxidation of pyridine attracts special attention, since it
may and very likely will happen concurrently with epoxidation.
The oxidation of Py to PyO occurs mainly when B is present,
suggesting that pyridine oxidation from A‚L is much less
effective than that from B.
This factor prevented a meaningful analysis of the kinetic data.
1
The H NMR spectrum of this system showed a significant
amount of methanol, confirming the decomposition of MTO.
This result is different from the system lacking Py′ in these
respects: (1) B was formed from MTO and H2O2 in 95% yield,
since MTO remained intact; (2) addition of unsubstituted Py
also gave a 95% yield of peroxorhenium species. From that
we conclude the coordination of Py to MTO must protect the
catalyst from deactivation by the base-mediated pathway.
A plausible mechanism to account for the kinetic enhance-
ment by bases is presented in Scheme 3. The left side of the
scheme is the same as that proposed for aqueous acidic solution.
The right depicts the basic pathway. Because H2O2 (pKa )
1
1.6) is more acidic than water (pKa ) 15.7, pKw ) 14.0), the
Effects of Pyridines on the Formation of A and B. It is
unlikely that the acceleration of these reactions arises only from
the coordination of Py to MTO. The bulky and thus noncoor-
dinating pyridine Py′ ) 2,6-di-tert-butyl-4-methylpyridine ac-
celerates these reactions as well. Although the MTO binding
constants for Py and PyO are nearly the same, the accelerating
effect for forming B is much larger for Py than for PyO. We
thus conclude that these accelerations arise from the Brønsted
equilibration between 3 and 4 should favor 4. Products seem
likely to arise from 3, however, because 3 leads directly to
-
MeOH, whereas 4 might release MeO . Once 4 forms, then A
will follow by the sequence 4 f A-OH f A-OH2 f A. The
formation of B follows from a closely analogous scheme.
-
Both H2O2 and HO2 produce the peroxorhenium complexes
-
A and B. But HO2 , the better nucleophile, should react more
rapidly with MTO, leading to a pH effect. Still, an explanation
must be produced for the significant decomposition of MTO
when the sterically encumbered Py′ is used.
-
basicity of pyridine: it increases the concentration of HO2 ,
which is much more nucleophilic and thus more reactive than
H2O2. Scheme 2 shows this effect.
According to this, the reaction rate will be
The binding of MTO to Py or PyO stabilizes the system,
suggesting that the reaction between MTO‚L with H2O2 (Scheme
1
) is also involved in the generation of the active form of the
+
d[MTO]
k
1
[H ] + k
1
′K
a
catalysts. Pyridine coordination may alter the rate of methyl
migration; also, the change of the coordination geometry about
rhenium may impede methyl migration.
-
)
[H O ] [MTO]
(8)
2
2 T
+
dt
[
H ] + Ka
Epoxidation Reactions in the Presence of Pyridine. 1H
NMR techniques were used to follow the epoxidation but with
two complications: (1) the catalyst is slowly decomposing, and
On the other hand, the rates of eq 1 are independent of acidity
at pH 1-3 in aqueous solution.9 These formulations are not
-
12
contradictory, however, since Ka is but 2.5 × 10 , such that
(2) pyridine is being oxidized at a comparable rate. To optimize
an acid effect should not be found at pH < 3, even if k1′ were
8
-1 -1 11
the information, a relatively reactive olefin, R-methylstyrene,
was used. We chose to use 0.83 M H2O2 for comparison with
as large as 4 × 10 L mol s .
A study of the effects of pyridine on the kinetics of the
formation of peroxorhenium species must allow for concomitant
oxidation of pyridine. Since Py′ and Py have comparable
(26) Anne, A.; Morioux, J.; Saveant, J.-M. J. Am. Chem. Soc. 1993, 115,
10224.

11340 J. Am. Chem. Soc., Vol. 120, No. 44, 1998
Wang and Espenson
Table 3. Initial Rates of Oxidation of Pyridine and
R-Methylstyrene in the Presence of 0.83 M H
CD CN ([H O] ) 3.3 M)
2
O
2
at 296 K in
3
2
Vi/10- L mol-1 s-1
5
MTO/ [Py]/ R-methylstyrene/
entry
mM
M
M
PyO
epoxide
1
2
3
4
5
6
7
8
9
0.66
2.0
6.6
4.0
4.0
4.0
1.3
1.3
1.3
0.090
0.090
0.090
0.33
0.25
0.16
0.058
0.058
0.058
0.52
0.52
0.52
0.20
0.20
0.20
0.12
0.071
0.020
2.22 ( 0.21 7.63 ( 0.097
7.05 ( 0.18 26.4 ( 0.77
19.9 ( 1.3
56.2 ( 2.1
42.2 ( 1.2
72.8 ( 2.8
19.2 ( 0.61
19.8 ( 0.61
27.1 ( 0.48 18.8 ( 0.53
3.13 ( 0.17 3.74 ( 0.053
3.17 ( 0.29 2.30 ( 0.037
3.06 ( 0.11 0.67 ( 0.008
Figure 5. Apparent rate constant for the epoxidation of 0.32 M
R-methylstyrene, in CD
3
CN at constant [Py] ) 41 mM, [Re]
T
) 3.3
mM, and 3.3 M H O, varying with the hydrogen peroxide concentration,
2
c. The fit takes the form k ) 1.20c/(1 + 3.92c).
trile, with k ) 0.086 L mol- s-1. On the other hand, in 1:1
1
CH3CN/H2O, the rate constant is larger, as expected: 0.47 L
-
1
-1
mol s ; epoxidations of alkenes are always faster in a
2
7
semiaqueous solvent.
The data just presented show that the epoxidation rate is
independent of [Py]; nonetheless, the pyridine-free epoxidation
rate is lower. Since the epoxide produced in the experiment
with pyridine, during the first 3 min of reaction, amounts to
some 5-15 times [B]0, a catalytic cycle is clearly effective.
The difference in k’s might be attributed to a faster reaction of
R-methylstyrene with A than B, as is the case for most styrenes
Figure 4. Plots of the initial rates (V
i
) for the oxidation of pyridine
(
circles) and R-methylstyrene (squares) against the product of two
concentrations, [Re] [X] (X ) [Py] or [R-methylstyrene] ). The
concentration of H was kept constant at 0.83 M ([H O] ) 3.3 M),
while [Py] was varied from 0.058 to 0.33 M and [R-methylstyrene]
from 0.020 to 0.52 M.
(the difference of the reactivities between A and B is not
T
0
0
27
significant). If so, and because the ratio of [A]ss/[B]ss should
increase as [H2O2] decreases, then the catalytic rate should
increase as [H2O2] decreases.
O
2 2
2
0
0
Just the opposite is so, as shown in Figure 5: the “rate
1
constant” for epoxidation at constant [Py] and [Re] rises to a
other work, but this precluded the use of nitromethane solvent,
where the system was not homogeneous. Thus, for these
studies, acetonitrile was the solvent; independent studies
confirmed the accelerating effect of pyridine coordination under
catalytic conditions. Again, the kinetic data were analyzed by
the method of initial rates. The reaction order of each pertinent
species was explored by examining the variation of the reactant
concentrations, as given in Table 3.
T
plateau with increasing [H2O2]. This treatment ignores the
speciation of the binary and ternary complexes of MTO/Py/
H2O2, all rate constants being expressed on the basis of [Re]T.
Evidence for A‚L has been presented; the current findings allow
us to conclude that A‚L is much less reactive in epoxidation
than B. To explore this issue further, the epoxidation reactions
were carried out as before with high and constant [H2O2], but
lower [Py] and varying [R-MeSty] and [Re]T. As before the
rate constants were calculated from the initial rates as vi/([Re]T-
In entries 1-3, the two substrates were held constant while
MTO] was varied; the initial rate of oxidation of each substrate
[
[R-MeSty]). These values rise with increasing [Py], as presented
was proportional to [MTO]0. Further, and to no surprise, the
rate of pyridine oxidation is first-order with respect to [Py]0,
and zero-order with respect to [R-MeSty]0, entries 4-9. These
entries also show that the rate of formation of the epoxide is
first-order with respect to [R-MeSty]0 and zero-order with
respect to [Py]0. To display these kinetic effects graphically,
for each substrate X ()R-MeSty and Py), a plot was constructed
in Figure 6. To account for this, we suggest that the (so far
undetected) species B‚Py is present, and that it epoxidizes the
alkene even faster than B.
Epoxidation Reactions in the Presence of Pyridine Deriva-
tives. The epoxidation of R-methylstyrene in the presence of
4
-methylpyridine, 3-chloropyridine, or pyridine-N-oxide was
X
also carried out in a way similar to the experiments presented
in Figure 6. As expected the oxidation of 4-methylpyridine by
the MTO/H2O2 system is faster than that of pyridine (0.59 vs
of the initial rate for each substrate (Vi ) against the product
[Re]T[X], using the data from Table 3. The excellent straight
lines at fixed hydrogen peroxide ) 0.83 M, are shown in Figure
-
1
-1
0
.42 L mol s ). Unlike the case with 4-methylpyridine, a
4
. The quantitative conclusions of this aspect are
much higher concentration of 3-chloro pyridine was needed to
achieve high selectivity for epoxide. Compared with the
epoxidation reactions in the absence of any ligands, the
formation of epoxide is faster in the presence of 4-methylpy-
ridine or 3-chloropyridine. The maximum rate was achieved
Vi ) k[MTO] [Py] ; k ) 0.42 ( 0.01 L mol s-1
PyO
-1
0
0
Epox
Vi
) k[MTO] [R-MeSty] ;
0 0
k ) 0.22 ( 0.04 L mol s-1
-1
(27) Al-Ajlouni, A.; Espenson, J. H. J. Am. Chem. Soc. 1995, 117, 9243-
As these data show, oxidation of pyridine is favored over
9250. The inhibitor present in R-methylstyrene, p-tert-butylcatechol, did
not affect the kinetics in the studies reported in this reference. As pointed
out by one reviewer, a problem might exist in the present work, but that
seems unlikely considering that the data for R-methylstryene were
independent of concentration (Table 3).
the pyridine-present epoxidation of R-methylstyrene. The rate
constant for the pyridine-present epoxidation of R-methylstyrene
is larger than that for the pyridine-free epoxidation in acetoni-

Epoxidation Reactions Catalyzed by MTO
J. Am. Chem. Soc., Vol. 120, No. 44, 1998 11341
2
,3
epoxide from a unreactive olefin.
As observed in the oxi-
2
5
dation of hydroxylamines, protonation of pyridine by adding
trifluoromethane sulfonic acid slowed the formation of pyridine-
N-oxide. When [Py]0 > 10[H ]0 ([Py]0 ) 0.20 M), as more
+
and more acid was added, the selectivity toward epoxide de-
creased while the reactivity increased slightly. For R-methyl-
styrene, whose epoxide is sensitive to the acid-catalyzed ring
opening reaction, best results were obatined when the ratio [Py]/
[acid] is about 2 for CH3COOH and about 120 for CF3SO3H.
Conclusions. The beneficial effects of pyridine can be traced
to these factors: (1) Selectivity for the epoxidation over
dihydroxylation is greatly enhanced due to the basicity of
1
-3,27
pyridine.
Although amines, stronger bases than pyridine,
Figure 6. Effect of pyridine on the epoxidation rate constant in
would seem to be a better choice in this regard, it has been
shown that the oxidation of amines by the MTO/H2O2 system
is generally much faster than most of the epoxidation reactions.
A bulky pyridine can give a high selectivity of epoxide;
however, the noncoordinating pyridine cannot stabilize the
MTO/HP system. (2) Stabilization of MTO against decomposi-
tion is seen, provided a sufficient pyridine concentration is used.
For a pyridine derivative, both the basicity and its binding
constant to MTO should be considered. Generally, as the
basicity increases, the binding constant increases and the catalyst
lifetime decreases. Our results indicate that the basicity of
pyridine is probably too strong for the current system and a
CD
3
CN at 296 K. Conditions: 0.66-6.6 mM Re
T
; 2.1-360 mM Py;
O.
2
0-560 mM R-methylstyrene; 0.83 M H
2
O
2
; 3.3 M H
2
when the concentration of 4-methylpyridine was greater than
0
0
.05 M; however, the maximum rate was not attained even at
.2 M 3-chloropyridine. No substantial difference was found
between the reactivities of the peroxorhenium species in the
presence of pyridine and 4-methylpyridine.
Although PyO does not accelerate the epoxidation of R-me-
thylstyrene, its presence does speed up the formation of A and
B. Since the donor capabilities of the oxygen atoms in PyO
and H2O are probably similar, the reactivities of their peroxo-
rhenium adducts are possibly alike as well. The acceleration
of the formation of peroxorhenium species by PyO unlikely
results from its Brønsted basicity (pKb ) 13.2). It has been
suggested on the basis of the IR stretching frequencies that the
+
Py/PyH buffer should be used instead of pyridine alone. (3)
Pyridine not only accelerates the generation of the peroxorhe-
nium catalysts but also accelerates the epoxidation reactions.
The nature of the latter effect of pyridine is not clear. It is
2
possible that the Re-O (η -O2) bond is weakened upon
RedO bond of MTO is weakened upon coordination of an O-
or N-containing ligand.17 Since the rate-controlling step for the
coordination of pyridine, and as a result the energy barrier for
3
0
the formation of epoxide is decreased. Acceleration of the
epoxidation reactions will benefit all olefins, and acceleration
of the regeneration of catalysts will speed up the epoxidation
of very reactive olefins.
formation of A is believed to be the formation of CH3ReO2-
OH)(OOH),2
8,29
(
weakening the RedO bond in the presence of
PyO is likely the main cause of the acceleration for the formation
of peroxorhenium species. The selectivity enhancement in the
presence of a large excess of pyridine-N-oxide possibly results
from its replacement of H2O on the H2O-solvated rhenium
species.
For the epoxidation of R-methylstyrene, the use of a mixture
of PyO and a bulky pyridine is better than the use of the bulky
pyridine alone. Also, the preliminary results indicate that the
use 2,6-dimethylpyrazine, which provides a protected basic
function group and a binding site, is better than the use of 2,6-
dimethylpyridine. However, none of these is superior to the
use of pyridine.
Acknowledgment. This research was supported by a grant
from the National Science Foundation (CHE-9007283). Some
experiments were conducted with the use of the facilities of
the Ames Laboratory. We are grateful to Dr. A. Yudin and Dr.
K. B. Sharpless for helpful discussions and to the latter for
making available to us the material in ref 1 prior to publication.
JA9813414
(
29) Espenson, J. H.; Abu-Omar, M. M. AdV. Chem. Ser. 1997, 253,
9
9-134.
(
30) The following results favor the involvement of the peroxorhenium
Better Choice than Pyridine Alone? Oxidation of pyridine
has to be considered, particularly when one tries to prepare an
complex with a coordinated pyridine, Re(η2-O )Py, instead of the ReO-
2
(PyO) complex. The reaction of pyridine-N-oxide with R-methylstyrene in
the presence of MTO did not yield the corresponding epoxide in 24 h. Also,
pyridine-N-oxide does not accelerate the epoxidation of R-methylstyrene
by MTO/H2O2; see ref 3 as well.
(28) Pestovsky, O.; van Eldik, R.; Huston, P.; Espenson, J. H. J. Chem.
Soc., Dalton Trans. 1995, 133.