Isomerization during olefin metathesis: An assessment of potential catalyst culprits

Article abstract of DOI:10.1002/cctc.201300886

Two ruthenium hydride complexes commonly proposed as agents of unintended isomerization during olefin metathesis are examined for their activity in isomerization of estragole, a representative allylbenzene. Neither proves kinetically competent to account for the levels of isomerization observed during cross-metathesis of estragole by the second-generation Grubbs catalyst. A structure-activity analysis of selected ruthenium hydride complexes indicates that higher isomerization activity correlates with a more electrophilic metal center. It wasn't me: Two Ru hydrides thought to trigger double-bond migration during olefin metathesis are examined for their isomerization activity. Neither can account for the high levels of undesired isomerization seen during self-metathesis of estragole, a model allylbenzene substrate. Higher activity is found to correlate with a less electron-rich Ru center. Copyright

Full text of DOI:10.1002/cctc.201300886

CHEMCATCHEM
COMMUNICATIONS
DOI: 10.1002/cctc.201300886
Isomerization During Olefin Metathesis: An Assessment of
Potential Catalyst Culprits
Carolyn S. Higman, Lucie Plais, and Deryn E. Fogg*[a]
Two ruthenium hydride complexes commonly proposed as
agents of unintended isomerization during olefin metathesis
are examined for their activity in isomerization of estragole,
a representative allylbenzene. Neither proves kinetically com-
petent to account for the levels of isomerization observed
during cross-metathesis of estragole by the second-generation
Grubbs catalyst. A structure–activity analysis of selected ruthe-
nium hydride complexes indicates that higher isomerization
activity correlates with a more electrophilic metal center.
Grubbs and Mol groups found that closely related carbonyl
complexes form on exposure of the Grubbs catalysts to oxy-
gen.[17f,h] The latter findings raised the possibility that incom-
plete air exclusion could contribute to isomerization.
Although both Ru-1[16] and Ru-2[22,23] have been shown to
be isomerization-active, it is unclear whether they are kinetical-
ly competent to account for the levels of isomerization ob-
served during metathesis reactions. Here, we have examined
this point by explicitly comparing their isomerization activity
with rates of isomerization observed during self-metathesis of
the same substrate under identical conditions. As a further
goal, we sought to establish clear-cut structure–activity correla-
tions for a series of selected Ru hydride complexes, which
could ultimately aid the identification of plausible culprits.
For both purposes, we chose to study estragole (1) as a rep-
resentative, functionalized allylbenzene of keen interest as a re-
newable platform chemical.[24] Recent parallel work has pointed
out the high-value of products accessible from 1 and isomeric
phenylpropenoids through metathesis.[25–27] As noted above, al-
lylbenzene substrates readily undergo isomerization to bring
the double bond into conjugation with the phenyl ring, to the
extent that they have been used as test substrates to assess
potential isomerization inhibitors.[9] In a representative recent
study, Bruneau and co-workers reported just 34% cross-meta-
thesis (CM) on treating 2-methoxy-4-allyl-phenol (eugenol)
with methyl acrylate, when using GII as catalyst.[7] Isomerized
species, including secondary metathesis products, accounted
for the balance.
Double-bond isomerization is a common side-reaction in olefin
metathesis, even for simple aliphatic olefins.[1–4] Isomerization
is particularly problematic for substrates containing allylic
amine, ether, or aromatic groups.[5–10] Comparative studies indi-
cate that the problem is most acute for the second-generation
Grubbs catalyst [Ru(=CHPh)Cl2(H2IMes)(PCy3)2] (GII; H2IMes=
N,N-bis(2,4,6-trimethylphenyl)-4,5-dihydroimidazol-2-ylidene),
Cy=cyclohexyl, as compared to the first-generation catalyst.[11]
Isomerization-active species have been proposed to originate
in contaminants introduced during catalyst synthesis,[12,13]
use,[14] or in intrinsic catalyst deactivation pathways.[15,16] Most
commonly cited in recent reports is the potential involvement
of two hydride complexes ([Ru2(m-C)Cl3(H)(H2IMes)2] (Ru-1) and
[RuCl(CO)H(H2IMes)(PCy3)] (Ru-2)], accessible by decomposition
of GII (Scheme 1).[17] Dinuclear Ru-1 was shown by Hong and
Consistent with the Bruneau report, we observed only minor
amounts of self-metathesis (SM) on treating 1 with GII at 408C
in toluene (1 mol% GII, 0.2m 1; Figure 1). The yield of homo-
coupled estragole (2) reached a maximum of 41%, then dimin-
ished, owing to competing isomerization, metathesis, or
Scheme 1. Complexes proposed to account for undesired isomerization
during metathesis by using GII.
Grubbs to form on thermolysis of the resting-state methyli-
dene complex (albeit slowly),[16] whereas several reports de-
scribe routes to hydridocarbonyl complexes of type Ru-2 from
GII.[18–20] Notably, Percy and co-workers observed Ru-2 during
ring-closing metathesis of 1,7-octadiene by GII,[21] and the
[a] C. S. Higman, L. Plais, Prof. D. E. Fogg
Department of Chemistry
and Centre for Catalysis Research & Innovation
University of Ottawa
Ottawa, ON, K1N 6N5 (Canada)
E-mail: dfogg@uottawa.ca
Figure 1. Rates of SM vs. isomerization on treating 1 with GII (0.2m 1,
1 mol% GII); based on replicate runs, Æ3%.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3548 – 3551 3548

CHEMCATCHEM
COMMUNICATIONS
www.chemcatchem.org
both.[28] Additional products arose from isomerization of 1 into
anethole (3), CM-isomerization (4), SM-isomerization (5), and
isomerization-SM (6), among other, less obvious pathways.[29]
Having established a baseline for isomerization during meta-
thesis, we turned to the question of whether Ru-1 and/or Ru-2
were sufficiently reactive to account for the levels of isomeriza-
tion observed. These experiments were performed under iden-
tical conditions of olefin concentration, temperature, and sol-
vent, using the maximum proportion of hydride complex at-
tainable assuming 100% transformation of the GII charge in
Figure 1 (i.e., 0.5 mol% for the dimer Ru-1 or 1 mol% for
Ru-2). Although clearly much higher than the proportion of
any hydride species observed during metathesis, these load-
ings represented an inarguable upper limit. Nonetheless, both
catalysts exhibited marginal isomerization activity (Figure 2).
Figure 3. Isomerization of 1 by using various Ru hydrides (0.2m 1, 1 mol%
Ru; yield of 3 at 0.5 h); based on replicate runs, Æ3%.
(e.g., an unsaturated olefin or aryl site within the substrate
itself). More fundamentally, the use of PPh3 allowed us to
probe the correlation between isomerization activity and
ligand donicity.
Given the low activity observed for Ru-2 at 408C (Figure 2),
we evaluated the activity of these five complexes at 808C.
Consistent with prior reports,[11] levels of isomerization were
considerably higher for H2IMes complex Ru-2 than its PCy3 an-
alogue Ru-2“ or, notably, IMes analogue Ru-2’. Dramatically
higher activity, however, was seen for the PPh3 derivatives
Ru-3 and Ru-4, suggesting that ligands of weaker donor ability
favored isomerization. Decomposed Ru species from which the
strong NHC and/or PCy3 donors had been scavenged may thus
have been responsible for the high levels of isomerization ac-
tivity shown in Figure 1.
Figure 2. Isomerization of 1 to 3 by using Ru-1 or Ru-2 (0.2m 1, 1 mol%
Ru); based on replicate runs, Æ3%.
For Ru-1, isomerization reached 6% in 5 h and only 14% after
24 h.[30] The low activity of Ru-1, compounded by its slow for-
mation (reported as ꢀ50% after 3 d at 558C in the absence of
substrate),[16] was strong evidence against its culpability in the
levels of isomerization shown in Figure 1. Similarly feeble activ-
ity (maximum 8% isomerization) was seen for Ru-2, despite
the higher catalyst loading.
The correlations above have important mechanistic implica-
tions. Whereas the higher activity of Ru-3 versus Ru-2 could re-
flect the greater lability of PPh3 than PCy3, comparison of Ru-3
and Ru-4 tells a different story. In experiments performed at
408C to maximize discrimination, the carbonyl complex Ru-3 is
observed to be nearly three times more active than Ru-4, de-
spite the fact that the p-acid ligand attenuates PPh3 lability.
Higher activity thus appears to be associated with a more elec-
tron-deficient Ru center (a factor that could contribute to the
impressive activity of a cationic Ru complex reported by Grot-
jahn and co-workers).[27,32] This, in turn, suggests that olefin
binding occurs during or before the rate-determining step. Al-
though early work on dihydride catalysts related to 3 also sug-
gested an associative pathway,[33] isomerization by catalysts of
type 2 has been presumed to proceed through a dissociative
mechanism.[21]
Given the implausibility of these catalysts as candidates for
the undesired isomerization reactions, we sought to clarify the
nature of the ligands associated with higher isomerization ac-
tivity. To this end, we screened a set of structurally related hy-
dridochloro complexes in the isomerization of 1. Whereas re-
views by the Schmidt and Krompiec groups examine the
impact of substrate functional groups on isomerization activi-
ty,[20,31] systematic comparisons of ligand effects are more limit-
ed. As noted above, however, multiple studies suggest that
second-generation [i.e., Ru–N-heterocyclic carbene (NHC)]
metathesis catalysts trigger more extensive isomerization side-
reactions than their first-generation analogues.[11]
To clarify this point, we evaluated rates of isomerization at
a range of concentrations of 1 (20 mm, 0.20m, or 1.0m) with
Ru-2 as catalyst. An approximately first-order dependence on
olefin concentration was observed, with the onset of satura-
tion kinetics near 1m. This indicated a mechanism that is asso-
ciative in olefin, despite the bulk associated with the PCy3 and
H2IMes groups already present on the catalyst. Added PCy3 in-
hibited the reaction, however, consistent with a subsequent
step involving the rate-determining loss of PCy3 (as also dem-
onstrated in prior studies).[21,33]
To examine this point with well-defined catalysts, we com-
pared the isomerization activity of the complex [RuCl(CO)H-
(PCy3)2] (Ru-2“) with that of the NHC derivatives Ru-2 and
[RuCl(CO)H(IMes)(PCy3)] (Ru-2’). Also examined were two com-
plexes with PPh3 [RuCl(CO)H(PPh3)3] (Ru-3) and [RuClH(PPh3)3]
(Ru-4); Figure 3]; this labile ligand was chosen as a proxy for
unknown weak donors potentially present during metathesis
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3548 – 3551 3549

CHEMCATCHEM
COMMUNICATIONS
www.chemcatchem.org
0.018 mmol, 1 mol%) and the reaction heated at 808C. Aliquots re-
moved at specific time intervals were quenched by exposure to air,
diluted with CH2Cl2, and analyzed by GC-FID.
Two lead candidates have been favored in the recent litera-
ture as probable triggers for unintended double-bond isomeri-
zation during olefin metathesis. We have provided strong evi-
dence against their involvement. Indeed, the correlation be-
tween weaker donor ligands and higher isomerization activity
tends to suggest that the isomerization-active species may
arise from more extensive decomposition of the metathesis
catalyst, in which the PCy3 and/or NHC ligands have been scav-
enged. One intriguing possibility is that the observable hydride
species play no active role, with the true culprits instead being
species that operate via a p-allyl mechanism. Finally, an asso-
ciative pathway has been uncovered in the isomerization
mechanism, even with rather bulky ligands on the metal. This
corroborates earlier work showing similar behavior for Ru-PPh3
complexes. If this pathway is general, isomerization may be
most problematic for metathesis reactions performed at high
olefin concentrations, including neat olefin.
Acknowledgements
This work was funded by The Natural Sciences and Engineering
Research Council of Canada (NSERC), who are also thanked for
a PGS-D award to C.S.H.
Keywords: catalysis
·
hydride
·
isomerization
·
olefin
metathesis · ruthenium
[1] K. J. Ivin, J. C. Mol, Olefin Metathesis and Metathesis Polymerization, Aca-
demic Press, New York, 1997.
[2] D. Bourgeois, A. Pancrazi, S. P. Nolan, J. Prunet, J. Organomet. Chem.
2002, 643–644, 247–252.
[3] S. E. Lehman, J. E. Schwendeman, P. M. O’Donnell, K. B. Wagener, Inorg.
Chim. Acta 2003, 345, 190–198.
[4] F. C. Courchay, J. C. Sworen, I. Ghiviriga, K. A. Abboud, K. B. Wagener, Or-
ganometallics 2006, 25, 6074–6086.
Experimental Section
[5] S. Krompiec, M. Krompiec, R. Penczek, H. Ignasiak, Coord. Chem. Rev.
2008, 252, 1819–1841.
General
Reactions were performed under an inert atmosphere by using
standard Schlenk and glove-box techniques. Toluene was dried by
using a solvent purification system (Anhydrous Engineering) and
stored over 4 ꢁ molecular sieves (Linde) under N2. 1 (Sigma–Al-
drich, 98%) was degassed by consecutive freeze–pump–thaw
cycles and filtered through neutral alumina under N2. Decane
(Sigma–Aldrich, anhydrous, 99%) and potassium trispyrazolylborate
(TCI Chemicals, 97%) were used as received. Complexes GII,[34]
Ru-1,[16] Ru-2,[35] Ru-2’,[35] Ru-2’’,[35] Ru-3,[36] and Ru-4[37] were pre-
pared according to literature methods. Yields in catalytic runs were
measured on an GC (Agilent 7890A) equipped with a flame ioniza-
tion detector (FID) and polysiloxane column (Agilent HP-5) at an
inlet split ratio of 10:1 and inlet temperature of 2508C. Retention
times for 1, 3, and 2 were confirmed by comparison with pure
samples. Yields were determined from integrated peak areas vs.
decane, and corrected for the response factors for decane and ana-
lyte. Calibration curves (peak areas vs. concentration) were con-
structed in the relevant concentration regime to account for the
dependence on detector response for all analytes.
´
[6] N. Kuznik, S. Krompiec, Coord. Chem. Rev. 2007, 251, 222–233.
[7] H. Bilel, N. Hamdi, F. Zagrouba, C. Fischmeister, C. Bruneau, RSC Adv.
2012, 2, 9584–9589.
[8] J. Moꢂse, S. Arseniyadis, J. Cossy, Org. Lett. 2007, 9, 1695–1698.
[9] S. H. Hong, D. P. Sanders, C. W. Lee, R. H. Grubbs, J. Am. Chem. Soc.
2005, 127, 17160–17161.
[10] K. Endo, M. B. Herbert, R. H. Grubbs, Organometallics 2013, 32, 5128–
5135.
[11] Selected examples: a) A. Fꢃrstner, O. R. Thiel, L. Ackermann, H.-J.
Schanz, S. P. Nolan, J. Org. Chem. 2000, 65, 2204–2207; b) F. C. Courchay,
J. C. Sworen, K. B. Wagener, Macromolecules 2003, 36, 8231–8239;
c) S. D. Kamau, P. Hodge, A. J. Hall, S. Dad, A. Ben-Haida, Polymer 2007,
48, 6808–6822; d) P. C. M. van Gerven, J. A. A. W. Elemans, J. W. Gerrit-
sen, S. Speller, R. J. M. Nolte, A. E. Rowan, Chem. Commun. 2005, 3535–
3537; e) A. V. Chuchuryukin, P. A. Chase, H. P. Dijkstra, B. Suijkerbuijk,
A. M. Mills, A. L. Spek, G. P. M. van Klink, G. van Koten, Adv. Synth. Catal.
2005, 347, 447–462; f) T. R. Hoye, M. A. Promo, Tetrahedron Lett. 1999,
40, 1429–1432; g) R. Ahuja, S. Kundu, A. S. Goldman, M. Brookhart, B. C.
Vicente, S. L. Scott, Chem. Commun. 2008, 253–255; h) S. D. Edwards, T.
Lewis, R. J. K. Taylor, Tetrahedron Lett. 1999, 40, 4267–4270.
[12] A. E. Sutton, B. A. Seigal, D. F. Finnegan, M. L. Snapper, J. Am. Chem. Soc.
2002, 124, 13390–13391.
[13] K. F. W. Hekking, F. L. van Delft, F. Rutjes, Tetrahedron 2003, 59, 6751–
6758.
Representative procedure for cross-metathesis
[14] T. Nicola, M. Brenner, K. Donsbach, P. Kreye, Org. Process Res. Dev. 2005,
9, 513–515.
[15] W. Janse van Rensburg, P. J. Steynberg, W. H. Meyer, M. M. Kirk, G. S.
Forman, J. Am. Chem. Soc. 2004, 126, 14332–14333.
[16] a) S. H. Hong, M. W. Day, R. H. Grubbs, J. Am. Chem. Soc. 2004, 126,
7414–7415; b) S. H. Hong, D. P. Sanders, C. W. Lee, R. H. Grubbs, J. Am.
Chem. Soc. 2007, 129, 7961–7968.
[17] Hong and Grubbs reported 76% conversion of allylbenzene following
treatment with 1.5 mol% Ru-1 for 24 h in refluxing CD2Cl2. See
Ref. [16]. For reports invoking the involvement of Ru-1 in isomerization,
see: a) S. Shahane, L. Toupet, C. Fischmeister, C. Bruneau, Eur. J. Inorg.
Chem. 2013, 54–60; b) G. B. Djigouꢄ, M. A. R. Meier, Appl. Catal. A 2009,
368, 158–162; c) R. Hemelaere, F. Carreaux, B. Carboni, J. Org. Chem.
2013, 78, 6786–6792. For isomerization reactions invoking Ru-2, see:
Ref. [8]; d) S. Hanessian, S. Giroux, A. Larsson, Org. Lett. 2006, 8, 5481–
5484; e) B. Schmidt, J. Org. Chem. 2004, 69, 7672–7687; f) M. B. Dinger,
J. C. Mol, Organometallics 2003, 22, 1089–1095; g) D. Banti, J. C. Mol, J.
Organomet. Chem. 2004, 689, 3113–3116; h) T. M. Trnka, J. P. Morgan,
M. S. Sanford, T. E. Wilhelm, M. Scholl, T.-L. Choi, S. Ding, M. W. Day, R. H.
Grubbs, J. Am. Chem. Soc. 2003, 125, 2546–2558.
A Schlenk tube was loaded with 1 (270 mg, 1.8 mmol), decane (in-
ternal standard; 259 mg, 1.8 mmol) and toluene (9.4 mL). An ali-
quot was removed and assessed by using GC-FID to establish the
starting ratio of 1/decane. Catalyst GII (15.5 mg, 0.018 mmol,
1 mol%) was then added. The Schlenk tube was removed immedi-
ately from the glovebox and attached to the Schlenk line. The cap
was replaced with a septum for sampling with a syringe. The
Schlenk tube was then submerged in a thermostatted oil bath at
408C. Aliquots were taken at specific time intervals, quenched with
potassium trispyrazolylborate (solution in THF, 10 mg/1 mL,
10 equiv), diluted with CH2Cl2, and analyzed by using GC-FID.
Representative procedure for isomerization
A Schlenk tube was charged with 1, decane, and toluene (quanti-
ties as above), after which an aliquot was removed to establish ini-
tial GC-FID integrations. Catalyst Ru-2 was then added (14.1 mg,
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3548 – 3551 3550

CHEMCATCHEM
COMMUNICATIONS
www.chemcatchem.org
[18] N. J. Beach, J. A. M. Lummiss, J. M. Bates, D. E. Fogg, Organometallics
2012, 31, 2349–2356.
[19] M. B. Dinger, J. C. Mol, Eur. J. Inorg. Chem. 2003, 2827–2833.
[20] B. Schmidt, Eur. J. Org. Chem. 2004, 1865–1880.
[21] I. W. Ashworth, I. H. Hillier, D. J. Nelson, J. M. Percy, M. A. Vincent, Eur. J.
Org. Chem. 2012, 5673–5677.
[30] For the Ru-1-catalyzed reaction after 24 h, a 5% discrepancy emerged
between the conversion of 1 (19%) and the yield of 3 (14%). Of the
four additional products observed, one minor constituent (2%) was
identified as CM product 2, suggesting that Ru-1 is not completely
metathesis-inactive or that an alkylidene is regenerated from the bridg-
ing carbido ligand.
[31] S. Krompiec, N. Kuznik, M. Krompiec, R. Penczek, J. Mrzigod, A. Torz, J.
Mol. Catal. A-Chem. 2006, 253, 132–146.
[22] M. Arisawa, Y. Terada, K. Takahashi, M. Nakagawa, A. Nishida, J. Org.
Chem. 2006, 71, 4255–4261.
[32] D. B. Grotjahn, C. R. Larsen, J. L. Gustafson, R. Nair, A. Sharma, J. Am.
Chem. Soc. 2007, 129, 9592–9593.
[23] J. R. Clark, J. R. Griffiths, S. T. Diver, J. Am. Chem. Soc. 2013, 135, 3327–
3330.
[33] S. Komiya, A. Yamamoto, J. Mol. Catal. 1979, 5, 279–292.
[34] P. Schwab, R. H. Grubbs, J. W. Ziller, J. Am. Chem. Soc. 1996, 118, 100–
110.
[35] N. J. Beach, U. L. Dharmasena, S. D. Drouin, D. E. Fogg, Adv. Synth. Catal.
2008, 350, 773–777.
[36] R. A. Schunn, E. R. Wonchoba, G. Wilkinson, Inorg. Synth. 1972, 13, 131–
134.
[37] N. Ahmad, J. J. Levison, S. D. Robinson, M. F. Uttley, Inorg. Synth. 1974,
15, 45–64.
[24] L. M. de Espinosa, M. A. R. Meier, Top. Organomet. Chem. 2012, 39, 1–
44.
[25] S. Manzini, D. J. Nelson, S. P. Nolan, ChemCatChem 2013, 5, 2848–2851.
[26] J. A. M. Lummiss, K. C. Oliveira, A. Pranckevicius, A. Santos, E. N. dos
Santos, D. E. Fogg, J. Am. Chem. Soc. 2012, 134, 18889–18891.
[27] C. R. Larsen, D. B. Grotjahn, J. Am. Chem. Soc. 2012, 134, 10357–10360.
[28] GII has been shown to effect metathesis of unhindered terminal olefins
with internal olefins related to 2. See, for example: a) J. Czaban, B. M.
Schertzer, K. Grela, Adv. Synth. Catal. 2013, 355, 1997–2006; b) A. Fꢃrst-
ner, O. R. Thiel, L. Ackermann, Org. Lett. 2001, 3, 449–451.
[29] For a useful representation of the plethora of isomerization products
accessible, see Ref. [28a].
Received: October 17, 2013
Published online on November 21, 2013
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3548 – 3551 3551