On the mechanism of the initiation reaction in Grubbs-Hoveyda complexes

Article abstract of DOI:10.1021/ja208967h

Grubbs-Hoveyda-type complexes with variable 4-R (complexes 1: 4-R = NEt₂, OiPr, H, F, NO₂) and 5-R substituents (complexes 2: 5-R = NEt₂, OiPr, Me, F, NO₂) at the 2-isopropoxy benzylidene ether ligand and with variable 4-R substituents (complexes 3: 4-R = H, NO₂) at the 2-methoxy benzylidene ether ligand were synthesized and the respective Ru(II/III) redox potentials (ranging from ΔE = +0.46 to +1.04 V), and UV-vis spectra recorded. The initiation kinetics of complexes 1-3 with the olefins diethyl diallyl malonate (DEDAM), butyl vinyl ether (BuVE), 1-hexene, styrene, and 3,3-dimethylbut-1-ene were investigated using UV-vis spectroscopy. Electron-withdrawing groups at the benzylidene ether ligands were found to increase the initiation rates, while electron-donating groups lead to slower precatalyst activation; accordingly with DEDAM, the complex 1(NO ₂) initiates almost 100 times faster than 1(NEt₂). The 4-R substituents (para to the benzylidene carbon) were found to have a stronger influence on physical and kinetic properties of complexes 1 and 2 than that of 5-R groups para to the ether oxygen. The DEDAM-induced initiation reactions of complexes 1 and 2 are classified as two-step reactions with an element of reversibility. The hyperbolic fit of the k_obs vs [DEDAM] plots is interpreted according to a dissociative mechanism (D). Kinetic studies employing BuVE showed that the initiation reactions simultaneously follow two different mechanistic pathways, since the k_obs vs [olefin] plots are best fitted to k_obs = k_D·k₄/k _-D·[olefin]/(1 + k₄/k_-D·[olefin]) + k_I·[olefin]. The k_I·[olefin] term dominates the initiation behavior of the sterically less demanding complexes 3 and was shown to correspond to an interchange mechanism with associative mode of activation (I_a), leading to very fast precatalyst activation at high olefin concentrations. Equilibrium and rate constants for the reactions of complexes 1-3 with the bulky PCy₃ were determined. In general, sterically demanding olefins (DEDAM, styrene) and Grubbs-Hoveyda type complexes 1 and 2 preferentially initiate according to the dissociative pathway; for the less bulky olefins (BuVE, 1-hexene) and complexes 1 and 2 both D and I _a are important. Activation parameters for BuVE reactions and complexes 1(NEt₂), 1(H), and 1(NO₂) were determined, and ΔS^? was found to be negative (ΔS ^? = -113 to -167 J·K^-1·mol ^-1) providing additional support for the I_a catalyst activation.

Full text of DOI:10.1021/ja208967h

Article
pubs.acs.org/JACS
On the Mechanism of the Initiation Reaction in Grubbs−Hoveyda
Complexes
Vasco Thiel, Marina Hendann, Klaus-Jurgen Wannowius, and Herbert Plenio*
̈
Organometallic Chemistry, FB Chemie, Petersenstr. 18, 64287 Darmstadt, TU Darmstadt, Germany
S
* Supporting Information
ABSTRACT: Grubbs−Hoveyda-type complexes with variable 4-R (complexes 1:
4-R = NEt2, OiPr, H, F, NO2) and 5-R substituents (complexes 2: 5-R = NEt2,
OiPr, Me, F, NO2) at the 2-isopropoxy benzylidene ether ligand and with variable
4-R substituents (complexes 3: 4-R = H, NO2) at the 2-methoxy benzylidene ether
ligand were synthesized and the respective Ru(II/III) redox potentials (ranging
from ΔE = +0.46 to +1.04 V), and UV−vis spectra recorded. The initiation kinetics
of complexes 1−3 with the olefins diethyl diallyl malonate (DEDAM), butyl vinyl
ether (BuVE), 1-hexene, styrene, and 3,3-dimethylbut-1-ene were investigated using UV−vis spectroscopy. Electron-withdrawing
groups at the benzylidene ether ligands were found to increase the initiation rates, while electron-donating groups lead to slower
precatalyst activation; accordingly with DEDAM, the complex 1(NO2) initiates almost 100 times faster than 1(NEt2). The 4-R
substituents (para to the benzylidene carbon) were found to have a stronger influence on physical and kinetic properties of
complexes 1 and 2 than that of 5-R groups para to the ether oxygen. The DEDAM-induced initiation reactions of complexes 1
and 2 are classified as two-step reactions with an element of reversibility. The hyperbolic fit of the kobs vs [DEDAM] plots is
interpreted according to a dissociative mechanism (D). Kinetic studies employing BuVE showed that the initiation reactions
simultaneously follow two different mechanistic pathways, since the kobs vs [olefin] plots are best fitted to kobs = kD·k4/
k−D·[olefin]/(1 + k4/k−D·[olefin]) + kI·[olefin]. The kI·[olefin] term dominates the initiation behavior of the sterically less
demanding complexes 3 and was shown to correspond to an interchange mechanism with associative mode of activation (Ia),
leading to very fast precatalyst activation at high olefin concentrations. Equilibrium and rate constants for the reactions of
complexes 1−3 with the bulky PCy3 were determined. In general, sterically demanding olefins (DEDAM, styrene) and Grubbs−
Hoveyda type complexes 1 and 2 preferentially initiate according to the dissociative pathway; for the less bulky olefins (BuVE,
1-hexene) and complexes 1 and 2 both D and Ia are important. Activation parameters for BuVE reactions and complexes
1(NEt2), 1(H), and 1(NO2) were determined, and ΔS‡ was found to be negative (ΔS‡ = −113 to −167 J·K−1·mol−1) providing
additional support for the Ia catalyst activation.
Scheme 1. Precatalysts for Olefin Metathesis Reactions
INTRODUCTION
■
Olefin metathesis has become a powerful tool in both organic1
and polymer synthesis.2 It is the preferred methodology for the
construction of carbon−carbon double bonds,3 for hetero-
cycles4 and in total synthesis,5 for green chemistry,6 for protein
modifications7 or in pharmaceutical chemistry.8 In addition to
tungsten- and molybdenum-based complexes,9 a large number
of different ruthenium complexes are known to catalyze such
transformations.10 Obviously, a detailed understanding of the
mechanism of olefin metathesis with such complexes is
essential. Systematic mechanistic studies on the initiation
reactions of Grubbs first- and second-generation (Scheme 1,
A) complexes established that the dissociation of the phosphine
ligand from the ruthenium center constitutes the rate-limiting
step.11 The knowledge of the initiation mechanism aided the
development of fast-initiating ruthenium complexes useful for
the synthesis of low-dispersity polymers (Scheme 1, B)2a,12 or
of slowly initiating complexes C, which are suitable for olefin
metathesis reactions of sterically demanding substrates.13 The
synthesis of rapidly initiating ruthenium precatalysts (Scheme 1,
D) by Piers and van der Eide14 enabled efficient initiation
reactions at low temperatures. At −50 °C ruthenacyclobutane
was shown to exist as a key intermediate in olefin metathesis
and was studied in detail by NMR spectroscopy.15 More
recently exhaustive low-temperature NMR experiments allowed
mapping of the key steps in ring-closing metathesis reactions
Received: September 23, 2011
Published: December 21, 2011
© 2011 American Chemical Society
1104
dx.doi.org/10.1021/ja208967h | J. Am. Chem.Soc. 2012, 134, 1104−1114

Journal of the American Chemical Society
Article
and obtaining an energy profile of the initiation reaction.25
The energy barriers along the RCM path were shown to be small,
consequently the catalyst utilized has an inherent high reactivity.
In order to change the activity of a precatalyst for given olefin
metathesis transformations, two approaches are useful: (a) to
change the nature of the active species in the catalytic cycle by
modifying the nature of the NHC ligand and (b) to adjust the
initiation rate of the precatalyst. The first approach led to the
development of numerous complexes with modified NHC
ligands or anionic ligands other than chloride.10b However,
approach (b), which relies primarily on changing the initiation rate
of the precatalyst by varying the nature of the leaving group, can
also be highly efficient, despite the fact that following the initia-
tion reaction, this approach probably generates the same active
species.16 Both methods led to the development of a large number
of new precatalysts,17 which initiate much more rapidly14,18 or
much more slowly (latent catalysts) than precatalyst A.19
Hoveyda et al. and Blechert et al. reported on stable
ruthenacarbenes derived from Grubbs-type complexes, in which
benzylidene and PCy3 are replaced by a bidentate benzylidene
ether ligand (E in Scheme 1).20 Such complexes do not contain
phosphines, are air and moisture resistant, and have been tested
successfully in numerous olefin metathesis reactions.21 Dorta
et al.22 and Grubbs et al.23 demonstrated that at very high
olefin concentrations, Grubbs−Hoveyda-type complexes can be
extremely efficient in ring closing metathesis (RCM).24 This
behavior is less easily accommodated with a purely dissociative
mechanism. It was shown later that the olefin is involved in the
rate-limiting step of precatalyst initiation, and consequently,
high olefin concentrations lead to rapid catalyst initiation and
excellent catalytic activity.25
A popular modification of Grubbs−Hoveyda complexes,
leading to faster initiating precatalysts is the introduction of a
5-NO2 group as a substituent at the benzylidene ether ligand.18a
It is believed that the diminished electron density at the ether
oxygen leads to a labile Ru−O interaction and consequently
faster initiation. The effect of substituents at the benzylidene
ether ligand in Grubbs−Hoveyda type complexes on the
catalytic performance was studied by Blechert et al. in a sys-
tematic manner.26 Complexes with different R groups in the 3-,
4-, or 5- position of the benzylidene ether ligand were tested in
several olefin metathesis reactions. The conversion after 4.5 h
and the time required to achieve 50% conversion were eval-
uated. The influence of substituents on substrate conversion
was found to correlate with the respective σ+- Hammett
parameter of R.
the Grubbs−Hoveyda complex, based on the Clar rule, indicated
that complexes with a delocalized cyclic structure “ruthenafurane”
provided (Scheme 2, F and H) provided slowly initiating complexes.
The absence of a delocalized structure in the five-membered ring
leads to rapidly initiating precatalysts (Scheme 2, G).
Recently a DFT study by Solans-Montfort et al. provided addi-
tional support for the relationship between the extent of delocalized
structure and catalyst reactivity.29 More specifically the catalytic
activity was found to be correlated with the bond length between
the carbene carbon and the α-carbon in the phenyl ring. Based on a
dissociative mechanism, no correlation between the energy of the
ruthenium−oxygen interaction and catalyst activity was observed.
A different view was provided by Vougioukalakis and Grubbs
who, based on negative activation entropies for the initiation
reaction, suggested an associative mechanism to be operative
for the initiation reaction of Grubbs−Hoveyda-type com-
plexes.30 In a preliminary communication we reported on
conclusive evidence for the participation of the incoming olefin
and the outgoing ether unit in the rate-limiting step,25a which
led us to propose an interchange mechanism. Additional
evidence had suggested that the nature of the rate-limiting step
may change at high olefin concentrations.25a More recently
Hillier, Percy et al. performed a combined experimental and
quantum chemical study.31 For the initiation reaction of
Grubbs−Hoveyda-type complexes involving ethene as the
olefinic substrate, an associative mechanism was proposed. In
the case of sterically more demanding olefins, such as ethyl
vinyl ether, an interchange mechanism provided the best fit for
experimental and calculated activation parameters. The
controversial results on the mechanism of the inititation
reaction in Grubbs−Hoveyda type complexes prompted us to
investigate this reaction in much more detail. Consequently, we
wish to report here on the results of these studies.
RESULTS AND DISCUSSION
■
Synthesis of Grubbs−Hoveyda-Type Complexes. Com-
plexes 1 and 2 (Scheme 3) with different groups in the 4- and
Scheme 3. Grubbs−Hoveyda-Type Complexes with
Different R4, R5 Groups Named according to 1(R4), 2(R5),
and 3(R4): 1(NEt2), 1(OiPr), 1(H), 1(F), 1(NO2), 2(NEt2),
2(OiPr), 2(Me), 2(F), 2(NO2), 3(H), and 3(NO2)
Barbasiewicz and Grela et al. explained why certain structural
modifications (Scheme 2) lead to faster or slower initiating
Scheme 2. Grubbs−Hoveyda-Type Complexes with
Naphthyl Groups
5-position of the benzylidene ether ligand were synthe-
sized according to literature procedures.18a,c,26,32 The 4- and 5-
positions were chosen to minimize steric effects on the initia-
tion reaction and focusing on electronic effects instead. The
respective substituents were selected to cover a large range of
electronic effects; several of the substituents are characterized
by very different meta- and para-Hammett parameters.33 The
steric bulk of complexes 1 and 2 with 2-isopropoxy groups was
modified by synthesizing closely related complex 3 with the
smaller 2-methoxy ether groups. The new complexes 2(NEt2)
and 3(NO2) were prepared according to a general procedure
precatalysts and what the origin of this behavior is.17b,18a,27
Specifically it was considered that there is an aromaticity-
controlled activity of ruthenium metathesis catalysts.28 A simple
topological analysis of several naphthalene-based analogues of
1105
dx.doi.org/10.1021/ja208967h | J. Am. Chem.Soc. 2012, 134, 1104−1114

Journal of the American Chemical Society
Article
recently reported.34 In contrast, to all of the other complexes
utilized in the present study, 3(NO2) decomposes in solution.
This complex is difficult to purify by chromatography and
consequently 3(NO2) is only obtained in ca. 25% yield.35
Electrochemistry. In order to obtain information on the
electron density at the ruthenium, the Ru(II/III) redox
potentials of the various complexes 1−3 were determined by
cyclic voltammetry (Table 1). For complexes 1 the redox
Table 1. Ruthenium (II/III) Redox Potentials for Complexes
a
1−3
Figure 2. UV−vis spectra of the 4-R substituted Grubbs−Hoveyda-
type complexes 1 in toluene.
complex
redox potential (V)
Ea − Ec (mV)
1(NEt2)
1(OiPr)
1(H)
0.460
0.705
0.837
0.859
1.038
0.908
0.678
0.823
0.821
1.001
0.886
1.09
78
66
63
66
69
72
72
72
69
69
77
>95
1(F)
1(NO2)
2(F)
2(NEt2)
2(OiPr)
2(Me)
2(NO2)
3(H)
3(NO2)
a
Ea − Ec is the difference between anodic and cathodic peak potential;
Figure 3. UV−vis spectra of the 5-R substituted Grubbs−Hoveyda-
type complexes 2 in toluene.
solvent: CH2Cl2, scan rate: 100 mVs−1, supp. electrolyte 0.1 M
NBu4PF6 ref. vs FcMe8 E1/2 = +0.010 V.
370−500 nm (Table 2), which are likely due to metal-to-ligand
charge transfer (MLCT) into the π* orbital of the RuCHR
potentials range from +0.460 to +1.038 V and for complexes 2
from 0.678 to 1.001 V.36 Obviously, the influence of the 4-R
group on the Ru(II/III) redox potential is stronger than that of
the 5-R group. Since the 4-R substituents are located para to
the benzylidene carbon and meta to the ether oxygen, it is likely
that the Ru(II/III) redox potentials are primarily influenced via
the benzylidene carbon and to a lesser degree via the ether
oxygen donor. This view is supported by the fact that there
is a reasonable correlation (R2 = 0.92) of the respective
σ -Hammett parameters relative the benzylidene carbon and
the redox potentials of complexes 1 and 2 (Figure 1).37 The
a
Table 2. UV−vis Data of Complexes 1−3
complex
λmax (nm) ε (L·mol−1·cm−1
)
λmax (nm) ε (L·mol−1·cm−1
)
1(NEt2)
1(OiPr)
1(H)
429
20090
12050
8630
535, 639
546, 635
554, 673
520, 642
581, 717
561, 665
585, 683
551, 663
517, 626
551, 635
522, 626
548, 702
250, 90
100, 50
120, 60
190, 100
210, 180
90, 40
388
376
1(F)
374
7660
1(NO2)
2(NEt2)
2(OiPr)
2(Me)
2(F)
419
11470
8080, 3180
6960
363, 450
398
120, 60
130, 90
160, 80
110, 70
150, 80
240, 150
384
9040
383
8540
2(NO2)
3(H)
370
7800
368
6600
3(NO2)
407
9537
a
Spectra recorded in toluene (10−4 M [Ru]) and 1.00 cm path length.
Data for weaker absorbances obtained from Gauss fits.
bond.39 This MLCT band provides an excellent spectroscopic
handle for monitoring the reactions of the respective Grubbs−
Hoveyda-type complexes. Changes in the 4-R group lead to
larger variations of λmax (374−429 nm) and ε (7658−20 092
L·mol−1·cm−1) than with the 5-R group (370−388 nm, 6964−
9043 L·mol−1·cm−1). The three 4-R groups which can enter
into direct resonance with the ruthenium display the highest
extinction coefficients. Two additional much weaker absorbances
(ε = 40−250 L·mol−1·cm−1) are observed in complexes 1−3
between 500−700 nm. Due their low extinction coefficients,
these bands probably originate from d−d transitions.
Figure 1. Plot of Hammett σ-parameter of the R4 (σpara+, para-) and R5
(σmeta) group in complexes 1 and 2 relative to the benzylidene carbon.
analogous plot using Hammett constants relative to the ether
oxygen is uncorrelated (R2 = 0.40) (Supporting Information,
SI-54).38 Replacing an 2-isopropoxy group in complexes 1 by a
2-methoxy substituent in complexes 3 has a small influence on
the Ru(II/III) redox potential.
UV−vis Spectra. The UV−vis spectra of complexes 1−3
(Figures 2 and 3) are characterized by one or two strong
absorbances (ε = 7000−20 000 L·mol−1·cm−1) between
Initiation Reaction of Complex 1(H) with Diethyl
Diallyl Malonate (DEDAM), Butyl Vinyl Ether (BuVE), and
1-Hexene and Discussion. In order to obtain information
1106
dx.doi.org/10.1021/ja208967h | J. Am. Chem.Soc. 2012, 134, 1104−1114

Journal of the American Chemical Society
Article
on the mechanism of precatalyst activation, the reactions of
complex 1(H) with the olefinic substrates DEDAM, BuVE, and
1-hexene were studied under pseudo first-order conditions
([olefin] ≫ [1(H)]) as a function of the substrate
concentration over 2 orders of magnitude for [DEDAM] =
0.01 − 2.0 mol·L−1, [BuVE], and [1-hexene] = 0.01−3.5
mol·L−1 at constant [1(H)] = 1.0 × 10−4 mol·L−1 in toluene
solvent at 303 K. The kinetic evaluation of precatalyst activation
is based on time-dependent UV−vis spectra at the maximum of
the peak at 376 nm.40 The absorbance vs time data for the
olefin initiation were fitted to a single exponential or a
hyperbolic function (see Supporting Information, S28−S29) to
obtain the respective kobs according to the kinetic model
reported previously.25a
enhanced in the presence of high olefin concentrations.22,25a,30
This accelerated product formation appears implausible, when
at the same time a nonproductive equilibrium is populated with
excess olefin. Furthermore there is no spectroscopic evidence
for a GH·olefin intermediate. This argument also applies to
pathway (a), which is less likely, since within the mixing time
no absorbance jump in the UV−vis spectra was observed,
which was carefully checked with twin-chamber cuvettes
(Supporting Information, Figure SI-42). An associative pathway
can result in a sudden shift of the UV−vis absorbances at the
beginning of the data collection, due to the rapid establishment
of the pre-equilibrium of olefin and the respective Grubbs−
Hoveyda complex. The kobs vs [DEDAM] data appear to reach
a ceiling rate at high olefin concentrations. It is likely that this
occurs when breaking of the ether oxygen to ruthenium bond
becomes rate limiting. At this point of the discussion, a disso-
ciative mechanism for the initiation of complexes 1(H) appears
to be the most likely choice. Later more arguments will be pre-
sented, which further support this view. We note that this expla-
nation is compatible with the mechanism established by Sanford,
Grubbs, et al. for Grubbs II type complexes.11a However, at this
stage of the discussion, there appears to be an inconsistency with
the mechanistic models proposed by Vougioukalakis and
Grubbs,30 Percy and Hillier et al.,31 and us.25a Obviously additional
detailed studies are required to resolve this.
v = − d[Ru]/dt = k ·[Ru]
(1)
obs
The dependencies of the observed pseudo first-order rate con-
stants kobs on the olefin concentrations are shown in Figure 4;41
for all complexes a weakly curved dependence is obtained.42
BuVE and 1-Hexene. The two additional olefins studied in
the precatalyst activation of complex 1(H) provide different kobs
vs [olefin] plots. Unlike the kobs vs [DEDAM] plot, the kobs vs
[BuVE] or [1-hexene] data for complex 1(H) in Figure 4
cannot be fitted satisfactorily with (eq 2), since the curve
segment at concentrations higher than ca. 0.3 M [BuVE] is
linear and depends on the olefin concentration. An excellent fit
of the data is obtained when the hyperbolic function (eq 2) is
extended with an additional linear term, with a first-order
dependence on the [olefin]:
Figure 4. Plot of kobs vs [olefin] (olefin = DEDAM, BuVE, 1-hexene)
for complex 1(H).
DEDAM. The kobs vs [DEDAM] data for complex 1(H) were
successfully fitted to a hyperbolic function (eq 2).
k
= m ·[olefin]/(1 + b·[olefin]) + m [olefin]
0 1
(3)
obs
In order to illustrate the individual contributions of the
linear and the hyperbolic terms for the kobs vs [BuVE] data,
both constituting terms are plotted in Figure 5. The red curve
k
= m ·[olefin]/(1 + b·[olefin])
0
(2)
obs
The parameter m0 stands for the initial slope and b is a measure
of the curvature. A previously employed linear function25a,31
also provides excellent fit parameters for the present data, but
the fit residuals show systematic deviations, which are apparent
when a large number of data points in the appropriate
concentration range are recorded.
The suitability of (eq 2) for the kobs vs [DEDAM] data suggest
to view this initiation reaction as a two-step reaction with an
element of reversibility.43 Three different mechanistic scenarios
have to be considered in this respect. The formal kinetics are out-
lined below (GH stands for the unactivated precatalyst, GH·olefin
for a complex with coordinated olefin and ruthenium
coordination number 6, and G represents a ruthenium complex
with the dissociated ether oxygen with CN 4, see also Scheme 5):
Figure 5. Fit of kobs vs [BuVE] plot according to (eq 3, red curve) for
complex 1(H) displayed in the range [BuVE] = 0−2.55 mol·L−1 (kobs
measured up to [BUVE] = 3.5 mol·L−1). The contribution of eq 2
(dashed line) and m1·[BuVE] (solid line) is shown.
(a) Associative scenario: GH + olefin ⇆ GH·olefin and
GH·olefin → products
(b) Dead-end setting involving a nonproductive equilibrium:
GH + olefin ⇆ GH·olefin and GH + olefin ⇆ products
(c) Dissociative scenario: GH ⇆ G and G + olefin →
products.
according to (eq 3) comprises the sum of the linear and the
hyperbolic curves, the dashed black line represents the con-
tribution of the hyperbolic term in (eq 3), and the solid line
stands for the linear term in (eq 3). The linear segment is
the major contributor to the red curve at olefin concentrations
The dead end scenario (b) is considered to be unlikely, since
it is well-known that olefin metathesis product formation is
1107
dx.doi.org/10.1021/ja208967h | J. Am. Chem.Soc. 2012, 134, 1104−1114

Journal of the American Chemical Society
Article
>0.5 M. Based on this extended model for the initiation of
complex 1(H) with BuVE or 1-hexene using a hyperbolic and a
linear term, the initiation reactions with BuVE and 1-hexene
cannot follow a simple dissociative mechanism. The use of a fit
function with two additive terms implies that the initiation
reaction must simultaneously follow two parallel mechanistic
pathways.
Initiation Reaction of Complexes 1−3 with BuVE and
1-Hexene. In order to learn whether the two potential modes
of precatalyst activation are general features of such complexes,
a much more detailed investigation of the initiation behavior,
employing 12 different complexes 1(NEt2), 1(OiPr), 1(H), 1(F),
and 1(NO2); 2(NEt2), 2(OiPr), 2(Me), 2(F), and 2(NO2); and
3(H) and 3(NO2), was done. The initiation reactions with BuVE
and 1-hexene were studied at 15 different olefin concentrations in
the range of 0.01−3.5 mol·L−1 under conditions similar to those
reported for complex 1(H). The respective plots of kobs vs
[BuVE] for the Grubbs−Hoveyda complexes 1 and 3 are shown
in Figure 6, those of complexes 2 in Figure 7, and the kobs vs [1-
hexene] in the Supporting Information, Figure SI-45. The same
basic features are found for all kobs vs [BuVE] plots of complexes
1 and 2 (Figures 6 and 7): (i) the kobs vs [BuVE] curve is almost
compared to 1(H) and a 5-fold increase with respect to
1(NO2). The most electron-withdrawing groups (4-F and 4-
NO2) lead to the most rapidly initiating complexes 1, while the
most strongly donating group (4-NEt2) is the slowest. In the
five-substituted complexes 2 (Figure 7), the substituents effects
on kobs are not according to the electron-donating capacity with
respect to the para-position. Instead the 5-OiPr substituent,
which is electron donating with respect to the para-position but
electron-withdrawing with respect to the meta-position, shows
faster initiation than 2(NO2). The same explanation applies to
the relatively fast initiation of 2(NEt2).
The kobs vs [1-hexene] plots for complexes 1(NO2), 1(NEt2),
and 1(H) (Supporting Information, Figure SI-45) are best
fitted with the hyperbolic + linear function (eq 3). Again with
the sterically less demanding precatalysts 3(H) and 3(NO2), a
linear fit is more suitable than (eq 3).
Initiation Reaction of Complexes 1−3 with DEDAM,
Styrene, and Neohexene. DEDAM. The precatalyst acti-
vation of the complexes 1(NEt2), 1(OiPr), 1(H), 1(F), and
1(NO2) and of 2(NEt2), 2(OiPr), 2(Me), 2(F), and 2(NO2),
and 3(H) and 3(NO2) with DEDAM was studied under the
conditions reported for 1(H). The dependencies of the
observed pseudofirst-order rate constants kobs on the olefin con-
centrations are shown in Figure 8 for the series of complexes 1
and 3 and in Figure 9 for complexes 2. In order to better
Figure 6. Plot of kobs vs [BuVE] for the four-substituted complexes 1
and 3.
Figure 8. Plot (log−log) of kobs vs [DEDAM] for the four-substituted
complexes 1 and 3.
Figure 7. Plot of kobs vs [BuVE] (display range 0−1.5 mol·L−1, data
range 0.01−3.5 mol·L−1) for the five-substituted complexes 2.
linear at [BuVE] > 0.3−0.5 mol·L−1, and there is a nearly linear
segment with a different slope at [BuVE] < 0.3 M. Consequently,
eq 3 provides excellent fits for the plots of all complexes 1 and 2.
(ii) On decreasing the bulk of the 2-alkoxy group from 2-
isopropoxy to 2-methoxy in complex 3, a pronounced increase in
the rate of the initiation reaction is observed, and the kobs vs
[olefin] plot is best fitted with kobs = m1[olefin].
Precatalyst activation for complexes 3 is much faster than for
complexes 1 and 2 (Figure 6) and at [BuVE] = 2.5 mol·L−1
complex 3(H) displays a 5-fold increase of kobs with respect to
1(H). Complex 3(NO2) shows a 12-fold increase of kobs
Figure 9. Plot of kobs vs [DEDAM] for the five-substituted complexes 2.
display the large differences in kobs for complexes 1 and 3, a
double logarithmic representation of the data was chosen. The
data in Figures 8 and 9 were fitted to kobs = m0·[DEDAM]/(1 + b·
[DEDAM]) (eq 2). Again the 2-methoxy-substituted com-
plexes 3 initiate much faster than all other 2-isopropoxy
complexes 1 and 2; at [DEDAM] = 1.0 mol·L−1, there is a
6-fold increase in kobs with 3(NO2) compared to 1(NO2). This
1108
dx.doi.org/10.1021/ja208967h | J. Am. Chem.Soc. 2012, 134, 1104−1114

Journal of the American Chemical Society
Article
rate acceleration is even more pronounced at higher [BuVE]
concentration when the curve for 1(NO2) is almost flat, while
that of 3(NO2) still increases. The range of initiation rates
covered by the various Grubbs−Hoveyda-type complexes is
impressive: 3(NO2) is ca. 1000 times faster than the slowest
complex 1(NEt2).
The respective kobs vs [DEDAM] plots of the four-substituted
complexes 1 (Figure 8) are ordered along their electron-donat-
ing capacity. 1(NEt2) is an extremely slowly initiating complex,
being almost 100 times slower than 1(NO2). For the series of
complexes 2 (Figure 9) with variable 5-R substituents the
spread in the initiation rates is much smaller than for complexes
1. The weaker influence of the 5-R substituents was observed
before for other physical parameters (redox potentials: Table 1;
UV−vis spectra: Table 2) and appears to be typical. There
seems to be a (weak) meta-effect, since 2(OiPr) and 5(NEt2)
are slightly faster than expected based on their respective para-
donation.
(3.5 mol·L−1) and 1.4 ·10−5 s−1 (3.0 mol·L−1); 3(H) 9.6 ·10−5 s−1
(3.5 mol·L−1) and 8.6 × 10−5 s−1 (3.0 mol·L−1). Complex
1(NO2) undergoes faster precatalyst activation than 1(H) and
3(H). This is suggestive of a primarily dissociative mode of
activation.
Reaction of Grubbs−Hoveyda Complexes 1−3 with
PCy3. This reaction leads to the substitution of the oxygen
donor by the phosphine ligand (Scheme 4) and might serve as
Scheme 4. Reaction of Grubbs−Hoveyda Complex with
PCy3
a simple model for the first steps of the initiation reaction. The
structure of the PCy3 addition product in solution was previously
confirmed by nuclear Overhauser enhancement spectroscopy
experiments.25a With a view to the bulk of PCy3, this reaction
could primarily occur via a dissociative mechanism. In case this is
true, the rates for the dissociation reaction of the ruthenium
complexes with PCy3 and olefins should be comparable. The
respective equilibrium constants KNMR with the complexes
Styrene. The initiation reactions of styrene are instructive,
since this olefin appears to be a borderline case concerning the
dominant mechanism of precatalyst activation. This is visible in
the double logarithmic representation of the kobs vs [styrene]
plot (Figure 10). Complexes 1(H) and 1(NO2) appear to
1
1−3 were determined via H NMR spectroscopy. The rate
constants kD were obtained via UV−vis experiments (Table 3
and Supporting Information, Figure SI-52).
Table 3. Kinetic and Thermodynamic Parameters for the
Reaction of Complexes 1−3 with PCy3 in Toluene at 303 K
k5·k−D = kD/
K
kD·103
K
·103
k5·KD·103
b
NMR
NMRa
b
(M−1
)
(s−1
13.1 0.8
42
)
(mol·L−1·s)
(L·mol−1·s−1
)
1(NEt2)
1(OiPr)
1(H)
119
285
92
0.110
0.146
0.317
0.091
0.170
0.188
68
6
5
8
71
Figure 10. Double logarithmic plot of kobs vs [styrene] for the four-
substituted complexes 1 and 3.
29.2 3.3
66 10
50.5
2
1(F)
720
926
97
147 14
1(NO2)
2(NEt2)
2(OiPr)
2(Me)
2(F)
158 32
171
40
5
follow a primarily dissociative initiation reaction. At high olefin
concentration the contribution of the m1·[styrene] term for the
two complexes increases significantly relative to the hyperbolic
term, and this leads to a significant bending of the fit curve.
Such bent curve segments at intermediate styrene concen-
trations of ca. 0.02 − 0.3 mol·L−1 denote roughly equal contri-
butions of the two different mechanistic pathways for
precatalyst initiation. For complex 3(H) the bent curve
segment is followed by another nearly linear section at styrene
concentrations in excess of 0.7 mol·L−1 and a dominant
m1·[styrene] term. Obviously, the main mode of precatalyst
activation appears to change going from low to high styrene
concentrations. For 3(NO2) the dissociative initiation mecha-
nism appears to be less important, and the linear m1·[styrene]
term is dominant at all styrene concentrations studied.
Neohexene. The precatalyst activation of 1(H), 1(NO2),
and 3(H) with a sterically very demanding olefin was also
tested. As expected, the initiation reactions involving neohexene
turn out to be extremely slow; roughly 1000 times slower than
for 1-hexene at the same olefin concentration. The kobs were
determined at two neohexene concentrations (3.0 and 3.5
mol·L−1) and found to be: 1(NO2) 2.1 ·10−4 s−1 (3.5 mol·L−1)
and 1.9 ·10−4 s−1 (3.0 mol·L−1); 1(H) 2.4 ·10−5 s−1
18
c
8
5
c
c
180
106
194
516
88
−
−
−
32
7
0.306
0.189
0.076
1.8
76
72
39
4
5
4
37 10
39 30
2(NO2)
3(H)
156
6
59 24
70 18
3(NO2)
928
587 43
0.63
a
Determined via the ratio of the integrals of the benzylidene protons
in the 1H NMR spectrum, the estimated error of KNMR is 10%.
b
Determined via UV−vis experiments fitted to (eq 4) with olefin =
c
PCy3. Spectral changes too small for evaluation.
The ordering of the equilibrium constants KNMR and of kD is
roughly according to the electron-donating abilities. Again the
interpretation of the substituent effects is not straightforward,
since the coordinating properties of ruthenium and of the ether
oxygen can be influenced via a combination of meta- and para-
effects of the 4- and 5-substituents. The rate constants kD
obtained from the PCy3 substitution reactions and from the
BuVE or the 1-hexene experiments for the respective complexes
1 and 2 show good agreement (see Table 4) for the data in the
concentration range ([PCy3] = 0.001−0.3 mol·L−1).44 For the
reactions of 3(H) and 3(NO2) with PCy3 a significant increase
1109
dx.doi.org/10.1021/ja208967h | J. Am. Chem.Soc. 2012, 134, 1104−1114

Journal of the American Chemical Society
Article
plots can either be fitted with a hyperbolic function based on
(eq 2), a hyperbolic + linear function kobs = m0·[olefin]/(1 +
b·[olefin]) + m1[olefin] (eq 3) or a linear function through the
origin. This has important implications for the mechanism of
such reactions, which need to be discussed in detail. An outline
of the potential associative, interchange, and dissociative
pathways as well as the terminology of the rate constants is
given in Scheme 5. Earlier in this manuscript, the hyperbolic
Table 4. Kinetic Parameters for the Initiation Reaction
of Complexes 1−3 with BuVE and 1-Hexene in Toluene at
303 K
kD ·103 (s−1
)
k4·KD·103 (L·mol−1·s−1
)
kI·103 (L·mol−1·s−1
)
BuVE
1(NEt2)
1(OiPr)
1(H)
17.5 1.3
109 18
107 18
18.0 0.7
25.1 1.4
24.5 0.8
45
22
4
2
43
9
1(F)
83 21
59 19
210 106
84 51
58
63
8
6
Scheme 5. Reaction Scheme and Potential Intermediates in
the Initiation Reaction of Grubbs−Hoveyda Complexes via
Associative, Interchange, and Dissociative Pathway
1(NO2)
2(NEt2)
2(OiPr)
2(Me)
2(F)
33
54
13
24
25
3
8
4
3
9
113 22
155 48
35 24
39.8 1.5
55.5 3.5
30
2
109 34
102 75
61 10
49.6
56
2
2(NO2)
3(H)
5
21 18
151
169
402
9
a
5
4
c
c
a
3(NO2)
1-Hexene
1(NEt2)
1(H)
−
−
b
−
3
0.05
−
25
8
105 76
21
43
102
4
9
1(NO2)
3(H)
110 27
210 100
c
c
a
−
−
−
4
c
c
a
3(NO2)
−
431 69
a
b
Obtained from linear fit kI·[BuVE] or kI·[1-hexene]. Obtained from
linear fit m1·[1-hexene], it is not clear whether this an D or an I rate.
c
The errors given are the errors of the fit procedure. Predominant Ia
mechanism, kD cannot be determined.
in the substitution rates is observed relative to those of com-
plexes 1(H) and 1(NO2). This observation is not compatible
with an exclusively dissociative mechanism for the ligand
substitution reaction for these complexes, while the equilibrium
constants KNMR are the same for the related complexes 1 and 3
Monitoring the Rate of Styrene Release. Two of the
styrenes (4-nitro-2-isopropoxystyrene and 4-NMe2-2-isopro-
poxystyrene) used for this study are characterized by UV−vis
spectra with absorption maxima significantly different from
those of the respective Grubbs−Hoveyda complexes 1(NO2)
and 1(NEt2). Consequently, the amount of liberated styrene
during olefin metathesis as well as the kinetics of this reaction
can be easily monitored independent from changes occurring at
the respective precatalysts. The absorbance vs time curves for
the DEDAM initiation of 1(NO2) were simultaneously
determined at 303 K by monitoring the decay of the LMCT
band at 419 nm and the increase in the styrene band at 360 nm.
The respective kobs vs [DEDAM] plots provide nearly identical
rate constants. It can thus be concluded that all steps following
this first step in the initiation reaction are much faster. This also
holds true when the same kinetic experiments are performed at
253 K. Despite the low temperatures, the initiation is still fast
enough at high olefin concentrations to be monitored
conveniently. Even at such low temperatures, the disappearance
of the precatalyst band and the increase in the styrene
absorbance follow almost the same rate, even though small
differences begin to show up. After the initiation reaction, the
concentration of liberated styrene remains constant throughout
the metathesis reaction. Consequently, there is no evidence for
a return of the benzylidene ligand to the ruthenium.25b
term was assigned to a dissociative pathway; based on this, eq 2
should be rewritten:
k
= k ·k /k ·[olefin]/(1 + k /k ·[olefin])
D 4 −D 4 −D
(4)
obs
Based on the applicability of eq 3, it was concluded that
precatalyst activation occurs according to two parallel
mechanistic pathways and that there are certain combinations
of olefin and ruthenium complexes, in which one of the two
possible modes of activation prevails. The respective kobs vs
[olefin] plots for the activation of precatalysts 1 and 2 with
sterically demanding substrates (DEDAM, styrene) are best
fitted with (eq 4). For the reactions of smaller substrates (1-
BuVE, 1-hexene) with small precatalysts 3, a kobs = m1·[olefin]
applies. For reactions of bulky olefins and small precatalysts 3
the same relationship appears to be the best approach in the
majority of reactions, while for small olefins and complexes 1 or
2 eq 4 + m1[olefin] applies.
This information serves as a starting point for the following
discussion on the interpretation of the m1·[olefin] term. The
first-order behavior with respect to olefin and the pronounced
influence of steric bulk at the olefin and the ruthenium
complexes on the rates suggest that this step can either
correspond to an associative or an interchange pathway for the
initiation reaction (Figure 4). Those two steps cannot be
differentiated kinetically with certainty for the reactions studied.
However, an associative mechanism for precatalyst activation in
Grubbs II complexes appears to be less likely, even though this
Discussion of the Mechanism of Precatalyst Activation
in Complexes 1−3. Depending on the nature of the
precatalysts and of the olefins, the respective kobs vs [olefin]
1110
dx.doi.org/10.1021/ja208967h | J. Am. Chem.Soc. 2012, 134, 1104−1114

Journal of the American Chemical Society
Article
has been considered for Grubbs I complexes.45 Our decision is
based on the following arguments: DFT studies by Solans-
Montfort et al. could not locate an energetically acceptable
transition structure according to an associative mechanism.29b
Related computational studies by Hillier and Percy et al. sug-
gest an interchange mechanism for olefins other than ethene,
rather than the formation of an intermediate with increased
coordination number.31 In our previous studies we had also
noted the contribution of both reaction partners in the initi-
ation process, which had led us to propose an interchange
mechanism with associative mode of activation. Furthermore an
associative pathway requires a six-coordinate intermediate.
There are only a few solid-state structures with six-coordinate
ruthenium in closely related (NHC)RuCl2(CHR) com-
plexes,46 which are formed with small substrates (pyridine) or
chelating ligands. An olefin bound to ruthenium in a potential
intermediate with coordination number 6 will be sterically more
demanding than for example a pyridine ligand, regardless of
whether the olefin adds trans45 or cis to the benzylidene.47
Another less appealing feature of an associative pathway is that
it does not seem to lead to an intermediate, which is able to
directly enter the catalytic cycle.48 Apart from the obligatory
ether oxygen dissociation, an additional olefin rearrangement
seems to be required.49
A different interpretation of the linear segment in the kobs vs
[olefin] plots for BuVE initiation also has to be discussed.
Linearity can also be observed with a purely dissociative
mechanism, far from the saturation kinetics in the initial linear
segment of the hyperbola controlled by k4·KD. However, we
consider this alternative as unlikely. As mentioned before, it is
well-known that reduced steric bulk at the catalyst favors an
interchange over a dissociative pathway. Along the same line,
reduced bulk at the olefinic substrate should also favor the
interchange over the dissociative pathway. Finally the activation
entropy, which was found to be strongly negative (see
Determination of Activation Parameters for BuVE), clearly
favors the I over the D pathway. Based on these arguments, the
linearity of the kobs vs [olefin] plots is unlikely to be the result
of a dissociative pathway.
presented show that electron-withdrawing groups render an
electron-deficient ruthenium, which is more willing to accept an
(electron rich) olefin in an interchange pathway than an
electron-rich ruthenium. Thus we note an interesting
dichotomy: Rapidly initiating precatalysts are characterized by
an electron-deficient ruthenium center, while the ruthenium in
the catalytically active species ought to be electron rich, as
claimed by Straub.51 According to Jensen et al., an electron-rich
ruthenium also increases the stability of the high oxidation state
ruthenacyclobutane.52
The substituent-dependent variations of the rate constants kI
and kD are much less pronounced in the series of complexes 2
than in complexes 1. This phenomenon was observed before.
Again there appears to be a meta-effect, since an increase in kI is
also observed for 2(OiPr), whose 5-OiPr substituent is electron
withdrawing with respect to the meta-position.
In order to provide additional evidence for our interpretation
of a simultaneous interchange and dissociative pathway for the
initiation reactions utilizing Grubbs−Hoveyda complexes, the
effect of a structural modification in the Grubbs−Hoveyda
complexes was tested. For a given substrate the interchange
pathway is expected to be more favorable with less bulky
ruthenium precatalysts. Thus replacing the 2-isopropoxy group
by the much smaller 2-methoxy group should favor the
interchange over the dissociative pathway. Only a few such
complexes have been reported in the literature, and
consequently, their reactivity was rarely studied.27d,53 The kobs
vs [BuVE] plots for the reactions of complexes 3(H) and
3(NO2) are characterized by a very modest curvature and are
almost linear over the whole concentration range. For such
complexes the parameters defining the hyperbola cannot be
determined reliably, and a kobs = kI·[BuVE] fit is more
appropriate. Consequently, for complexes 3 there is a strong
increase in kI(3(H)) = 0.151 L·mol−1·s−1 compared to
kI(1(H))= 0.0245 L·mol−1·s−1 and kI(3(NO2))= 0.402
L·mol−1·s−1 compared to kI(1(NO2))= 0.063 L·mol−1·s−1.
This accelerated interchange pathway is easily explained by
the reduced steric bulk of complexes 3 compared to that of
complexes 1.
Based on the results reported here, we propose an
interchange mechanism with associative mode of activation
(Ia) to account for the linear term m1·[olefin] and a dissociative
mechanism to account for the hyperbolic term. Such a
changeover from a dissociative mechanism to a mechanism
with associative character is not unusual and has been observed
for various platinum and ruthenium complexes.50 The validity
of the arguments presented above will be further elaborated in
the following evaluation of the rate constants obtained,
according to this mechanistic model.
Discussion of the Kinetic Parameter for BuVE and 1-
Hexene Initiation. The kinetic parameters for the BuVE
induced precatalyst activation of complexes 1−3 were
calculated by fitting the kobs vs [BuVE] data according to eq
4 + kI[BuVE] or eq 4 + kI[1-hexene] (Table 4). The
interchange rate kI is the slope of the linear segment at high
olefin concentrations and can be obtained from the fits with
good precision. The data in Table 4 show that the rate of the
interchange pathway is accelerated in complexes with electron-
withdrawing groups. This is reflected in a good correlation of
the Hammett parameters and kI (Supporting Information,
Figure SI-56). The established view within the framework of a
dissociative mechanism is that decreasing the electron density
at the ether oxygen accelerates Ru−O dissociation. The results
The kinetic parameters of 1-hexene are comparable to those
of BuVE, the most important information being the
pronounced increase in the interchange rate kI on replacing
the 2-isopropoxy by a 2-methoxy group in complexes 3.
Determination of Activation Parameters for BuVE. In
an attempt to provide additional evidence for an interchange
pathway, the respective activation parameters for the initiation
reactions of complexes 1(H), 1(NEt2), and 1(NO2) with BuVE
were determined (Table 5) by plotting the respective kI values
in an Eyring plot. The ΔH‡ (kI) value for 1(H) thus obtained is
close to the value determined previously.30 The three activation
entropies derived from kI are strongly negative (−113 to −168
J·K−1·mol−1), which clearly points to an interchange (or an
associative) initiation mechanism. The activation entropies for a
dissociative mechanism are not meaningful; the opening or
closing of the chelate ring in Grubbs−Hoveyda complexes does
not change the number of molecules.54
Discussion of the Kinetic Parameters for DEDAM and
Styrene Initiation. The kobs vs [DEDAM] plots can be fitted
according to eq 4, and the data obtained via this approach are
given in Table 6. The excellent fit parameters of the hyperbolic
fit suggest that initiation reactions of complexes 1 and 2 with
DEDAM operate via a dissociative mechanism. Given the
higher steric demand of DEDAM as compared to BuVE, this is
1111
dx.doi.org/10.1021/ja208967h | J. Am. Chem.Soc. 2012, 134, 1104−1114

Journal of the American Chemical Society
Article
Table 5. Activation Parameters for the Initiation of Complexes 1(H), 1(NEt2), and 1(NO2) with BuVE
ΔS‡ (k4·KD)
ΔS‡ (kI)
ΔH‡ (kD)
ΔH‡ (k4· KD)
ΔH‡ (kI)
ΔG‡
(k4· KD)
ΔG‡
(kI)
293K
[J·K−1·mol−1
]
[J·K−1·mol−1
]
[kJ·mol−1
]
[kJ·mol−1
]
[kJ·mol−1
]
[kJ·mol−1
]
[kJ·2m93oKl−1
]
1(H)
−86
−110 15
−140 37
8
−130 26
−170 81
−110 25
67 6.3
68 5.5
79 8.7
56
3
43 3.6
34 4.2
46 5.3
81
79
79
4
6
9
82
83 10
79
7
1(NEt2)
1(NO2)
47 3.7
38 4.4
9
Table 6. Kinetic Parameters for the Initiation Reaction
of Complexes 1−3 with DEDAM and Styrene in toluene at
303 K
Comparison of the Olefins. In order to evaluate the
substitution strength of the olefins utilized for the initiation
reactions, the ratio of the respective KD·k4 with the precatalysts
1(NEt2), 1(H) and 1(NO2) was calculated for BuVE and
DEDAM. For identical ruthenium precatalysts, the values of KD
should be identical, and consequently, the KD·k4 ratio should
reflect the ratio of k4 (Scheme 5) for the two olefins. This step
is important for the initiation reaction, but it is also closely
related to the association of the olefin in the catalytic cycle of
the actual olefin metathesis transformation. The analysis of k4
should thus provide information on the ability of the respective
olefin and the ruthenium complex to undergo olefin metathesis
reactions. KD·k4 is the initial slope of the kobs vs [olefin] plots
and was determined via eq 4 (Table 6). Alternatively, at very
low olefin concentrations the initiation reaction should be
primarily dissociative, and then the ratio of the kobs can provide the
same ratio of k4. According to the two procedures, the following
k4(BuVE)/k4(DEDAM) values were calculated (bracketed
values are the respective ratios of kobs at [olefin] = 0.01
mol·L−1): 1(NEt2) = 160 (177); 1(H) = 4 (6.5); and 1(NO2) =
1.3 (1.1). In general, BuVE attacks faster than DEDAM. For
1(H) the difference in olefin binding is significant, and for the
electron-rich and poorly olefin binding complex 1(NEt2) the
olefins are strongly differentiated, and the binding of more
electron-rich olefin is preferred. The rapidly initiating and
electron-deficient complex 1(NO2) hardly distinguishes
between BuVE and DEDAM complexation in the initiation
reaction. Since the propagating species in the limiting steps of
the catalytic cycle of RCM reactions is of the RuCH2 type,55
the olefin binding capacity of the ruthenium species does not
change. However, in cross metathesis reactions of electron-rich
and electron-deficient olefins with different substitu-
tion strengths and different electron-densities of the resulting
propagating ruthenium carbenes, the ratio of k4 may well
determine product distribution.
k4/k−D·103
k4·KD·103
kD·103
kI·103
a
a
a
(L·mol−1
)
(L·mol−1·s−1
)
(s−1
)
(L·mol−1·s−1
)
DEDAM
d
1(NEt2) 312 94
1(OiPr) 178 138
0.69 0.04
3.6 0.3
2.2 0.5
20 14
−
d
−
d
1(H)
207 29
551 165
806 134
10.5 0.2
40.3 3.2
132.2 7.2
7.7 0.3
51
6
−
d
1(F)
73 18
−
d
1(NO2)
164 20
−
d
2(NEt2) 218 55
2(OiPr) 423 180
35
8
−
d
16.9 1.6
5.9 0.3
40 14
52 27
80 15
93 12
−
d
2(Me)
2(F)
113 61
430 97
576 93
−
d
34.3 1.8
53.6 2.3
−
d
2(NO2)
3(H)
−
d
d
d
b
−
−
−
−
−
51
1
d
d
d
b
3(NO2)
Styrene
1(H)
−
416
8
d
630 90
420 80
16.5 0.7
26
3
−
d
1(NO2)
3(H)
78
4
184 29
−
c
c
c
c
65800 11000
387 44
5.9 0.4
21 0.6
d
d
d
b
3(NO2)
−
−
−
324
4
a
b
c
Fitted with eq 4. Slope of the linear fit. Eq 4 + kI[styrene].
Predominant D or Ia mechanism, the respective rate constant cannot
d
be determined. DEDAM contains two double bonds per molecule.
reasonable, but with a view to the BuVE and 1-hexene exper-
iments, a single activation pathway appears to be less likely.
Since at realistic [DEDAM] the kobs values for DEDAM
initiation have not reached the respective ceiling rates, it cannot
be excluded that there is some contribution from kI·[DEDAM].
However, the attempted fitting of the kobs vs [DEDAM] to
(eq 4) + kI·[DEDAM] does not converge to a well-defined set of
parameters. To obtain an estimate of the potential involvement
of an interchange mechanism in the DEDAM reactions, the
dissociation rates kD for complexes 1 and 2 obtained from the
PCy3 initiation were used for a simplified fit procedure. This is
based on the assumption that the kD based on PCy3 for complexes
1 and 2 are primarily dissociative and thus independent from
SUMMARY AND CONCLUSIONS
■
The mechanism of the initiation reaction of electronically and
sterically modified Grubbs−Hoveyda-type complexes with the
olefins DEDAM, BuVE, 1-hexene, styrene, and neohexene was
investigated. Based on the detailed analysis of the initiation
kinetics, this reaction was shown to simultaneously follow two
parallel pathways: a dissociative (D) mechanism and an
interchange mechanism with associative mode of activation
(Ia). The preference for one of the two possible modes of
precatalyst activation critically depends on the steric and
electronic properties of the respective ruthenium complexes
and the olefin employed for the metathesis reaction:
the nature of the substrate. This modified fit according to kobs
=
kD(PCy3)·k4/k−D·[olefin]/(1 + k4/k−D·[olefin]) + (kI·[olefin])
(eq 4) confirmed a largely dissociative pathway, except for a few
of the electron-deficient complexes 1 and 2, the interchange
contribution appears to be negligible.
There is a large spread in the dissociative rate constant kD for
complexes 1 which correlate well with the respective Hammett
constants relative to the benzylidene carbon (Supporting
Information, Figure SI-55). The initiation of complexes 3(H)
and 3(NO2) with DEDAM and with styrene is much faster than
that of complexes 1(H) and 1(NO2). Due to the modest
curvature of the kobs vs [DEDAM] plots with complexes 3, the
hyperbola according to (eq 4) is less well-defined and a linear
fit of the data more appropriate.
(i) Electron-withdrawing groups at the benzylidene ether
group of the Grubbs−Hoveyda complexes render an
electron-deficient ruthenium and lead to an acceleration
of both the interchange and the dissociative mechanism.
Substituents at the 4-position of the benzylidene ether
ligand (para to the benzylidene carbon) in complexes 1
exert a stronger effect on the initiation rates than
1112
dx.doi.org/10.1021/ja208967h | J. Am. Chem.Soc. 2012, 134, 1104−1114

Journal of the American Chemical Society
Article
(5) (a) Hoveyda, A. H.; Malcolmson, S. J.; Meek, S. J.; Zhugralin,
substituents in the 5-position (para to the ether oxygen)
in complexes 2.
A. R. Angew. Chem., Int. Ed. 2010, 49, 34−44. (b) Furstner, A. Chem.
̈
Commun. 2011, 47, 6505−6511.
(ii) Decreasing the steric bulk of the Grubbs−Hoveyda
complexes, by replacing the 2-isopropoxy substituent in
complexes 1 and 2 by a smaller 2-methoxy group in
complexes 3, leads to a pronounced acceleration of the
olefin-dependent Ia mechanism and consequently to a
strong increase in the rate of precatalyst activation at high
olefin concentrations.
(iii) Electron-rich and sterically less demanding olefins, such
as 1-hexene or BuVE, prefer the Ia activation. For the
more bulky or less electron-rich olefins, such as DEDAM
and styrene, the dissociative pathway is more important
with bulky complexes 1 and 2. With the less bulky
complexes 3, the olefin DEDAM undergoes precatalyst
activation preferentially according to Ia, while styrene is a
borderline case with comparable D and Ia contributions.
(6) Meier, M. A. R. Macromol. Chem. Phys. 2009, 210, 1073−1079.
(7) Lin, Y. A.; Chalker, J. M.; Davis, B. G. ChemBioChem 2009, 10,
959−969.
(8) (a) Farina, V.; Shu, C.; Zeng, X.; Wei, X.; Han, Z.; Yee, N. K.;
Senanayake, C. H. Org. Process Res. Dev. 2009, 13, 250−254.
(b) Morzycki, J. W. Steroids 2011, 76, 949−966.
(9) Schrock, R. R. Chem. Rev. 2009, 109, 3211−3226.
(10) (a) Samojłowicz, C.; Bieniek, M.; Grela, K. Chem. Rev. 2009,
109, 3708−3742. (b) Vougioukalakis, G. C.; Grubbs, R. H. Chem. Rev.
2010, 110, 1746−1787.
(11) (a) Sanford, M. S.; Love, J. A.; Grubbs, R. H. J. Am. Chem. Soc.
2001, 123, 6543−6554. (b) Love, J. A.; Sanford, M. S.; Day, M. W.;
Grubbs, R. H. J. Am. Chem. Soc. 2003, 125, 10103−10109.
(12) Louie, J.; Grubbs, R. H. Angew. Chem., Int. Ed. 2001, 40, 247−
249.
(13) (a) Sashuk, V.; Peeck, L. H.; Plenio, H. Chem.Eur. J. 2010, 16,
In order to achieve rapid precatalyst initiation an electron-
deficient and sterically accessible ruthenium is ideal, while the
olefins should be electron-rich and sterically unhindered to
promote both the dissociative as well as the interchange
activation pathway. The combination of steric and electronic
effects leads to Grubbs−Hoveyda-type complexes with widely
differing rates of precatalyst activation. Complex 3(NO2) with
4-NO2 and 2-methoxy substituents initiates ca. 1000 times
faster with DEDAM than precatalyst 1(NEt2) with 4-NEt2 and
2-isopropoxy substituents.
It remains to be shown in the future whether such a toolbox
of catalysts with tunable initiation rates can be used for more
efficient olefin metathesis transformations, by using precatalysts
whose initiation rate is tailored to the needs of a specific
substrate. Accepting the dual mechanism for precatalyst
activation also helps to understand why different combinations
of precatalysts and substrate lead to widely different reactivities
in olefin metathesis reactions, despite the fact that the same
catalytically active species is formed.
3983−3993. (b) Vorfalt, T.; Leuthaußer, S.; Plenio, H. Angew. Chem.,
̈
Int. Ed. 2009, 48, 5191−5194. (c) Peeck, L. H.; Plenio, H.
Organometallics 2010, 29, 2761−2768. (d) Bantreil, X.; Randall,
R. A. M.; Slawin, A. M. Z.; Nolan, S. P. Organometallics 2010, 29.
(e) Wolf, S.; Plenio, H. J. Organomet. Chem. 2010, 695, 2418−2422.
(14) Romero, P. E.; Piers, W. E.; McDonald, R. Angew. Chem. 2004,
116, 6287−6291.
(15) (a) Romero, P. E.; Piers, W. E. J. Am. Chem. Soc. 2005, 127,
5032−5033. (b) Wenzel, A. G.; Grubbs, R. H. J. Am. Chem. Soc. 2007,
128, 16048−16049. (c) Romero, P. E.; Piers, W. E. J. Am. Chem. Soc.
2007, 129, 1968−1704. (d) van der Eide, E. F.; Romero, P. E.; Piers,
W. E. J. Am. Chem. Soc. 2008, 130, 4485−4491. (e) Rowley, C. N.; van
der Eide, E. F.; Piers, W. E.; Woo, T. K. Organometallics 2008, 27,
6043−6045. (f) Wenzel, A. G.; Blake, G.; VanderVelde, D. G.; Grubbs,
R. H. J. Am. Chem. Soc. 2011, 133, 6429−6439. (g) vanderEide, E. F.;
Piers, W. E. Nat. Chem. 2010, 2, 571−576.
(16) Bieniek, M.; Michrowska, A.; Usanov, D. L.; Grela, K. Chem.
Eur. J. 2008, 14, 806−818.
(17) (a) Diesendruck, C. E.; Tzur, E.; Lemcoff, N. G. Eur. J. Inorg.
Chem. 2009, 4185−4203. (b) Tzur, E.; Szadkowska, A.; Ben-Asuly, A.;
Makal, A.; Goldberg, I.; Wozniak, K.; Grela, K.; Lemcoff, N. Chem.
́
Eur. J. 2010, 16, 8726−8737.
ASSOCIATED CONTENT
* Supporting Information
Experimental procedures for the synthesis of ligands and
ruthenium complexes and characterization of complexes, copies
of NMR and mass spectra, formal kinetics, Eyring plots, and
cyclic voltammograms. This material is available free of charge
via the Internet at http://pubs.acs.org.
(18) (a) Michrowska, A.; Bujok, R.; Harutyunyan, S.; Sashuk, V.;
Dolgonos, G.; Grela, K. J. Am. Chem. Soc. 2004, 126, 9318−9325.
(b) Wakamatsu, H.; Blechert, S. Angew. Chem., Int. Ed. 2002, 41,
2401−2403. (c) Gułajski, Ł.; Michrowska, A.; Bujok, R.; Grela, K.
J. Mol. Cat. A 2006, 254, 118−123. (d) Choi, T.-L.; Grubbs, R. H.
Angew. Chem., Int. Ed. 2003, 42, 1743−1746.
■
S
(19) (a) Monsaert, S.; Vila, A. L.; Drozdzak, R.; Voort, P. V. D.;
Verpoort, F. Chem. Soc. Rev. 2009, 38, 3360−3372. (b) Diesendruck,
C. E.; Iliashevsky, O.; Ben-Asuly, A.; Goldberg, I.; Lemcoff, N. G.
Macromol. Symp. 2010, 293, 33−38. (c) Ben-Asuly, A.; Tzur, E.;
Diesendruck, C. E.; Sigalov, M.; Goldberg, I.; Lemcoff, N. G.
Organometallics 2008, 47, 6422−6425. (d) Gstrein, X.; Burtscher,
D.; Szadkowska, A.; Barbasiewicz, M.; Stelzer, F.; Grela, K.; Slugovc, C.
J. Polym. Sci., Part A 2007, 45, 3494−3500. (e) Ung, T.; Hejl, A.;
Grubbs, R. H.; Schrodi, Y. Organometallics 2004, 23, 5399−5401.
(20) (a) Kingsbury, J. S.; Harrity, J. P. A; Bonitatebus, P. J.; Hoveyda,
A. H. J. Am. Chem. Soc. 1999, 121 (-), 791−799. (b) Garber, S. B.;
Kingsbury, J. S.; Gray, B. L.; Hoveyda, A. H. J. Am. Chem. Soc. 2000,
122, 8168−8179. (c) Gessler, S.; Randl, S.; Blechert, S. Tetrahedron
Lett. 2000, 41, 9973−9976.
AUTHOR INFORMATION
Corresponding Author
plenio@tu-darmstadt.de
■
ACKNOWLEDGMENTS
■
We thank the DFG for support through grant Pl 178/8-3. We
are grateful to Dr. X. Solans-Montfort for useful discussions and
the Umicore AG for a donation of M31 complex.
REFERENCES
■
(21) Hoveyda, A. H.; Gillingham, D. G.; Veldhuizen, J. J. V.; Kataoka,
O.; Garber, S. B.; Kingsbury, J. S.; Harrity, J. P. A Org. Biomol. Chem.
2004, 2, 8−23.
(1) Hoveyda, A. H.; Zhugralin, A. R. Nature 2007, 450, 243−251.
(2) (a) Leitgeb, A.; Wappel, J.; Slugovc, C. Polymer 2010, 51, 2927−
2946. (b) Slugovc, C. Macromol. Rapid Commun. 2004, 25, 1283−
1297. (c) Mutlu, H.; de Espinosa, L. M.; Meier, M. A. R. Chem. Soc.
Rev. 2011, 40, 1404−1445. (d) Opper, K. L.; Wagener, K. B. J. Polymer
Sci. A 2011, 49, 821−831.
(22) Gatti, M.; Vieille-Petit, L.; Luan, X.; Mariz, R.; Drinkel, E.;
Linden, A.; Dorta, R. J. Am. Chem. Soc. 2009, 131, 9498−9499.
(23) Kuhn, K. M.; Bourg, J.-B.; Chung, C. K.; Virgil, S. C.; Grubbs,
R. H. J. Am. Chem. Soc. 2009, 131, 5313−5320.
(3) Nolan, S. P.; Clavier, H. Chem. Soc. Rev. 2010, 39, 3305−3316.
(4) Krische, M. J. Proc. Nat. Acad. Sci.U.S.A. 2010, 107, 3279−3280.
(24) Monfette, S.; Fogg, D. E. Chem. Rev. 2009, 109, 3783−3816.
1113
dx.doi.org/10.1021/ja208967h | J. Am. Chem.Soc. 2012, 134, 1104−1114

Journal of the American Chemical Society
Article
(25) (a) Vorfalt, T.; Wannowius, K. J.; Plenio, H. Angew. Chem., Int.
Ed. 2010, 49, 5533−5536. (b) Vorfalt, T.; Wannowius, K.-J.; Thiel, V.;
Plenio, H. Chem.Eur. J. 2010, 16, 12312−12315.
(43) Wilkins, R. G. Kinetics and Mechanism of Reactions of Transition
Metal Complexes, 2nd ed.; VCH: Weinheim, Germany, 1991.
(44) Measurements at PCy3 concentrations >0.3 M are less reliable
due to increasingly important but unknown secondary reactions. The
available data are not sufficient to resolve this issue, and hyperbolic
data analysis was employed for all PCy3 reactions.
(26) Zaja, M.; Connon, S. J.; Dunne, A. M.; Rivard, M.; Buschmann,
N.; Jiricek, J.; Blechert, S. Tetrahedron 2003, 59, 6545−6558.
(27) (a) Grela, K.; Harutyunyan, S.; Michrowska, A. Angew. Chem.,
Int. Ed. 2002, 41, 4038−4040. (b) Bujok, R.; Bieniek, M.; Masnyk, M.;
Michrowska, A.; Sarosiek, A.; Stepowska, H.; Arlt, D.; Grela, K. J. Org.
Chem. 2004, 69, 6894−6896. (c) Bieniek, M.; Bujok, R.; Cabaj, M.;
Lugan, N.; Lavigne, G.; Arlt, D.; Grela, K. J. Am. Chem. Soc. 2006, 128,
13652−13653. (d) Barbasiewicz, M.; Bieniek, M.; Michrowska, A.;
Szadkowska, A.; Makal, A.; Wozniak, K.; Grela, K. Adv. Synth. Catal.
2007, 349, 193−203.
(45) Adlhart, C.; Chen, P. J. Am. Chem. Soc. 2004, 126, 3496−3510.
́
(46) Bieniek, M.; Samojłowicz, C.; Sashuk, V.; Bujok, R.; Sledz, P.;
́
Lugan, N. l.; Lavigne, G.; Arlt, D.; Grela, K. Organometallics 2011, 30,
4144−4158.
(47) (a) Anderson, D. R.; Hickstein, D. D.; O’Leary, D. J.; Grubbs,
R. H. J. Am. Chem. Soc. 2006, 128, 8386−8387. (b) Stewart, I. C.;
Benitez, D.; O’Leary, D. J.; Tkatchouk, E.; Day, M. W.; Goddard,
W. A.; Grubbs, R. H. J. Am. Chem. Soc. 2009, 131, 1931−1938.
(48) Straub, B. F. Adv. Synth. Catal. 2007, 349, 204−214.
(49) Correa, A.; Cavallo, L. J. Am. Chem. Soc. 2006, 128, 13352−
13353.
(50) (a) Alibrandi, G.; Bruno, G.; Lanza, S.; Minniti, D.; Romeo, R.;
Tobe, M. L. Inorg. Chem. 1987, 26, 185−190. (b) vanEldik, R.;
Hubbard, C. D. Adv. Phys. Org. Chem. 2006, 41, 1−78.
(51) Straub, B. F. Angew. Chem., Int. Ed. 2005, 117, 5974−5978.
(52) Occhipinti, G.; Bjørsvik, H.-R.; Jensen, V. R. J. Am. Chem. Soc.
2006, 128, 6952−6964.
(53) (a) Grela, K.; Kim, M. Eur. J. Org. Chem. 2003, 963−966.
(b) Krause, J. O.; Nuyken, O.; Buchmeiser, M. R. Chem.Eur. J. 2004,
10, 2029−2035.
(54) Atkins, P. W. Physikalische Chemie; VCH: Weinheim, Germany,
1990.
(55) This concerns the intermolecular reaction with olefin and does
not apply to the ring closing step with a bound alkylidene ligand.
However, the intermolecular reaction with an olefin can be expected to
be much faster than the corresponding intramolecular reaction.
(28) Barbasiewicz, M.; Szadkowska, A.; Makal, A.; Jarzembska, K.;
Wozniak, K.; Grela, K. Chem.Eur. J. 2008, 14, 9330−9337.
(29) (a) Solans-Monfort, X.; Pleixats, R.; Sodupe, M. Chem.Eur. J.
2010, 16, 7331−7343. (b) Nunez-Zarur, F.; Solans-Monfort, X.;
Rodriguez-Santiago, L.; Pleixats, R.; Sodupe, M. Chem.Eur. J. 2011,
17, 7506−7520.
(30) Vougioukalakis, G. C.; Grubbs, R. H. Chem.Eur. J. 2008, 14,
7545−7556.
(31) Ashworth, I. W.; Hillier, I. H.; Nelson, D. J.; Percy, J. M.;
Vincent, M. A. Chem. Commun. 2011, 5428−5430.
(32) Sußner, M.; Plenio, H. Angew. Chem., Int. Ed. 2005, 44, 6885−
̈
6888.
(33) Hansch, C.; Leo, A.; Taft, R. W. Chem. Rev. 1991, 91, 165−195.
(34) Doppiu, A.; Winde, R.; Worner, E.; Rivas-Nass, A.; Slugovc, C.;
̈
Lexer, C. Method for preparation of ruthenium-based metathesis
catalyst with chelating alkylidene ligand. Patent WO 2010/127829,
November 11, 2010.
(35) Complex 3(NO2) contains a small amount (<5%) of the
stilbene originating from the self-metathesis of 2-methoxy-4-nitro-
styrene. Quantitive removal of the stilbene was not possible by
chromatography due to the instability of this ruthenium complex.
(36) (a) Sußner, M.; Plenio, H. Chem. Commun. 2005, 5417−5419.
̈
(b) Leuthaußer, S.; Schmidts, V.; Thiele, C. M.; Plenio, H. Chem.
̈
Eur. J. 2008, 14, 5465−5481. (c) Collins, M. S.; Rosen, E. L.; Lynch,
V. M.; Bielawski, C. W. Organometallics 2010, 29, 3047−3053.
(d) Wolf, S.; Plenio, H. J. Organomet. Chem. 2009, 694, 1487−1492.
(37) Anslyn, E. V., Dougherty, D. A. Fitted according to the Yukawa-
Tsuno equation. Modern Physical Organic Chemistry; University
Science Books: Herndon, VA, 2006.
(38) The use of σ+/σ− Hammett parameter for the substituents in
the 4- and 5-position also leads to good correlations with the Ru(II/
III) redox potentials.
(39) (a) Hansen, S. M.; Rominger, F.; Metz, M.; Hofmann, P.
Chem.Eur. J. 1999, 5, 557−566. (b) Getty, K.; Delgado-Jaime,
M. U.; Kennepohl, P. J. Am. Chem. Soc. 2007, 129, 15774−15776.
(40) The need for a very large excess of olefin precluded kinetic
studies using NMR spectroscopy.
(41) Higher DEDAM concentrations (up to 4 M) were also studied,
but the kobs values at higher than 2.5 M cease to follow the hyperbolic
function. It is important to know in this context that a 4 M DEDAM
solution in toluene consists of 96% DEDAM and only 4% toluene.
(42) A 2 M DEDAM solution in toluene contains nearly 50% weight
of the olefin. It might be argued that the change in the nature of the
solution (dielectric constant) at very high olefin concentrations exerts
a significant influence on the initiation behavior of the ruthenium
complexes. We do not believe this effect to be of significance (at least
up to 50% weight of olefin) based on the following arguments: (i) The
higher dielectric constant of DEDAM compared to toluene might lead
to an increase in the initiation rates. Instead a slowing down is
observed; (ii) the kobs vs [olefin] plot for styrene and DEDAM are
similar, even though concerning the dielectric constant the two olefins
are very different; (iii) the strict linearity of kobs vs [olefin] observed
for several initiation reactions with these and other olefins is hardly
compatible with significant solvent effects; and (iv) different
complexes lead to very different kobs vs [olefin] plots, for example,
for related complexes 1 and 3 with the same olefin.
1114
dx.doi.org/10.1021/ja208967h | J. Am. Chem.Soc. 2012, 134, 1104−1114