On the relationship between structure and reaction rate in olefin ring-closing metathesis

Article abstract of DOI:10.1039/c0cc02440f

In the RCM reactions of a series of simple α,ω-dienes, the relative order of reactivity has been unambiguously determined showing that cyclohexene forms faster than cyclopentene or cycloheptene. 1,5-Hexadiene inhibits the RCM of 1,7-octadiene; 1,5-hexadiene cannot progress to the RCM product (cyclobutene) but forms an unexpectedly stable cyclic η²-complex.

Full text of DOI:10.1039/c0cc02440f

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
On the relationship between structure and reaction rate in
olefin ring-closing metathesisw
Ian W. Ashworth,a Davide Carboni,b Ian H. Hillier,c David J. Nelson,b Jonathan M. Percy,*b
Giuseppe Rinaudob and Mark A. Vincentc
Received 8th July 2010, Accepted 17th August 2010
DOI: 10.1039/c0cc02440f
In the RCM reactions of a series of simple a,x-dienes, the
relative order of reactivity has been unambiguously determined
showing that cyclohexene forms faster than cyclopentene or
cycloheptene. 1,5-Hexadiene inhibits the RCM of 1,7-octadiene;
1,5-hexadiene cannot progress to the RCM product (cyclobutene)
but forms an unexpectedly stable cyclic g2-complex.
invoked to explain the appearance of the 6-membered ring
product. It is also interesting that various attempts to populate
6-membered ring species as thermodynamic products by
raising reaction temperatures and applying an overpressure
of ethene failed to deliver clean results. The presence of
functional groups on the substrates used for target synthesis
inevitably complicates the analysis of reaction outcomes.
We were therefore interested to discover if there were
significant rate differences in RCMs of simple a,o-dienes in
solution; rigorous characterisation of the behaviour of the
parent diene systems would then enable the assignment of
kinetic products to be made more confidently.
Alkene metathesis has become a vital part of the synthetic
repertoire, and has been applied in many syntheses of complex
natural products. Much of what we know about the relation-
ship between diene structure and reactivity has been inferred
from reaction yields from the huge body of results that has
accumulated in the area of target synthesis; quantitative
studies that relate structure and reactivity are very rare.1,2
Arguments based on kinetic versus thermodynamic control
have been advanced by Schmidt and Nave3 to explain reaction
outcomes from systems that can form 5- or 6-membered rings
competitively. For example, Scheme 1 shows the completely
selective double RCM of 1 to 2 when exposed to the first
generation Grubbs pre-catalyst 3. Because 3 is known to react
slowly with disubstituted alkenes (and particularly those with
an allylic oxygen function), it seems extremely unlikely that 2
could be a thermodynamic product. Substrate 4 with a free
hydroxyl group affords dihydrofuran 5 and dihydropyran 6 in
equal proportions and an hydroxyl group directing effect is
A
recent mass spectrometric study by Metzger and
co-workers4 identified ions on the cyclisation pathways of
simple a,o-dienes 7a–d catalysed by Grubbs’ first generation
pre-catalyst 3 (Scheme 2). Collisional energies were used to
estimate the rates at which the Z2-complexes 8a–d progressed
onward through the cycle. Metzger reported that the most
reactive system derived from 1,7-octadiene 7c (n = 2), which
goes on to form cyclohexene; the Z2-complex 8c progressed
more rapidly than those from 1,6-heptadiene 7b (n = 1) and
1,8-nonadiene 7d (n = 3), which were very similar. This is
contrary to what appears to be a generally held belief that
5-membered rings are the fastest to form by RCM.3,4
We have investigated the cyclisation rates of simple
a,o-dienes in solution using the Grubbs’ second generation
pre-catalyst 9 (Scheme 2) which is reactive enough to permit
the cycloalkenes to equilibrate fully.5 NMR spectroscopy was
the method of choice because it would allow concentrations to
be measured without the introduction of artefacts arising from
sample concentration or loss by evaporation. In the first
instance, a mixture of 7b (3.3 mM), 7c (3.4 mM) and 7d
(3.5 mM) in CDCl3 containing 1,3,5-trimethoxybenzene as the
internal standard,6 was prepared and exposed to pre-catalyst 9
(1 mol%) at 25 1C in the magnet of a 600 MHz spectrometer
(see the ESIw for full details). We were careful to determine
Scheme 1
a AstraZeneca Global Research and Development,
Silk Road Business Park, Charter Way, Macclesfield,
SK10 2NA, UK
b WestCHEM Department of Pure and Applied Chemistry,
Thomas Graham Building, 295 Cathedral Street, Glasgow G1 1XL,
UK. E-mail: jonathan.percy@strath.ac.uk;
Fax: +44 (0)141 548 4822; Tel: +44 (0)141 548 4398
c School of Chemistry, University of Manchester, Oxford Road,
Manchester M13 9PL, UK
w Electronic supplementary information (ESI) available: T1 values for
dienes and cycloalkenes, typical NMR spectra of reaction solutions,
concentration/time data and plots, Cartesian coordinates, and energies
of 10a–c, 11a–c, 12b, 12c, 13b and 13c. See DOI: 10.1039/c0cc02440f
Scheme 2
c
This journal is The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 7145–7147 7145

View Article Online
Fig. 1 Competition experiment in chloroform showing product
formation.
Fig. 2 Inhibition of cyclohexene formation by hexadiene.
T1’s for all the reaction components, and set D1 to 5T1 to
ensure accurate integration. Dichloromethane, rather than
chloroform, is the default solvent for RCM in the synthetic
literature; we elected to work in d-chloroform because it was
more economical, but selected experiments were also run in
d2-dichloromethane.
while the second mixed 7a and 7c (5 mM in 7c, 5.7 mM in 7a
with 0.93 mol% 9). While the conversion profile of cyclohexene
formation in the 7b/7c mixture was almost identical to that
from the experiment containing 7c alone (10 mM, 1 mol% 9),
cyclohexene formed significantly more slowly and reached
lower conversion when 7a was present in the reaction mixture
(Fig. 2).
The dispersion between key signals was excellent and we
were able to collect a full concentration/time profile for each
cycloalkene as the reactions proceeded to their end points
(Fig. 1). The identity of the most reactive diene was beyond
dispute; cyclohexene formed most rapidly from the reaction, in
agreement with Metzger’s earlier inference. Cyclopentene
formed less rapidly, and cycloheptene was the slowest of the
three. We carried out a series of single substrate experiments
(each in duplicate) and were able to reproduce the reactivity
order obtained from the three-substrate experiment, indicating
that the profiles do not represent an artefact of a competition
process. This finding clearly contradicts the assertion that
5-membered rings in general form fastest by RCM.
Electronic structure calculations at the M06-L/B2/CPCM//
M06-L/B2 level were used to probe further. We have pre-
viously shown that the M06-L functional is appropriate for
describing these systems.12 The basis used, B2, consisted of the
SDD ECP and corresponding basis set on Ru,13 with an
f-function of exponent 0.5780, and with a 6-311G** basis on
all other atoms. This calculation gave quite different geometries
for the two initial Z2-complexes (Fig. 3). The Z2-complex 11c
(Fig. 3b) lies 6.6 kcal molꢀ1 below the starting 14e alkylidene
complex. In this Z2-complex the p orbitals of the CQC and
RuQC double bonds point towards each other and are thus
ready to form the metallocycle, whereas in the Z2-complex
from 1,5-hexadiene (11a, Fig. 3a) the p orbitals are at right
angles, so this is not a reactive conformation. This complex lies
10.6 kcal molꢀ1 below the starting 14e alkylidene complex,
and is noticeably more stable than 11c. This arrangement of
alkene and alkylidene is very similar to the one found in an
untethered methylidene/ethene complex.14 As shown in Fig. 4,
Z2-complex 11a is competitive with the lowest point (metallo-
cyclobutane 12c at ꢀ10.2 kcal molꢀ1) on the energy surface
for cyclohexene formation. These results confirm the origin of
the effect observed by Metzger and begin to explain the very
low cross metathetical reactivity of 7a. They also support
the observation that 7a retards cyclohexene formation,
presumably by sequestering active catalyst in a non-productive
complex,15 competitively with productive binding to 7c. The
heptadiene metathesis begins with complex 11b; while this is
also a much more stable complex than 11c, it can progress and
therefore does not sequester catalyst. Inspection of the next
It was also possible to measure the solution concentration of
ethene, which reaches 90–95% of its maximum value before
slow egress begins.7 The second generation pre-catalyst 9 was
also present at the end of the reaction (from the 1H NMR
spectrum), so these mixtures may represent equilibrium
compositions because all the reaction components are present.
In both dichloromethane and chloroform, 7b and 7c were
converted completely within experimental error to their
cycloalkene products (all the reactions were slightly faster in
dichloromethane), whereas the reaction of 1,8-nonadiene 7d
forms small quantities (ca. 5%) of oligomeric material in
addition to cycloheptene in either solvent.8 There was a minor
difference in half-life in the two solvents for all the dienes, with
the chloroform reactions typically slightly slower.z
The behaviour of 1,5-hexadiene 7a, was also examined;
Metzger and coworkers found that the initial Z2-complex
derived from this diene and pre-catalyst 3 was much more
stable than expected. Other authors have suggested that
chelate complexes containing a hexadienyl motif are stabilised;
7a is known to undergo ADMET polymerisation more slowly
than 1,9-decadiene, for example.9–11
We therefore examined the behaviour of this shorter diene
under the conditions used to cyclise 7b–7d and found that 7a
was unreactive at 10 mM (1 mol% 9), with no change in the
1H NMR spectrum; even at 250 mM, conversion was very
slow, reaching only 20% after 8 hours. In two further experiments,
the first contained an equimolar (4.5 mM in each) mixture of
7b and 7c (1.1 mol% 9 based on total diene concentration),
Fig. 3 Optimised geometries (M06-L/B2) for (a) 11a and (b) 11c.
c
7146 Chem. Commun., 2010, 46, 7145–7147
This journal is The Royal Society of Chemistry 2010

View Article Online
Fig. 4 Partial free energy profiles for metathesis of hexadiene 7a (red), and heptadiene 7b and octadiene 7c (black) with pre-catalyst 9.
part of the energy surface predicts that cyclohexene will form
more quickly than cyclopentene if metallocyclobutane break-
down is rate-determining, consistent with the behaviour of the
dienes in the NMR experiments.y
higher dienes lies beyond the scope of this manuscript and will be
reported elsewhere.
1 A. H. Hoveyda and A. R. Zhugralin, Nature, 2007, 450, 243.
2 Recently, Monfette and Fogg have emphasised the importance of
thermodynamic control as
a general determinant of RCM
In conclusion, we have shown that the order of kinetic
reactivity (6 4 5 4 7) of simple a,o-dienes in RCM is not the
one assumed in the literature, or referred to in investigations of
more complex systems; cyclohexene has been shown to form
more rapidly than cyclopentene, with cycloheptene formation
significantly slower. Further, the 1,5-hexadienyl motif is
shown to be far from innocent in metathesis, with 1,5-hexadiene
inhibiting the rapid cyclisation to cyclohexene. While cyclo-
butene does not form, care should be taken in synthetic
planning lest the formation of stable Z2-complexes analogous
to 11a impede RCMs involving complex substrates.
outcome. See: S. Monfette and D. E. Fogg, Chem. Rev., 2009,
109, 3783.
3 B. Schmidt and S. Nave, Adv. Synth. Catal., 2007, 349, 215.
4 H. Y. Wang, W. L. Yim, T. Kluner and J. O. Metzger, Chem.–Eur. J.,
2009, 15, 10948.
5 C. W. Lee and R. H. Grubbs, Org. Lett., 2000, 2, 2145.
6 C. S. Adjiman, A. J. Clarke, G. Cooper and P. C. Taylor, Chem.
Commun., 2008, 2806.
7 For a recent measurement of ethene solubility, see: E. F. van der
Eide, P. E. Romero and W. E. Piers, J. Am. Chem. Soc., 2008, 130,
4485. The roles of ethene and the alkylidene population will be
discussed more fully elsewhere.
8 Kress has generated and characterised oligomeric species
by ROMP of cycloheptene; J. Kress, J. Mol. Catal. A, 1995,
102, 7.
9 Z. Wu, A. D. Benedicto and R. H. Grubbs, Macromolecules, 1993,
26, 4975.
10 K. B. Wagener, K. Brzezinska, J. D. Anderson, T. R. Younkin,
K. Steppe and W. DeBoer, Macromolecules, 1997, 30, 7363.
11 For an isolated Z2-complex containing a 1,5-hexadienyl motif, see:
J. A. Tallarico, P. J. Bonitatebus and M. L. Snapper, J. Am. Chem.
Soc., 1997, 119, 7157.
12 S. Pandian, I. H. Hillier, M. A. Vincent, N. A. Burton,
I. W. Ashworth, D. J. Nelson, J. M. Percy and G. Rinaudo, Chem.
Phys. Lett., 2009, 476, 37.
This work was supported by AstraZeneca (Industrial CASE
Award to DJN), the EPSRC Initiative in Physical Organic
Chemistry 2 (EP/G013160/1 and EP/G013020/1, fellowship to
DC) and the University of Strathclyde (fellowship to GR).
Notes and references
z Only the reaction of 7c in dichloromethane was clearly first order, so
we cannot use first order rate constants to quantify the differences
between the reactions at this stage. The minor difference in half-life
observed for all the dienes suggests that CDCl3 is a useful solvent for
the study of these reactions.
13 D. Andrae, U. Haeussermann, M. Dolg, H. Stoll and H. Preuss,
Theor. Chim. Acta, 1990, 77, 123.
y Metallocyclobutane breakdown was identified as the highest energy
barrier on the surface for cyclohexene formation in our earlier
14 B. F. Straub, Angew. Chem., Int. Ed., 2005, 44, 5974.
15 I. C. Stewart, B. K. Keitz, K. M. Kuhn, R. M. Thomas and
R. H. Grubbs, J. Am. Chem. Soc., 2010, 132, 8534.
16 E. L. Dias, S. T. Nguyen and R. H. Grubbs, J. Am. Chem. Soc.,
1997, 119, 3887.
publication.12 However, catalyst formation from
9 and initial
alkylidene transfer from 10 to 11 may be at least partially rate-
determining,16 making the identification of the rate determining step
quite complex. A full analysis of the energy surfaces of these and
c
This journal is The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 7145–7147 7147