Building stereoselectivity into a chemoselective ring-opening metathesis polymerization catalyst for alternating copolymerization

Article abstract of DOI:10.1002/anie.200906846

(Figure presented) This way and that: Ruthenium complexes with asymmetric bidentate phosphine ligands bearing two substituents that differ in size (green spheres) produce a totally alternating copolymer of norbor Figure Presented nene and eyelooetene by ring-opening metathesis polymerization (ROMP). The E/Z ratio can be Influenced systematically by changing the bulklness of the aryl sulfonate ligand (blue rectangles).

Full text of DOI:10.1002/anie.200906846

Communications
DOI: 10.1002/anie.200906846
Designed Stereoselectivity
Building Stereoselectivity into a Chemoselective Ring-Opening
Metathesis Polymerization Catalyst for Alternating
Copolymerization**
Sebastian Torker, Andre Mꢀller, and Peter Chen*
Chemoselectivity, regioselectivity, and stereoselectivity in
homogeneous catalysis are ordinarily difficult to separate
from one another. Some time ago, we reported the mecha-
nism-based design of a chemoselective ring-opening meta-
thesis polymerization (ROMP) catalyst 1 for the alternating
copolymerization of norbornene and cyclooctene wherein
two diastereomeric carbenes interchange by inversion at the
stereogenic ruthenium center after each productive meta-
thesis step.[1,2] Here we report ruthenium-based catalysts in
which chemo- and stereoselectivity are mechanistically
designed according to a simple scheme involving bulky
sulfonates 4a–4 f (Scheme 1) as substitutes for the usual
halide ligands, with the two types of selectivity determined by
the steric bulk in two orthogonal planes.
We have pushed the original catalystꢀs (1)[2] performance
in terms of sequence selectivity (75% alternation with strong
dependence on the temperature and the norbornene/cyclo-
octene ratio, see also Figure S7 in the Supporting Informa-
tion) to the limit by developing more bulky bidentate
phosphines. Complex 2 shows a higher activity for copoly-
merization, and the faster initiating (see Figure S6 in the
Scheme 1. Catalysts tested in co- and homopolymerizations.
Supporting Information) variant 5, in particular, is able to
produce a copolymer within seconds after catalyst addition, as
evident by the immediate formation of a gel. The less-strained
cyclooctene seems to react cleanly at a competitive rate with
only one diastereomeric form of the carbene, whereas
norbornene shows the same high activity towards both. A
polymer produced from a 1:5 mixture of norbornene and
cyclooctene contains about 10% polynorbornene units, as
expected for this high selectivity which seems to be only
controlled by strain[3] release of monomers. Unexpectedly, the
stereochemistry of the olefinic moieties was close to 90%
trans, which inspired us to design for Z (or cis) selectivity (see
bottom spectrum in Figure 1).
Therefore, we modified the parent catalyst 2 by replace-
ment of the chloride with more bulky anions. Catalysts 2a–2 f
could easily be prepared by overnight reaction of the halide-
containing precursor 2 with excess silver sulfonates 4a–4 f in
benzene, with clean conversion into new carbene species,
which were used in situ. For polymerization experiments, a
norbornene/cyclooctene ratio of 1:20 was chosen to ensure
almost complete alternation (solvent: CH2Cl2; norbornene/
catalyst ratio = 2000:1). The reactivity of the sulfonate com-
plexes is much lower, and the following relative rates are
found: 5/2/2a/2d = 1:(1/30):(1/300):(1/750) (see Table S4 in
the Supporting Information).
13C NMR analysis of the polymers[4] in CDCl3 reveals a
systematic trend (Figure 1). Almost no signals for polynor-
bornene can be detected around d = 133–134 ppm. Whereas
the chloride catalyst 2 mainly produces two sharp signals at
d = 134.9 and 128.5 ppm, which can be assigned to an
alternating unit with trans stereochemistry (ttt triads for
both termini), progressively increasing the bulkiness of the
sulfonates (2a–2d) increases the amount of cis double bonds
up to 51%. We explain this by the mechanism depicted in
Scheme 2. The substituents marked with green on the
phosphorus atom control the chemoselectivity by discrim-
inating cyclic monomers by their ring strain, as has been
[*] S. Torker, A. Mꢀller, Prof. Dr. P. Chen
Laboratorium fꢀr Organische Chemie, Eidgenꢁssische Technische
Hochschule (ETH)
Zꢀrich (Switzerland)
Fax: (+41)44-632-1280
E-mail: peter.chen@org.chem.ethz.ch
[**] We thank Dr. Bernd Schweizer and Dr. Michael Solar from the X-ray
crystallographic service of the ETH, and Dr. Philip Zumbrunnen and
Dr. Rainer Frankensteiner from the NMR service of the Laborato-
rium fꢀr Organische Chemie (ETH) for performing the corre-
sponding experiments.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200906846.
3762
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 3762 –3766

Angewandte
Chemie
zations with pure 2d, however, show a decrease in Z selec-
tivity from 51 to 36%, which is still better than with catalyst 2
or 2a (see Figures S11 and S12 in the Supporting Informa-
tion). The isolation and structural characterization of 2d
clearly shows that a ruthenium sulfonate with the sulfonate
group cis to the phosphine has been prepared in the in situ
synthesis. The crystal structures of complexes are shown for
2d (Figure 2), 2e (Figure 3), 2a, and 2 f (see Figures S1 and S5
in the Supporting Information). Solvent effects suggest an
explanation for the difference in Z selectivity for the com-
plexes with the bulkiest sulfonates. If one changes the solvent
from CH2Cl2 to hexane, in situ prepared 2e shows an increase
in Z selectivity from 19 to 42% (see Figure S14 in the
Supporting Information), whereas
a slight decrease is
Figure 1. 13C NMR spectra showing the alternating copolymerization of
norbornene and cyclooctene in CH2Cl2 with increasing cis content from
bottom to top using asymmetric catalysts 2–2d prepared in situ and
low selective catalyst 2e. (See 1H NMR spectra in Figures S8 and S14
in the Supporting Information.)
observed for in situ prepared 2d (51 to 42%; see Figure S13
in the Supporting Information). In the latter case this might
be due to a change in transition-state energies, whereas with
2e dissociation of the sulfonate ligand most likely happens to
a greater extent in the more polar solvent. Furthermore, the
shown previously.[1] The introduction of a
bulky sulfonate (marked with blue), on the
one hand, forces the propagating polymer
chain to turn away from it. On the other hand,
p complexation of the olefin to the metal
center trans to the phosphine ligand and
subsequent formation of a metallacyclobutane
will happen in such a way as to obtain a
compromise between a) the steric repulsion of
the cycloolefin and the sulfonate and b) the
steric interaction between the all-cis substitu-
ents in the metallacyclobutane structure.
Whether the former or latter interaction is
stronger will determine the stereochemical
outcome. According to that mechanistic pic-
ture, we believe that, in first generation
Grubbs systems, chemoselectivity is controlled
by changing substituents in the plane contain-
ing the carbene, the ruthenium atom, and the
phosphorus center, which during the catalytic
cycle, also contains the metallacyclobutane,
whereas the E/Z ratio is influenced by sub-
stituents in a plane perpendicular to the first
one. To test the orthogonality of the E/Z
control from the alternation, symmetric var-
iants 3–3c of the catalyst system were used for
homopolymerization of norbornene; in these
cases a similar trend was observed with slightly
less than 50% cis double bonds produced with
the most Z-selective catalyst in this series, 3c,
(see Figure S9 in the Supporting Information).
The situation, however, gets more compli-
cated if one tries the polymerizations with
analytically pure catalysts or even bulkier
sulfonates (2e). Detailed synthetic procedures
and NMR spectra are given in the Supporting
Information. Isolated 2a yields the same
polymer as 2a prepared in situ. Polymeri- Scheme 2. Mechanistic explanation of the stereochemistry of double bonds.
Angew. Chem. Int. Ed. 2010, 49, 3762 –3766
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3763

Communications
in the chemical shift to d = 16.43 ppm for the even more bulky
and more electron-donating 2e. A similar trend is given for
the chemical shifts of the carbene carbon atom: 2 (d =
284.58 ppm), 2a (d = 295.31 ppm), 2b (d = 296.17 ppm), 2c
(d = 297.01 ppm), 2d (297.00 ppm), 2e (d = 297.12 ppm), and
2 f (d = 295.69 ppm). One may therefore hypothesize, that
sulfonate dissociation begins to spoil the Z selectivity with 2d,
which is suppressed when there is an excess of the sulfonate
salt in the in situ experiment. An experimental test of the
conjecture was done by adding 0.15 equivalents of silver
sulfonate 4d to isolated complex 2d, which pushes the
Z selectivity in the polymerization back up to 51%. Never-
theless, the size of sulfonate 4d is most probably close to the
limit, where dissociation can still be controlled, since we
observe low selectivity with the even bulkier 4e in CH2Cl2.
Several examples of chemoselective metathesis have been
reported in the literature. Stereogenic-at-metal ruthenium
complexes bearing enantiomerically pure bidentate N-heter-
ocyclic carbene/binaphthol ligands have been used by Hov-
eyda et al. to effectively catalyze the ring-opening cross-
metathesis (ROCM) of norbornene derivatives with two
equivalents of styrene.[5,6] Since chiral molybdenum catalysts
would readily give polymers under these conditions, effects
other than pure ring strain[3] have to account for this high
chemoselectivity. Since this catalyst is stereogenic at the
ruthenium center, one would expect, based on our previous
mechanistic picture,[1] two different diastereomeric carbenes
to take part in the catalytic cycle. If the highly strained
norbornene would react faster than styrene with both sides, a
ROMP polymer would be the expected result. It seems,
however, that the release of ring strain is only needed to effect
a change in the carbeneꢀs position to the energetically
disfavored side. Once there, the norbornene unit on the
carbene provides too much steric crowding so that only the
smaller styrene can react.
Figure 2. Crystal structure of complex 2d (ORTEP plot, 20% proba-
bility ellipsoids). Crystal structures for complexes 2a and 2 f are given
in the Supporting Information.
Figure 3. Crystal structure of complex 2e (ORTEP plot, 20% probabil-
ity ellipsoids).
À
crystal structure of 2e shows a distinct elongation of the Ru
O3SAr bond versus that in 2d [2.141(4) ꢁ versus 2.076(2) ꢁ].
The great steric bulk of the ortho-tert-butyl groups in 2e
results in the sulfur atom also being twisted out of the
aromatic plane.
Similar, achiral catalysts have also been used to introduce
chemoselectivity into polymers, taking advantage of a suitable
set of monomers which have a higher tendency to form
alternating linkages by kinetic control rather than undergoing
homopolymerization based on thermodynamic aspects (strain
release).[7–11] An early example is the alternating copolymer-
ization of cyclopentene and norbornene using RuCl3 in the
presence of phenol as a co-catalyst or solvent.[8] Hydrogen-
bonded solvent cages around the active site were proposed as
an explanation, which prevents the more reactive but bulkier
norbornene from performing two consecutive metathesis
steps. An example of this extreme case in chemoselective
copolymerization, where the rates of homopolymerization for
both monomers tend to zero, is shown for the alternating
copolymerization of the enantiomers of 1-methylnorbornene
catalyzed by ReCl5.[10] Blechert, Buchmeiser, and co-workers
recently reported the synthesis of a highly alternating
norbornene/cyclooctene copolymer with a cis content of
approximately 50%, comparable to what we obtained with
catalyst 2d, using Grubbs-type initiators containing an
unsymmetrical, chiral N-heterocyclic carbene ligand.[12] The
high tendency for alternation was explained by an enhanced
cyclooctene insertion rate into a norbornene-initiator-derived
terminus and, vice versa, an enhanced norbornene insertion
À
Another indication for a weaker Ru O3SAr bond comes
from comparing the chemical shifts (CH2Cl2) of the carbene
proton (see Table S3 in the Supporting Information).
Whereas the chloride catalyst 2 displays a doublet at d =
15.65 ppm, the corresponding signals are shifted downfield
for 2a to d = 16.29 ppm and even further for 2b (d =
16.44 ppm), 2c (d = 16.54 ppm), and 2d (d = 16.55 ppm).
One would expect that a larger electron donation by larger
aliphatic groups on the sulfonate would result in an upfield
shift of this proton, but exactly the opposite is observed. This
finding may be explained by a looser binding of the sulfonate
ligand, which as a consequence makes the ruthenium center
more electropositive. Only if one compares the chemical shift
of 2a (d = 16.29 ppm) versus 2 f (d = 16.34 ppm) can one see
that the more electron rich tosylate produces an upfield shift
compared to the benzenesulfonate because the substitution in
the para position leads to no increased steric interaction to
counteract the electronic (inductive) effect. In general, one
probably has to consider both effects, since there is a decrease
3764
www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 3762 –3766

Angewandte
Chemie
rate into a cyclooctene-initiator-derived termi-
nus. In other words they speculated that the
norbornene-initiator-derived terminus makes
the catalyst more crowded, thereby providing
steric energy for a faster cyclooctene insertion,
but too bulky, with the result that it slows down
consecutive norbornene incorporation. No
explanation for the stereochemistry was given.
Furthermore they recently reported for the
same reaction high selectivity towards alterna-
tion and even higher Z selectivity (ca. 60%) by
applying complexes containing nonchiral
unsymmetrical
N-heterocyclic
carbene
ligands.[13]
More than a decade ago, Grubbs and co-
workers studied the influence of ligands in first
generation ruthenium systems: “Phosphines
(donor), which are larger and more electron
donating, and likewise halogens (acceptor),
Scheme 3. Stereogenic-at-metal molybdenum complex 6 published by Schrock and
Hoveyda for the Z-selective polymerization of norbornene derivatives and proposed
metallacyclobutane structure 6 MCB with all-cis substituents.
which are smaller and more electron withdrawing, lead to
more active catalysts.”[14] For this reason, we chose electron-
withdrawing sulfonates as substitutes for chloride. It is at this
point worth noticing that a Hoveyda–Grubbs type ruthenium
complex with two trifluoromethanesulfonate ligands and a
mixed anionic complex with one sulfonate and one chloride
ligand have already been shown to be active catalysts.[15]
Inspired by mechanistic work by Eisenstein and co-
workers,[16,17] the research groups of Schrock and Hoveyda
recently published stereoselective molybdenum imido cata-
lysts[18,19] with high Z selectivity that are very effective in
enantioselective ring-opening/cross-metathesis[20] and also
ROMP of norbornene derivatives (6).[21] Apparently there
seem to be similarities between molybdenum and ruthenium
systems, even though the molybdenum complexes (for
example 6) have a tetrahedral structure (Scheme 3). Theo-
retical investigations[16,17] on molybdenum, tungsten, and
rhenium systems suggest that the more active catalysts
require an acceptor ligand (an alkoxide) in combination
with a donor ligand (a pyrrolide) to lower the energetic
barrier that is governed by geometrical distortion during
formation of the olefin p complex. The most stable p com-
plexes are formed by a trans approach of the olefin with
respect to the donor ligand. The subsequent cycloaddition
produces the metallacyclobutane trans to the donor (6 MCB).
The high Z selectivity reached with stereogenic-at-molybde-
num–imido complexes arises from the combination of a big
phenoxide acceptor ligand (in the catalyst cis to the metal-
lacyclobutane) with a small (close to the metal) adamanty-
limido ligand, forcing the sterically favored syn-alkylidene to
react with an olefin to give an all-cis metallacyclobutane
(Scheme 3). Similar, the bulky sulfonates in the complexes in
the present report would be cis to the metallacycle
(Scheme 2).
the stereochemistry.[24] Recently Verpoort and co-workers
reported a remarkable steric impact on the E/Z selectivity in
the cross-metathesis of allylbenzene and acrylonitrile.[25] For a
more detailed discussion of these last three examples, see the
Supporting Information.
In conclusion, we have shown that in first generation
Grubbs systems chemoselectivity is controlled by changing
the substituents in the plane containing the carbene, the
ruthenium atom, and the phosphorus atom, whereas the E/
Z ratio is influenced by substitution in a perpendicular plane
with respect to the first one. This puts a bulky anionic
substituent cis to the metallacycle. Current work is based on
designing further ligands that fulfil the above-mentioned
requirements to extend the applications to a broad range of
metathesis tasks.
Received: December 4, 2009
Published online: April 15, 2010
Keywords: alternating copolymers · copolymerization ·
.
metathesis · ruthenium · Z selectivity
[1] M. Bornand, P. Chen, Angew. Chem. 2005, 117, 8123 – 8125;
Angew. Chem. Int. Ed. 2005, 44, 7909 – 7911.
[2] M. Bornand, S. Torker, P. Chen, Organometallics 2007, 26, 3585 –
3596.
[3] K. B. Wiberg, Angew. Chem. 1986, 98, 312 – 322; Angew. Chem.
Int. Ed. Engl. 1986, 25, 312 – 322.
[4] K. J. Ivin, J. C. Mol, Olefin Metathesis and Metathesis Polymer-
ization, Academic Press, San Diego, 1997.
[5] A. H. Hoveyda, D. G. Gillingham, J. J. Van Veldhuizen, O.
Kataoka, S. B. Garber, J. S. Kingsbury, J. P. A. Harrity, Org.
Biomol. Chem. 2004, 2, 8 – 23.
[6] J. J. Van Veldhuizen, D. G. Gillingham, S. B. Garber, O.
Kataoka, A. H. Hoveyda, J. Am. Chem. Soc. 2003, 125, 12502 –
12508.
[7] M. F. Ilker, E. B. Coughlin, Macromolecules 2002, 35, 54 – 58.
[8] B. Al Samak, V. Amir-Ebrahimi, D. G. Corry, J. G. Hamilton, S.
Rigby, J. J. Rooney, J. M. Thompson, J. Mol. Catal. A 2000, 160,
13 – 21.
Another study in which the stereochemistry of alkene
metathesis was investigated was carried out about 30 years
ago by Casey et al., who studied the degenerate metathesis of
terminal alkenes,[22,23] and at the same time Basset and co-
workers speculated that large ligands in the metal sphere may
direct the approach of the incoming olefin and thus influence
Angew. Chem. Int. Ed. 2010, 49, 3762 –3766
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3765

Communications
[9] V. Amir-Ebrahimi, J. J. Rooney, J. Mol. Catal. A. 2004, 208, 115 –
[17] X. Solans-Monfort, E. Clot, C. Coperet, O. Eisenstein, J. Am.
Chem. Soc. 2005, 127, 14015 – 14025.
121.
[10] J. G. Hamilton, K. J. Ivin, J. J. Rooney, L. C. Waring, J. Chem.
Soc. Chem. Commun. 1983, 159 – 161.
[18] S. J. Malcolmson, S. J. Meek, E. S. Sattely, R. R. Schrock, A. H.
Hoveyda, Nature 2008, 456, 933 – 937.
[11] A. R. Song, K. A. Parker, N. S. Sampson, J. Am. Chem. Soc.
2009, 131, 3444 – 3445.
[12] K. Vehlow, D. Wang, M. R. Buchmeiser, S. Blechert, Angew.
Chem. 2008, 120, 2655 – 2658; Angew. Chem. Int. Ed. 2008, 47,
2615 – 2618.
[13] M. Lichtenheldt, D. Wang, K. Vehlow, I. Reinhardt, C. Kꢂhnel,
U. Decker, S. Blechert, M. R. Buchmeiser, Chem. Eur. J. 2009,
15, 9451 – 9457.
[19] E. S. Sattely, S. J. Meek, S. J. Malcolmson, R. R. Schrock, A. H.
Hoveyda, J. Am. Chem. Soc. 2009, 131, 943 – 953.
[20] I. Ibrahem, M. Yu, R. R. Schrock, A. H. Hoveyda, J. Am. Chem.
Soc. 2009, 131, 3844 – 3845.
[21] M. M. Flook, A. J. Jiang, R. R. Schrock, P. Mꢂller, A. H.
Hoveyda, J. Am. Chem. Soc. 2009, 131, 7962 – 7963.
[22] C. P. Casey, H. E. Tuinstra, J. Am. Chem. Soc. 1978, 100, 2270 –
2272.
[14] E. L. Dias, S. T. Nguyen, R. H. Grubbs, J. Am. Chem. Soc. 1997,
119, 3887 – 3897.
[23] C. P. Casey, S. W. Polichnowski, A. J. Shusterman, C. R. Jones, J.
Am. Chem. Soc. 1979, 101, 7282 – 7292.
[15] J. O. Krause, O. Nuyken, K. Wurst, M. R. Buchmeiser, Chem.
Eur. J. 2004, 10, 777 – 784.
[24] J. L. Bilhou, J. M. Basset, R. Mutin, W. F. Graydon, J. Am. Chem.
Soc. 1977, 99, 4083 – 4090.
[16] A. Poater, X. Solans-Monfort, E. Clot, C. Coperet, O. Eisenstein,
J. Am. Chem. Soc. 2007, 129, 8207 – 8216.
[25] N. Ledoux, A. Linden, B. Allaert, H. V. Mierde, F. Verpoort,
Adv. Synth. Catal. 2007, 349, 1692 – 1700.
3766 www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 3762 –3766