Do (pentaarylcyclopentadienyl)molybdenum(vi) dioxo species catalyse alkene epoxidations? Insights from kinetics data

Article abstract of DOI:10.1039/b705571d

A (perarylcyclopentadienyl)molybdenum(vi) dioxo complex is shown to catalyse epoxidation of cyclooctene by t-butylhydroperoxide, but decomposition to a much more active non-cyclopentadienyl containing catalyst occurs as the reaction proceeds. The Royal So

Full text of DOI:10.1039/b705571d

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/dalton | Dalton Transactions
Do (pentaarylcyclopentadienyl)molybdenum(VI) dioxo species catalyse alkene
epoxidations? Insights from kinetics data†‡
Michael Pratt, Jason B. Harper and Stephen B. Colbran*
Received 12th April 2007, Accepted 11th May 2007
First published as an Advance Article on the web 22nd May 2007
DOI: 10.1039/b705571d
A
(perarylcyclopentadienyl)molybdenum(VI) dioxo com-
Despite intensive study, the mechanism(s) of the catalyses remains
unclear.1,3–5
plex is shown to catalyse epoxidation of cyclooctene by
t-butylhydroperoxide, but decomposition to a much more
active non-cyclopentadienyl containing catalyst occurs as the
reaction proceeds.
Given the marked dependence of catalytic efficiency on the
cyclopentadienyl substituent, we found it surprising that no
(C5Ar5)MoVIO2X (Ar = aryl group) complex had been screened for
epoxidation catalysis. The purpose of this research was to perform
such a screen and was motivated in part by the fact that some
time ago we demonstrated (C5Ar5)MoO2X complexes may exhibit
exceptional stability.6 For example this stability allowed isolation
Mononuclear organomolybdenum(VI) dioxo species are amongst
the most important olefin epoxidation catalysts, and are the
subject of a very recent, comprehensive review by Ku¨hn et al.1
Prominent among these are (cyclopentadienyl)molybdenum(VI)
dioxo derivatives. The first organometallic molybdenum(VI) dioxo
species, CpMoO2Cl, reported by Cousins and Green in 1964,2
is a prototypical member of this class of complexes. Trost and
Bergman first demonstrated in 1991 that (cyclopentadienyl)-
molybdenum(VI) dioxo complexes do catalyse olefinepoxidations;3
they showed that Cp*MoO2Cl catalyses epoxidation of olefins
that do not include a conjugated electron-withdrawing group
such as cyclooctene, geraniol and 1,2,4,5-tetramethylcyclohexa-
1,4-diene, and (stereoselectively) E/Z-1,2-diphenylpropene by
organic hydroperoxides such as t-butylhydroperoxide (TBHP),
cumene hydroperoxide and n-alkylhydroperoxides. Reaction with
either H2O2 or Ph3COOH, however, lead to formation of the
p-peroxide Cp*Mo(O2)OCl, which was fully characterised and
shown to be a totally inactive, ‘dead-end’ species.3 Roesky and co-
workers subsequently confirmed the inactivity of Cp*Mo(O2)OCl
and described its X-ray crystal structure.4
Recently, studies of olefin epoxidations by alkylperoxides have
been extended by Herrmann, Ku¨hn, Poli, Roma˜o and their co-
workers to (C5R5)Mo(O)2X (R = H, alkyl, benzyl, or SiR3; X =
Cl, or alkyl) catalysts.5 These reports reveal a strong dependence
of the catalytic efficiency on the cyclopentadienyl substituents
(R). For instance, under catalysis conditions, Green’s original
complex, CpMo(O)2Cl, decomposed completely with no epoxida-
tion observed whereas (C5Bz5)Mo(O)2Cl (Bz = benzyl) performed
best. This co◦mplex catalysed the epoxidation of cyclooctene by
TBHP at 55 C to 100% conversion after 4 h and at an optimal
turnover frequency of 20 000 h−1 for 0.01 mol% catalyst loading.5c
At the latter low catalyst loading, however, catalyst deactivation
leading to a decreasing conversion was observed after 1 h.
5
of the new alkoxide derivatives Cp‡Mo(O)2(OR) (Cp‡ = g -
(2,5-dimethoxyphenyl)tetraphenyl cyclopentadienyl anion).6 The
results reported herein provide a cautionary warning about the
role of (C5R5)MoVI(O)2X species as epoxidation catalysts.
To facilitate comparison with previously reported catalyses we
concentrated on, and describe here, the epoxidation of cyclooctene
by TBHP in CDCl3 using Cp‡Mo(O)2Br as the catalyst. Bergman,
Herrmann, Ku¨hn, Poli, Roma˜o and their respective co-workers
used the same alkene, oxidant and solvent in their respective
studies.1,3–5 The progress of cyclooctene epoxidation was moni-
tored using 1H NMR spectroscopy at 305 K, e.g., Fig. 1S.‡ At this
comparatively low temperature, the reaction proceeded relatively
slowly with quantitative epoxidation of cyclooctene only observed
after ca. 19 h. The reaction was entirely selective for epoxidation
with no by-products (e.g. the 1,2-diol) observed. As a comparison,
no reaction was observed after 48 h between cyclooctene and
TBHP (1 : 2 mol ratio) alone or between cyclooctene and
Cp‡Mo(O)2Br (100 : 1 mol ratio) in air. Thus, Cp‡Mo(O)2Br is
incapable of catalysing either aerobic oxygenation or oxidation of
cyclooctene; for catalysis of the epoxidation both the complex and
TBHP are essential.
Plots of the consumption of cyclooctene and the production
of the cyclooctene oxide product against time are depicted in
Fig. 1(a). As the temperature used is comparatively low—the
School of Chemistry, University of New South Wales, Sydney, NSW, 2052,
Australia. E-mail: s.colbran@unsw.edu.au; Fax: +61 2 9385 6141
† The experiments were performed by M.P. as a third-year chemistry
undergraduate project at UNSW.
‡ Electronic supplementary information (ESI) available: NMR spec-
tra and details of measurements and kinetics modelling. See DOI:
10.1039/b705571d
Fig. 1 Plots of (a) conversion of cyclooctene vs. time (points = experi-
mental data; lines = fits) and (b) concentration profiles of the Mo species
obtained from the fit of these data (see text).
2746 | Dalton Trans., 2007, 2746–2748
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
aforementioned previously reported studies were all at 55 ◦C
or higher—the rates are slower allowing the subtleties of the
reaction kinetics to be more easily discerned. The kinetics are
clearly bimodal with an initiation phase characterised by slow
conversion (= ‘slow phase’) apparent at early stages (<104 s)
in the catalysis overtaken by a more rapid process (= ‘faster
1
phase’) at later times. Close inspection of the H NMR spectra
early in the catalysis reveals the four methoxy proton singlets
Scheme 1 Mechanistic scheme used to model the kinetic data (k1 ≈ k2 >
5 × 10−5 s−1; 35 kA ≈ kC = 7.3 ( 0.3) × 10−5 s−1; kB ≈ 0 s−1).
(at d 3.54, 3.31, 3.10, 2.96) that are a convenient spectroscopic
5
tag for a (g -Cp‡)Mo species,6,7 Fig. 2S.‡ The methoxy peaks
are shifted from those of Cp‡Mo(O)2Br (at d 3.39, 3.29, 3.14,
2.98), and separate experiments reveal that addition of TBHP is
responsible. Our previous observation6 of rapid reaction between
Cp‡Mo(O)2Br and aliphatic alcohols (ROH) to afford the corre-
sponding alkoxides Cp‡Mo(O)2(OR) supports the idea that the
r-peroxide Cp‡Mo(O)2(OOBut) could form, eqn (1)
formed by loss of the cyclopentadienyl (Cp‡) ligand. In the kinetic
model, Scheme 1, species C is the active catalyst that dominates
the faster phase (35 kA ≈ kC = 7.3 ( 0.3) × 10−5 s−1). Recharging
the catalysis mixture with cyclooctene and TBHP at this point
leads to immediate catalysis consistent with species C remaining
present and active at the end of the run.
Cp‡Mo(O)2Br + HOOBut = Cp‡Mo(O)2(OOBut) + HBr (1)
Two insights follow from this information. First, (perarylcy-
clopentadienyl)molybdenum(VI) dioxo species do catalyse alkene
epoxidation. Second, the cyclopentadienyl ligand is lost from
molybdenum as the catalysis proceeds, leading to the in situ
generation of a much more efficient catalyst. The latter finding
signals a warning: loss of the cyclopentadienyl ligand from a
(cyclopentadienyl)molybdenum(VI) dioxo ‘catalyst’, even if only
slight, should always be considered in catalyses of alkene epoxi-
dation, because the thus-formed molybdenum species may be the
most active catalyst. Certainly, Poli has revealed a varied and rich
aqueous chemistry for Cp*MoVI oxo species, including remarkable
stability under forcing conditions such as at the extremes of pH.9
However, Poli and co-workers have also shown loss of the Cp*
ligand may occur under certain conditions,10 so the presumption of
complete stability for an (alkylcyclopentadienyl)molybdenum(VI)
oxo catalyst during an epoxidation catalysis may be unfounded.
Loss of the alkylcyclopentadienyl ligand, even in trace amount, to
afford a more active molybdenum (per)oxo species could account
for the difficulties that have been encountered in reconciling
kinetic data5 with trends in parameters such as the steric bulk
of the ring substituents and either the proclivity or the ability
of the cyclopentadienyl to undergo catalysis-facilitating ring
slippage.
(in this scenario, the HBr released would be consumed by addition
to cyclooctene). However, the p-peroxide Cp‡Mo(O2)OBr should
also be considered, eqn (2),
Cp‡Mo(O)2Br + HOOBut = Cp‡Mo(O2)(O)Br + HOBut
(2)
given that the ‘dead-end’ p-peroxide CpMo(O2)OCl is isolated
from reaction of CpMo(O)2Cl and TBHP (see above). From the
presently available data, we can not distinguish between these
(or other) alternatives. It is noteworthy that, as the catalysis
5
proceeds the peaks attributed to this (g -Cp‡)Mo species rapidly
drop in intensity during the first 104 s (i.e. during the slow
phase), and disappear completely within 2.4 × 104 s at which
point only 60% conversion has occurred. The NMR spectra
then show no peaks either for a Cp‡Mo species or for the free
cyclopentadiene Cp‡H (methoxy peaks6 at d 3.31, 3.07). The faster
phase continues to 100% conversion, thus a molybdenum species
without a cylopentadienyl ligand must be responsible for it.8
Multiple simulations of the full kinetic data using various
underlying mechanistic schemes were attempted.‡ The data for
the faster phase, after 15 000 s, fits well to a single exponential,
Fig. 1(a). This suggests a first order process dominates the
catalysis during this phase. The fits of the faster phase data to
single exponentials afford ssubstrate = sproduct = 13 660
570 s. A
minimum of three new species are required to satisfactorily fit
the complete kinetic data and the fact that 1H NMR spectra
at the beginning of a catalytic run reveal total and immediate
transformation of Cp‡Mo(O)2Br into a new Cp‡Mo species, A,
upon mixing all reagents. If species A is (comparatively) weakly
active for epoxidation of cyclooctene, the slow phase rate data are
satisfactorily accounted for. We could not fit the complete rate
data by assuming that species A slowly converts into a more active
catalyst. Rather conversion of A into an intermediate species,
B, that is inactive for epoxidation (kB ≈ 0), which in turn is
converted into a third species, C, is the simplest kinetic scheme
that satisfactorily models the data, Scheme 1. As species B is
predicted to not build to an appreciable concentration, Fig. 1(b),
Notes and references
1 F. E. Ku¨hn, A. M. Santos and M. Abrantes, Chem. Rev., 2006, 106,
2455.
2 M. Cousins and M. L. H. Green, J. Chem. Soc., 1964, 1567.
3 M. K. Trost and R. G. Bergman, Organometallics, 1991, 10, 1172.
4 D. Chakraborty, M. Bhattacharjee, R. Kra¨tzner, R. Siefken, H. W.
Roesky, I. Uson and H.-G. Schmidt, Organometallics, 1999, 18, 106.
5 (a) F. E. Ku¨hn, A. M. Santos, A. D. Lopes, I. S. Gonc¸alves, E.
Herdtweck and C. C. Roma˜o, J. Mol. Catal. A: Chem., 2000, 164, 25;
(b) F. E. Ku¨hn, A. M. Santos, I. S. Gonc¸alves, C. C. Roma˜o and A. D.
Lopes, Appl. Organomet. Chem., 2001, 15, 43; (c) M. Abrantes, A. M.
Santos, J. Mink, F. E. Ku¨hn and C. C. Roma˜o, Organometallics, 2003,
22, 2112; (d) J. Zhao, A. M. Santos, E. Herdtweck and F. E. Ku¨hn,
J. Mol. Catal. A: Chem., 2004, 222, 265; (e) A. M. Martins, C. C.
Romao, M. Abrantes, M. C. Azevedo, J. Cui, A. R. Dias, M. T. Duarte,
M. A. Lemos, T. Lourenco and R. Poli, Organometallics, 2005, 24, 2582;
(f) F. E. Ku¨hn, A. M. Santos and W. A. Herrmann, Dalton Trans., 2005,
2483; (g) C. Freund, M. Abrantes and F. E. Ku¨hn, J. Organomet. Chem.,
2006, 691, 3718.
1
the failure to observe it in H NMR spectra acquired during the
catalysis provides no further information as to whether or not it is a
Cp‡Mo species. By the end of the catalysis run, however, species C
is predictedtobetheonlyMo-containing species present, Fig. 1(b).
Based on the 1H NMR spectra acquired at this point, species C is
6 W. M. Harrison, C. Saadeh, S. B. Colbran and D. C. Craig, J. Chem.
Soc., Dalton Trans., 1997, 3785.
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 2746–2748 | 2747
©

View Article Online
7
1H NMR spectra show two pairs of methoxy singlets; one pair each for
the two rotamers with the 2-methoxy group distal and proximal to the
metal centre; see ref. 6.
Cp‡Mo(VI) precursor which seems unlikely under the oxidising
reaction conditions (excess peroxide), and 3) we have previously
observed a facile loss of the Cp‡ ligand from a Mo(VI) species, see
ref. 6.
8 A possiblity suggested by a referee is that C is a paramagnetic
Cp‡Mo species; this seems unlikely for three reasons: 1) all peaks
in the NMR spectra are sharp at all stages of the reaction, 2)
build-up of such a species would require a clean reduction of the
9 R. Poli, Chem.–Eur. J., 2004, 10, 332.
10 E. Collange, L. Metteau, P. Richard and R. Poli, Polyhedron, 2004, 23,
2605.
2748 | Dalton Trans., 2007, 2746–2748
This journal is
The Royal Society of Chemistry 2007
©