Rhodium-carbon bond energies in Tp′Rh(CNneopentyl)(CH2X)H: Quantifying stabilization effects in M-C bonds

Article abstract of DOI:10.1021/ja400966y

A series of substituted methyl derivatives of the type Tp′ Rh(CNneopentyl)(CH₂X)H (CH₂X = CH₂C(i - O)CH₃, CH₂Ci - CCH₃, CH ₂O-t-Bu, CH₂CF₃, CH₂F, CHF ₂) was synthesized either by photolysis of Tp′Rh(CNneopentyl) (PhNCNneopentyl) in neat CH₃X or by exchange with the labile hydrocarbon in Tp′Rh(CNneopentyl)(n-pentyl)H or Tp′Rh(CNneopentyl) (CH₃)H. Only a single product was observed in each case. Clean reductive elimination was observed for all compounds in C₆D ₆. Structures of these complexes and their corresponding chlorinated derivatives have been characterized by NMR spectroscopy, elemental analysis, and X-ray crystallography. Relative Rh-C bond energies are calculated using previously established kinetic techniques, and two separate linear correlations are observed versus known C-H bond strengths, one for the parent hydrocarbons, and one for the substituted hydrocarbons. Both correlations have slopes of 1.4, and are separated vertically by 7.5 kcal mol^-1 (-CH₂X above -C_xH_y). In addition, it is now clear that preferences for linear vs branched olefin insertion products in substituted derivatives can be predicted on the basis of the strengths of the β-C-H bonds. The DFT calculations of the metal-carbon bond strengths in these Rh-CH₂X derivatives with α-substitution show a trend that is in good agreement with the experimental results.

Full text of DOI:10.1021/ja400966y

Article
pubs.acs.org/JACS
Rhodium−Carbon Bond Energies in Tp′Rh(CNneopentyl)(CH2X)H:
Quantifying Stabilization Effects in M−C Bonds
Yunzhe Jiao, Meagan E. Evans, James Morris, William W. Brennessel, and William D. Jones*
Department of Chemistry, University of Rochester, Rochester, New York 14627, United States
S
* Supporting Information
ABSTRACT: A series of substituted methyl derivatives of the type Tp′Rh(CNneopentyl)(CH2X)H
(CH2X = CH2C(O)CH3, CH2CCCH3, CH2O-t-Bu, CH2CF3, CH2F, CHF2) was synthesized
either by photolysis of Tp′Rh(CNneopentyl)(PhNCNneopentyl) in neat CH3X or by exchange with
the labile hydrocarbon in Tp′Rh(CNneopentyl)(n-pentyl)H or Tp′Rh(CNneopentyl)(CH3)H. Only
a single product was observed in each case. Clean reductive elimination was observed for all
compounds in C6D6. Structures of these complexes and their corresponding chlorinated derivatives
have been characterized by NMR spectroscopy, elemental analysis, and X-ray crystallography. Relative Rh−C bond energies are
calculated using previously established kinetic techniques, and two separate linear correlations are observed versus known C−H
bond strengths, one for the parent hydrocarbons, and one for the substituted hydrocarbons. Both correlations have slopes of 1.4,
and are separated vertically by 7.5 kcal mol−1 (−CH2X above −CxHy). In addition, it is now clear that preferences for linear vs
branched olefin insertion products in substituted derivatives can be predicted on the basis of the strengths of the β-C−H bonds.
The DFT calculations of the metal−carbon bond strengths in these Rh−CH2X derivatives with α-substitution show a trend that
is in good agreement with the experimental results.
INTRODUCTION
■
One of the most important properties of homogeneous
organometallic catalysts is their ability to control regioselectiv-
ity in alkene insertion reactions. For example, the hydro-
formylation of olefins by cobalt carbonyl shows preferential
formation of linear over branched aldehydes (∼4:1), and use of
a bulkier rhodium catalyst can improve on this selectivity
(∼16:1).1 Likewise, hydrocyanation of butadiene can be
tailored to give as high as 98% linear addition product using
a nickel diphosphine catalyst.2 In these and other reactions, the
kinetic and/or thermodynamic preferences for forming
branched vs linear metal−alkyl bonds are critical to controlling
the overall reaction selectivity. This in turn requires an
understanding of relative metal−carbon bond strengths in a
variety of substituted alkyl derivatives.
primary metal−alkyl bond is stronger than a secondary metal−
alkyl bond.12
There have been a number of reports, however, where this
thermodynamic selectivity can be reversed. Reger also reported
that heating the β-cyanoethyl derivative of the iron complex in
eq 1 to 95 °C led to the conversion to the branched α-
cyanoethyl complex (eq 3).13 Likewise, the β-cyanoethyl
derivative of the palladium complex in eq 2 rearranges
quantitatively to the α-cyanoethyl isomer at 120 °C (eq 4).
This change in selectivity for the branched isomer was
attributed to the electronic effect of the cyano group on the
carbon attached to the metal overcoming the steric effects
disfavoring formation of a branched isomer. Harvey has
reported, however, that the preference for primary over
secondary alkyl derivatives is also primarily an electronic effect,
as a result of the fact that primary carbanions are favored over
secondary carbanions.14
For the parent unsubstituted alkyl derivatives, there is strong
evidence that the linear, primary metal−alkyl bond is preferred
thermodynamically over the branched, secondary metal−alkyl
bond. It was observed in the 1970s that hydrozirconation of
isomers of octene led to the sole formation of the n-
octylzirconium product.3 Kochi observed that a tert-butyl
gold(III) complex isomerized spontaneously to the isobutyl
isomer at 25 °C with a half-life of 17 h.4 Reger observed in
1980 that the secondary alkyl complex CpFe(CO)(PPh3)(sec-
butyl) rearranged quantitatively to the linear isomer upon
heating to 63 °C (eq 1).5 A similar preference for the linear
over the branched isomers was seen in a series of palladium−
alkyl complexes at 75 °C (eq 2),6 and similar results were
observed with platinum complexes at 130 °C.7 These and other
observations8−10 of similar preferences, along with the
corresponding reactivities of metal−alkyl carbonyls to undergo
CO insertion,11 have led to the general conclusion that a
Received: January 28, 2013
Published: April 23, 2013
© 2013 American Chemical Society
6994
dx.doi.org/10.1021/ja400966y | J. Am. Chem. Soc. 2013, 135, 6994−7004

Journal of the American Chemical Society
Article
We also hypothesized that these α-substituent effects, as well
as in related α-mesityl and methallyl derivatives, led to a
correlation of Rh−C bond strengths that was distinct from the
correlation seen with unsubstituted hydrocarbons.22 Figure 1
Another example of where the linear/branched product
distribution is reversed is oftentimes seen in styrene insertion
reactions. For example, Holland reported that insertion of
styrene into an iron-nacnac hydride initially gave the linear β-
phenethyl product that rearranged to give the thermodynami-
cally preferred α-phenethyl isomer (eq 5).15 Similarly,
Figure 1. Comparison of rhodium−carbon and hydrogen−carbon
bond strengths. The slopes of the upper and lower lines are 2.13 and
1.59, respectively. The data shown in four open red squares are for the
primary C−H bond in propionitrile, butyronitrile, valeronitrile, and
capronitrile, top to bottom, using DFT calculated C−H bond
strengths for these four substrates, but experimental bond strengths
for all other C−H bond strengths. Reprinted with permission from ref
22. Copyright 2009 American Chemical Society.
Consiglio reported that insertion of styrene into a palla-
dium−acyl bond led to the branched insertion product rather
than the linear insertion product (eq 6).16 Related observations
have been made for other styrene insertions to give branched
isomers.17,18
shows the proposed correlations for the two classes of
substrates. The lower line contains only sp3 and sp2
hydrocarbons. The upper line includes only substituted methyl
derivatives of the type Tp′Rh(CNneopentyl)(CH2X)H. Only
three data points were used to formulate this second
correlation, and it was deemed necessary to collect additional
data to support the hypothesis that these substituted derivatives
followed a separate trend. In this contribution, additional data
are obtained for the compounds where X = CH2C(O)CH3,
CH2CCCH3, CH2O-t-Bu, CH2CF3, CH2F, and CHF2. These
data are then combined with recently obtained data for terminal
alkyne sp C−H activation23 and pentafluoroarene activation24
to provide a unified picture of the factors that control metal−
carbon bond strengths. Some surprises were uncovered.
We have been interested in determining quantitatively the
effect of substituents on α-substituted alkyl groups during C−H
activation reactions of substituted aliphatic hydrocarbons using
the [Tp′Rh(CNneopentyl)] fragment, for which extensive
thermodynamic data are available for the parent hydrocarbon
activations.19 We have discovered that 1-chloroalkanes show
strong selectivity for the activation of a methyl group at the
remote end of the chain, and that chloromethane undergoes
C−H but not C−Cl oxidative addition (eq 7).20 The product,
RESULTS
■
Synthesis and Characterization of Tp′Rh-
(CNneopentyl)(CH2 X)H. Irradiation of Tp′Rh-
(CNneopentyl)(PhNCNneopentyl) (1) has proven to be an
efficient way to form the coordinatively unsaturated fragment
[Tp′Rh(CNneopentyl)] which then reacts with most hydro-
carbons to give clean C−H oxidative addition products.25
Kinetic products are obtained, and the observed stabilities of
the products toward reductive elimination at room temperature
range from minutes (cyclohexane) to hours (methane) to
months (benzene).19 As mentioned above, only a few
substrates of the type CH3X have been examined where X
represents a functional group. Scheme 1 shows three additional
substrates that have been examined.
Tp′Rh(CNneopentyl)(CH2Cl)H is reluctant to undergo
reductive elimination, which was attributed to the electron
withdrawing nature of the α-substituent. Studies were also done
on a series of alkyl nitriles, where it was also found that α-cyano
substitution gave products that would only undergo reductive
elimination at elevated temperatures (eq 8).21 This effect was
attributed to a strengthening of the metal−carbon bond due to
substitution by the electron−withdrawing cyano group.
Photolysis of 1 in neat acetone resulted in a color change of
the solution from bright yellow to pale yellow after only 3 min
1
at −20 °C. H NMR spectroscopic analysis of the product
shows clean formation of Tp′Rh(CNneopentyl)(CH2C(
O)CH3)H (2) with a hydride resonance at δ −14.778 (d, JRh−H
6995
dx.doi.org/10.1021/ja400966y | J. Am. Chem. Soc. 2013, 135, 6994−7004

Journal of the American Chemical Society
Article
Scheme 1. Products from the Photolysis of 1 in Various
Substrates
Figure 2. Thermal ellipsoid drawing of Tp′Rh(CNneopentyl)(CH2C-
(O)CH3)Cl (2-Cl). Hydrogen atoms have been omitted for clarity.
Ellipsoids are shown at the 50% level.
NMR spectrum of 4 displays a hydride doublet at δ −14.403
(JRh−H = 25.2 Hz) and a methylene doublet at δ 4.888 (JRh−H
=
16.1 Hz). The bulky tert-butyl group and proximity of the
methyl to the electron-rich oxygen atom ensures the exclusive
C−H activation at the methyl group. Treatment of 4 with CCl4
produces Tp′Rh(CNneopentyl)(CH2O-t-Bu)Cl (4-Cl), which
was fully characterized by NMR spectroscopy, elemental
analysis, and X-ray structure determination (Figure 3). The
= 19.8 Hz). This small Rh−H coupling constant is typical in α-
substituted alkyl hydride complexes (cf. J = 20 Hz for
Tp′Rh(CNneopentyl)(CH2CN)H,21 vs J = 24−25 Hz for
unsubstituted alkyl and aryl hydride species19a). The diaster-
eotopic methylene resonances of RhCH2C(O)CH3 were
shifted downfield relative to free acetone and appear at δ 2.871
and 2.993. Similarly, the resonance for the ketone methyl peak
was shifted downfield to δ 2.469. Treatment of 2 with CCl4
gave Tp′Rh(CNneopentyl)(CH2C(O)CH3)Cl (2-Cl),
which was fully characterized by NMR spectroscopy, elemental
analysis, and X-ray structure determination (Figure 2). The
structure shows an octahedral geometry with a Rh1−C7
distance of 2.109 (3) Å, typical for a Rh−C(sp3) bond (cf.
d(Rh−C) = 2.105 (4) Å in Tp′Rh(CNneopentyl)(n-pentyl)-
Cl19a).
Photolysis of 1 in 2-butyne at −20 °C led to the rapid
formation of Tp′Rh(CNneopentyl)(CH2CCCH3)H (3)
with a hydride resonance at δ −14.567 (d, JRh−H = 21.9 Hz)
Figure 3. Thermal ellipsoid drawing of Tp′Rh(CNneopentyl)(CH2O-
t-Bu)Cl (4-Cl). Hydrogen atoms have been omitted for clarity.
Ellipsoids are shown at the 50% probability level.
1
in the H NMR spectrum. In contrast to the activation of
acetone, decomposition was observed if the irradiation was
continued longer than 3 minutes or at a temperature above −20
°C. In addition, no evidence of an η2-coordinated intermediate
was observed either during irradiation or upon heating of 3 in
benzene (vide infra), which suggests C−H activation has a
lower barrier than coordination to the triple bond and that the
C−H activation product is thermodynamically more stable than
the π-complex.
structure shows an octahedral geometry with a Rh1−C22
distance of 2.042 (7) Å, somewhat short for a Rh−C(sp3) bond
and more similar to a Rh−C(sp2) distance (c.f. dRh−C = 2.021
(6) Å in Tp′Rh(CNneopentyl)(CHCHCMe3)Cl,19b 2.063
(5) Å in Tp′Rh(CNneopentyl)(C6F5)Cl,24 and 2.050 (9) in
Tp′Rh(CNneopentyl)(c-propyl)Cl.26
Photolysis of 1 in 2-methoxy-2-methylpropane required a
longer irradiation time for completion (30 min) and produced
Photolysis of 1 in/with fluoroalkanes showed mainly
decomposition products. Therefore Tp′Rh(CNneopentyl)(n-
1
only one product, Tp′Rh(CNR)(CH2O-t-Bu)H (4). The H
6996
dx.doi.org/10.1021/ja400966y | J. Am. Chem. Soc. 2013, 135, 6994−7004

Journal of the American Chemical Society
Article
pentyl)H, Tp′Rh(CNneopentyl)(c-C6D11)D, or Tp′Rh-
(CNneopentyl)(CH3)H were prepared in situ by irradiation
of 1 in neat n-pentane or cyclohexane-d12 or by the reaction of
Tp′Rh(CNneopentyl)(CH3)Cl with Cp2ZrH2.26 The labile
complexes Tp′Rh(CNneopentyl)(R)H (R = n-pentyl, c-C6D11
or CH3) serve as thermal precursors to the reactive
intermediate, [Tp′Rh(CNneopentyl)], which then inserts into
the C−H bonds of fluoroalkanes (Scheme 2). The mild
exchange reactions usually take overnight to several days to go
to completion.
24.7 Hz, JF−H = 8.6 Hz). The fluorine resonance for 6 was
observed in the 19F NMR spectrum as a triplet of doublet of
doublets with one Rh−F and two H−F couplings. The triplet
H−F coupling was due to the two methylene hydrogen atoms,
with the other H−F coupling being due to the hydride (JH−F
8.6 Hz).
=
An unexpected hydride resonance at δ −14.14 with JRh−H
=
25.2 Hz was observed in both of the above reactions with yields
of 28% and 36%, respectively (Figure 4). The 19F NMR
Scheme 2. Products from Exchange Reactions with
Substrates
1
Figure 4. H NMR of hydride region for reductive elimination of a
mixture of Tp′Rh(CNneopentyl)(CH2F)H (6) and Tp′Rh(CNR)-
(CH2OCH3)H (7) in C6D6 at 66.9 °C. Small quantities of o-, m-, and
p-carbodiimide activation products can also be identified.
spectrum shows no corresponding fluorine resonance for this
byproduct. This hydrido compound was found to produce
Tp′Rh(CNneopentyl)(C6D5)D by reductive elimination in
C6D6 cleanly. The half−life of 2.0 h at 66.9 °C suggested
that it was generated from activation of an impurity in the
CH3F. These properties of the unknown product are similar to
those of Tp′Rh(CNR)(CH2O-t-Bu)H, in which the hydride
resonance appeared at a similar chemical shift (δ −14.403) with
the same coupling constant. Further examination of the gas
used in the experiment (1H NMR and GC−MS) allowed the
impurity to be identified as dimethyl ether, present at ∼16% of
the commercial CH3F sample. Therefore this hydride product
was assigned as Tp′Rh(CNR)(CH2OCH3)H (7) from
activation of dimethyl ether.
Exchange of Tp′Rh(CNneopentyl)(CH3)H with CH3CF3
led to the clean formation of Tp′Rh(CNneopentyl)(CH2CF3)
H (5). 5 was also prepared via reaction with Tp′Rh-
(CNneopentyl)(n-pentyl)H under a pressure of CH3CF3, but
was accompanied by trace amounts of o-, m-, and p-
carbodiimide activation products (<5%).20b,25 No C−F
activation product was detected by NMR spectroscopy. The
hydride resonance for 5 appeared as a doublet at δ −14.344
(JRh−H = 21.3 Hz). The hydride doublet also indicates the
absence of coupling between the hydride and any fluorine(s).
The 19F NMR spectrum shows a triplet due to coupling with
the two neighboring methylene hydrogen atoms. Due to the
presence of many couplings, 1H NMR analysis of the methylene
resonance was ambiguous, but a quintet was seen in the
13C{1H} NMR spectrum at δ 8.31 (JRh−C = JF−C = 26.9 Hz).
Moreover, an HSQC spectrum indicates that the hydrogen
resonances associated with this methylene group were obscured
Irradiation of Tp′Rh(CNneopentyl)(carbodiimide) in C6D12
followed by addition of CH2F2 led to formation of a mixture of
Tp′Rh(CNneopentyl)(CHF2)H (8) and the products of
carbodiimide activation (31:69). The hydride resonance
appeared at δ −14.095 as a doublet of doublets, suggesting
coupling with only one of the two diastereotopic fluorine
atoms. Two distinct fluorine resonances were observed at δ
−11.480 and δ −17.238 with different coupling patterns, the
former showing the same fluorine−hydride coupling constant
1
as that for the hydride resonance in the H NMR spectrum
1
by the pz-CH3 resonances (δ 2.134−2.146) in the H NMR
spectrum (see Supporting Information [SI]).
(11.7 Hz). The hydrogen of the difluoromethyl group could
not be readily identified, as many couplings were present.
An attempt to activate the C−H bond of CF3H under similar
conditions as those employed for the above studies failed to
show evidence for oxidative addition. Perhaps the bulk of the
three fluorines prevents formation of the alkane σ-complex that
precedes C−H activation.9,27
Reductive Elimination of Tp′Rh(CNneopentyl)(R)H.
The rates for reductive elimination of complexes 2−7 were
determined by monitoring the first order disappearance of the
Tp′Rh(CNneopentyl)(CH2F)H (6) was prepared via
exchange reaction of Tp′Rh(CNneopentyl)(CH3)H under a
pressure of CH3F in C6D12. However, the reaction is quite slow
and the yield was only 23% after two weeks. The exchange with
more labile Tp′Rh(CNneopentyl)(n-pentyl)H led to a higher
conversion, 50% after 5 days. Small amounts of carbodiimide
activation products were seen (∼10%). The hydride resonance
for 6 appeared as a doublet of doublets at δ −14.221 (JRh−H
=
6997
dx.doi.org/10.1021/ja400966y | J. Am. Chem. Soc. 2013, 135, 6994−7004

Journal of the American Chemical Society
Article
a
Table 1. Rates of Reductive Elimination of RH from Tp′Rh(CNneopentyl)(R)H in C6D6
g
kre(RH), s−1
τ1/2
ΔGre kcal·mol−1
⧧
R
T (°C)
b
C6H5
66.9
66.9
66.9
66.9
66.9
66.9
66.9
100.0
80
3.78 × 10−7
1.11 (2) × 10−5
3.26 (4) × 10−5
3.21 (7) × 10−4
8.31 (22) × 10−6
3.54 (12) × 10−6
9.7 (11) × 10−5
1.28 (4) × 10−5
∼4 × 10−5
20.8 d
17.3 h
5.9 h
29.98 (5)
27.71 (1)
26.98 (1)
25.43 (1)
27.90 (2)
28.48 (2)
26.24 (8)
30.36 (1)
∼ 27.9
CH2C(O)CH3, 2
CH2CCCH3, 3
CH2O-t-Bu, 4
CH2CF3, 5
0.6 h
23.2 h
54.4 h
2.0 h
CH2F, 6
CH2OCH3, 7
CHF2, 8
15.0 h
∼ 4.8 h
73.2 h
5.6 h
c
CH2Cl
d
CH2CN
100
2.62(7) × 10−6
3.41(13)× 10−5
2.46(4) × 10−7
31.36 (2)
23.52 (3)
36.81 (1)
e
CH3
23
f
C6F5
138.5
32.5 d
a
b
Errors are reported as standard deviations. Rate for reductive elimination was calculated from Eyring plot data in ref 28, ΔGre⧧ = 37.8 − T × 23 =
c
d
e
f
29.98 kcal mol−1 at 340 K. Data for CH2Cl from ref 20b. Data for CH2CN from ref 21. Data for CH3 from ref 19a. Data for C6F5 from ref 24.
g
Error in ΔG⧧ calculated using σG = (RT/kre)σk.
re
hydride resonance in C6D6 at 66.9 °C by 1H NMR
spectroscopy (see SI, Tables S1−S13). Generally, reductive
eliminations of substituted methanes CH3X from 2-7 were
much slower than the elimination of methane from Tp′Rh-
(CNneopentyl)(CH3)H at room temperature (τ1/2 ≈ 5 h). The
reductive elimination experiment with 8 was performed at
100.0 °C because no appreciable decrease of the corresponding
hydride resonance was detected at 66.9 °C after one week. The
corresponding activation barriers for reductive elimination in
2−8 are ∼2−7 kcal·mol−1 higher than that in Tp′Rh-
(CNneopentyl)(CH3)H. The rate of CH2F2 reductive elimi-
nation for 8 is comparable to that for Tp′Rh(CNneopentyl)-
(C6H5)H, which indicates that the difluoro−substitution greatly
stabilizes the compound vs Tp′Rh(CNneopentyl)(CH3)H.
Competitive Selectivity Experiments. The relative
selectivity of the fragment [Tp′Rh(CNneopentyl)] for C−H
activation of the various substrates was determined by
photolysis of 1 in a mixture of two substrates. The ratio of
(CNneopentyl)], which underwent activation of only one type
of C−H bond in each substrate. As shown in Table 1,
substituted products 2−8 underwent reductive elimination
much more slowly (τ1/2 = 0.6 h to 37 d at 66.9 °C) than the
unsubstituted methyl hydride complex (τ1/2 = 5.6 h at 23 °C).
These results generally indicate that α-substitution at carbon
with an electron withdrawing group increases the ‘stability’ of
the alkyl hydride species, or more properly, decreases the
lability of the alkyl hydride.29 Among 2−8, elimination of tert-
butyl methyl ether from 4 is the fastest (τ1/2 = 0.6 h), followed
by the less bulky dimethyl analogue 7 (τ1/2 = 2.0 h). An
unsaturated substituent improves the stability to some extent as
shown in complexes 2 and 3 (τ1/2 = 17.3 and 5.9 h). Fluoro
substitution had a stronger effect on the stability in the order
CH2CF3 < CH2F < CHF2, in which substitution at the β-site
increases the stability to a lesser extent than substitution at the
α-site, and di-fluorination is far superior to mono-fluorination.
These observations are in agreement with earlier studies of
alkylnitriles and alkyl chlorides, where α-substitution had the
most dramatic effect on stability.20,21 The difference between
fluorination and chlorination on stabilization is minimal − the
barrier for elimination of CH2F vs CH2Cl is only ∼0.6 kcal
mol−1 higher for the former. A cyano substituent dramatically
stabilizes the corresponding cyanomethyl hydride complex
compared to an α-keto (2) or α-alkynyl (3) group probably
due to its higher electron-withdrawing capability.
1
the two substrates was measured by H NMR analysis before
irradiation. The competition experiments were carried out with
only 5 min of photolysis at low temperature so that the product
ratio represented the kinetic products of the reaction. The
product distribution was determined on the basis of the relative
1
areas of the corresponding hydride resonances by H NMR
spectroscopy in deuterated solvent (C6D6 or C6D12). The
relative competitive rates k2/k1 reported in Table 2 were
calculated on a per−molecule basis using eq 9, where I2/I1 is
the integration area of the hydride resonances and n1/n2 is the
mole ratio of two substrates (subscript 2 refers to benzene and
subscript 1 represents the other competing substrate). The
The kinetic selectivity of the coordinatively unsaturated
intermediate [Tp′Rh(CNneopentyl)] toward the substrates R−
H discussed above follows the order:
⧧
Ph > CH2C≡ CCH3 > CH2OCH3 > CH3
> CH2C(=O)CH3 > CH2F > CH2OtBu
> CH2CF > C6F5 > CHF
differences in free energies of activation ΔΔGoa can be
calculated using eq 10. The results of competition between 2−8
and benzene are included in Table 2 as well as previous results
with methane and pentafluorobenzene.
3
2
⎛
⎞⎛
⎞
Generally, activation of many of the substituted methanes is
kinetically less preferable than activation of methane. This
could be attributed to the steric effect of α-substitution as the
order of selectivity follows approximately the size of the
substituent attached to the α-carbon. Steric effects may
interfere with the formation of an alkane σ-complex prior to
C−H oxidative cleavage. It is interesting to compare activation
of methyl tert-butyl ether vs dimethyl ether, in which the free
energy of activation (ΔGoa⧧) of the former is ∼0.4 kcal·mol−1
higher than that of the latter. This kinetic difference can be
k2
k1
I2 n1
=
⎜
⎝
⎟⎜
⎟
⎠
I
n
2
1 ⎠⎝
(9)
⧧
ΔΔGoa = RT ln(k2/k1)
(10)
DISCUSSION
■
As shown in Schemes 1 and 2, both photochemical and thermal
methods can be used to generate the fragment [Tp′Rh-
6998
dx.doi.org/10.1021/ja400966y | J. Am. Chem. Soc. 2013, 135, 6994−7004

Journal of the American Chemical Society
Article
attributed to statistical factors favoring the activation of
dimethyl ether (RT ln 2 = 0.4 kcal mol−1).
With the data for reductive elimination and kinetic selectivity
shown in Tables 1 and 2, the relative free energies for
D (Rh−C) = [ΔH(Rh−R) − ΔH(Rh−Ph)]
rel
= [D(R−H) − D(Ph−H)] − ΔG° +
RTln(6/#H)
(12)
Table 2. Kinetic Selectivity Data Determined from
Competition Experiments
Note that all of the Rh−C bond strengths for these
substituted derivatives are weaker than the Rh−Ph bond
(except for perfluorophenyl). Furthermore, when the sub-
stituents are compared as Rh−CH2−X derivatives of the Rh−
CH3 parent complex, it can be seen that fluorine substituents
(or CF3) increase the Rh−C bond strength whereas groups that
weaken the Rh−C bond can engage in resonance or have
weakly electron withdrawing groups (e.g., −OR).
a
c
⧧
ΔΔGoa
b
entry
substrates
T (°C)
k2/k1
(kcal mol−1
)
1
2
3
4
5
6
7
8
9
benzene: acetone
benzene: 2-butyne
benzene: MeOtBu
benzene: CH3OCH3
benzene: CH3CF3
benzene: CH3F
8
10
8
3.71 (19)
2.17 (11)
4.53 (23)
2.33 (12)
18.2 (9)
0.73 (3)
0.44 (3)
0.84 (3)
0.48 (3)
1.63 (3)
0.81 (3)
2.33 (3)
0.70 (3)
1.92 (3)
10
10
10
10
22
10
Plotting of these relative Rh−C bond energies vs C−H bond
energies along with previously determined data shows two
distinct correlations22 as shown in Figure 5. The lower line
(blue squares) connects sp, sp2, and sp3 hydrocarbons with no
functional groups and has a slope of 1.38 (3). These substrates
include terminal alkynes, terminal alkenes, benzene, and
alkanes. Siegbahn noted this same trend in DFT calculations
of metal−carbon bond strengths for second row transition
metal atoms.31 The upper line (red triangles) connects
substituted methanes. The slope of 1.40 (14) is indistinguish-
able from that for the unsubstituted hydrocarbons, and is offset
vertically by +7.5 kcal mol−1 from the hydrocarbon line. This
offset represents the increase in the metal−carbon bond strength
due to the presence of the substituent relative to the M−C bond
strength that would be predicted based upon the strength of the C−
H bond that was broken. However, the effect of the substituent, and
in particular the substituents that are in hyperconjugation with the
M−C bond, is to weaken the M−CH2X bond relative to M−CH3.
In the case of CH2F and CF2H, the M−C bonds are actually
stronger than the M−CH3 bond. CH2CF3, with its three β-
fluorines, shows a very slight increase in metal−carbon bond
strength. This result is consistent with the observation by Reger
that (Me2NCS2)(PEt3)Pd(CH2CH2CF3) and (Me2NCS2)-
(PEt3)Pd[CH(CH3)(CF3)] form in a 1:1 equilibrium, whereas
(Me2NCS2)(PEt3)Pd(CH2CH2CN) isomerizes completely to
(Me2NCS2)(PEt3)Pd[CH(CH3)(CN)],6 implying that CF3
substitution on a methyl group has only a minor effect on
the M−C bond strength.
4.24 (21)
62.9 (31)
3.31 (17)
benzene: CH2F2
d
benzene: methane
e
benzene: C6F5
30.2 (15)
a
b
Each sample was irradiated for 5 min. Errors in rate ratio estimated
at 5% for proton NMR integration, giving σG = (RT/ratio)σratio
=
c
0.05RT ≈ 0.03 kcal mol−1 (see Figures S-21 to S-26 in SI). A positive
d
value denotes that benzene is kinetically favored. Data are calculated
from methane vs pentane at room temperature and pentane vs
e
benzene at −15 °C from ref 22. Data are from ref 24.
compounds 2−8 vs Tp′Rh(CNneopentyl)(Ph)H were calcu-
lated using eq 11 (a positive ΔG° implies benzene is
thermodynamically favored). Since the temperature at which
the reductive elimination rates were measured for the various
RH substrates varied from 22 to 138 °C, the ΔG⧧ for benzene
reductive elimination at each temperature was calculated from
the known activation parameters for use in eq 11.28,30 The
relative metal−carbon bond energies compared to D(Rh−Ph)
were then calculated on the basis of the experimental values of
carbon−hydrogen bond energies D(R−H) using eq 12. Note
that the statistics for the entropic contribution to ΔΔG⧧ are
removed by inclusion of the RT ln(6/#H) term where #H is
the number of C−H bonds available for oxidative addition
(Table 3).
Tp′Rh(CNR′)(Ph)H + R−H
Keq
HoooooooooooooI Tp′Rh(CNR′)(R)H + Ph−H
For comparison with the experimental system, DFT
calculations were performed on a broad scope of substituted
and unsubstituted substrates using the simplified model
R′=neopentyl
ΔG° = ΔG⧧re(PhH) + ΔΔG⧧ − ΔG⧧re(RH)
(11)
oa
a
Table 3. Kinetic and Thermodynamic Data for Tp′Rh(CNneopentyl)(R)H
b
RH
no. of H
ΔΔG⧧
ΔG°
D(R−H)
Drel(Rh−C)
oa
benzene
CH3CF3
CH2F2
CH3F
6
3
2
3
4
3
6
6
3
6
1
0.00
1.63 (3)
2.33 (3)
0.81 (3)
0.70 (3)
1.48 (3)
0.73 (3)
0.48 (3)
0.84 (3)
0.44 (3)
1.92 (3)
0.00 (5)
3.71 (10)
1.19 (9)
112.9
106.7
103.2
101.3
105.0
94.8
0.0
−9.5
−10.2
−13.5
−15.8
−17.0
−19.9
−21.0
−24.9
−25.6
+11.2
2.31 (10)
8.17 (11)
−0.66 (10)
3.00 (9)
c
CH4
c
CH3CN
acetone
96.0
dimethyl ether
methyl tert-butyl ether
2-butyne
4.22 (16)
5.39 (9)
96.1
e
93.0
3.44 (9)
90.7
f
C6HF5
−6.57 (9)
116.5
a
b
All values aare in kcal·mol−1. Hydrocarbon C−H bond strengths are from Luo, Y.-R. Comprehensive Handbook of Chemical Bond Energies; CRC
c
c
e
Press: Boca Raton, FL, 2007. Data from ref 22. Data from ref 21. The bond strength for methyl tert-butyl ether is not experimentally known;
therefore, the value for methyl ethyl ether was used instead. Data from ref 24.
f
6999
dx.doi.org/10.1021/ja400966y | J. Am. Chem. Soc. 2013, 135, 6994−7004

Journal of the American Chemical Society
Article
Figure 6. DFT calculated plot of relative M−C bond strengths vs C−
H bond strengths for Tp′Rh(CNMe)(R)H. The lower line is fit to the
Figure 5. Plot of relative experimental M−C bond strengths vs C−H
bond strengths. The solid line is fit to the hydrocarbons and aliphatic
■
■
hydrocarbons (blue , y = 1.593x − 179.6), and the upper line is fit to
nitriles −(CH2)n−CN (n = 2−5) (blue , y = 1.376x − 159.5), and
▲
▲
the −CH2X and CHF2 substrates (red , y = 1.457x − 156.2). Data
the dashed line is fit to the −CH2X substrates and −CHF2 (red , y =
for C6F5H and CH3CF3 activation is also shown (Δ), but not included
in the fits. M06-2X method and basis set 6-31g** for first row atoms
and pseudopotentials, additional functions optimized by Stuttgart
group for atoms beyond the second row (see ref 23 for details on the
choice of method). Experimental C−H bond strengths were used for
all substrates except the alkynes. Alkyne C−H bond strengths were
calculated (B3LYP) since experimental values are unavailable or have
large errors.23 The vertical separation of the lines at DC−H = 100 kcal
mol−1 is 9.7 kcal mol−1.
1.4024x − 154.6). Also shown are −C6F5 and −CH2CF3 (Δ), which
are not included in either fit. Experimental C−H bond strengths were
used for all substrates except the alkynes and nitriles (except
acetonitrile). Alkyne and nitrile C−H bond strengths were calculated
(B3LYP) since experimental values are unavailable or have large
errors.23 The vertical separation of the lines at DC−H = 100 kcal mol−1
is 7.5 kcal mol−1.
fragment [HB(3,5-dimethylpyrazolyl)3Rh(CNMe)] (see SI for
details). A plot of calculated relative Rh−C bond strengths vs
the experimental (or calculated for alkyne) C−H bond
strengths within these substrates also shows two distinct linear
correlations with slopes of 1.59 and 1.46 for the analogous two
sets of compounds (Figure 6). There is generally good
agreement with the observed experimental trends in Rh−C
bond strengths, but DFT overestimates the range of Rh−C
bond strengths by 4−15%.
bonds can be excluded or at least have minimal effect here, as
bulkier Rh−CHF2 is still stronger than Rh−CH2F. The increase
in bond strength is therefore associated with an increase in the
ionic character of the metal−carbon bonding. Sakaki has
reported calculations on Pd(PH3)2(CH3)H and Pd-
(PH3)2(CH2CN)H and their (H2PCH2CH2PH2) analogues.
He concluded that charge transfer from Pd to the CH2CN
ligand was greater than in the CH3 ligand, due to the mixing of
the π*(CN) and σ*(C−H) orbitals to give a lower energy
acceptor orbital in CH3CN vs the σ*(C−H) in CH4.34 In
contrast to the present study, however, Sakaki calculated that
the Pd−CH2CN bond was some 11−16 kcal mol−1 stronger
than the Pd−CH3 bond.
There are other, more general aspects of this study that are
worth mentioning. First, the thermodynamics of a hydrocarbon
exchange reaction are controlled by the bonds that are broken
and formed. Here, acetonitrile activation is strongly thermody-
namically preferred over methane activation (by ∼9 kcal mol−1)
despite the fact that the product Tp′Rh(CNneopentyl)-
(CH2CN)H has a slightly weaker Rh−C bond (by 1.2 kcal
mol−1). One way to think about this is that in an equilibrium
between methane and acetonitrile, cleavage of the acetonitrile
C−H bond (94.8 kcal mol−1) means that a methane C−H
bond (105.0 kcal mol−1) has been left intact. It is this difference
that gives the large thermodynamic driving force favoring
acetonitrile activation.
While the majority of products form Rh−CH2X bonds that
are weaker than the Rh-methyl bond, the bonds are not as weak
as one would anticipate by comparing H−CH2X and H−CH3
bond energies. The C−H bonds are weak in many of the
substrates because the substituent can stabilize the radical after
the rupture of C−H bond through resonance. However, the
ionic contribution to the M−R bond is more important, as the
ionization potential is smaller for a metal than for hydrogen,
which indicates the charge distribution (ionicity) dominates the
bond strength, as pointed out by Siegbahn,31 Harvey,14 and
Clot.32 This could also account for the strengthening of M−R
bonds through purely inductive effects as in M−CH2F and M−
CHF2 (although Siegbahn’s calculations of bond strengths with
rhodium atoms indicates little effect of fluorine substitution on
bond strength33). Therefore either unsaturated substituents or
electronegative functional groups can contribute to strengthen-
ing the M−C bond by polarizing electron density from the
metal center to the carbon atom, leading to Rh−C bonds that
are stronger than expected on the basis of the C−H bond that
was broken. The current results also indicate the difluoromethyl
hydride complex contains a more ionic Rh−C bond than that
in fluoromethyl hydride species. The postulation that direct
repulsion from alkyl groups will weaken the resulting M−C
Another seemingly contradictory point that arises from this
study is the observed ‘stability’ of Tp′Rh(CNneopentyl)-
(CH2CN)H vs Tp′Rh(CNneopentyl)(CH3)H. The former
has a half−life of 3 days at 100 °C whereas for the latter it is
only 5 h at 25 °C. Why should the molecule with the weaker
7000
dx.doi.org/10.1021/ja400966y | J. Am. Chem. Soc. 2013, 135, 6994−7004

Journal of the American Chemical Society
Article
Rh−C bond appear to be much more ‘stable’ than the molecule
with the stronger Rh−C bond? The answer lies in the
recognition that heating a molecule to induce reductive
elimination represents a measure of the kinetics of that
reaction, not the thermodynamics of the elimination. That is,
we really mean that Tp′Rh(CNneopentyl)(CH2CN)H is less
labile than Tp′Rh(CNneopentyl)(CH3)H when we observe its
reluctance to lose acetonitrile. The reason for this can be seen
by comparison of the transition states for reductive elimination.
In the case of Tp′Rh(CNneopentyl)(CH2CN)H, a σ-C,H−
CH3CN complex is formed with a weak C−H bond (95 kcal
mol−1) as acetonitrile is lost. In the case of Tp′Rh-
(CNneopentyl)(CH3)H, a σ-C,H−CH4 complex is formed
with a strong C−H bond (105 kcal mol−1) as methane is lost.
The barrier for methane loss is lower than for acetonitrile
because of the greater strength of the C−H bond that is formed
in the transition state, lowering its energy.
Examination of Figure 5 shows that primary carbon−rhodium
bonds are stronger than secondary carbon−rhodium bonds.
Now consider the insertion of acrylonitrile as shown in eqs 3
and 4. Is the branched product favored over the linear product
because the metal−carbon bond is stronger? No. In fact, the
metal−carbon bond is weaker in the branched product because
homolysis leads to a radical that is resonance stabilized, vs a
primary radical in the linear product. The same applies to
styrene insertion in eq 5. Figure 5 also shows that Rh−CH2CN
and Rh−CH2Ph are weaker than Rh−CH3. So why are the
branched products favored in the nitrile−substituted (and
phenyl−substituted) products if the metal−carbon bonds are
weaker? The answer is shown in eqs 14 and 15, where the C−H
Several groups have reported related relationships between
product stabilities and the difficulty of C−H activation, which
was quantitatively expressed by plotting metal−carbon bond
energies in the products versus the corresponding known
carbon−hydrogen bond strengths in the substrates. The
examination by Bryndza and Bercaw in 1987 of the correlation
between metal−heteroatom and heteroatom−hydrogen bond
energies indicates that the difference between H−X and H−Y
BDEs is the same as the difference in M−X and M−Y BDEs
(i.e., related by a slope of 1.0).35 This is reasonable only if the
equilibrium constants Keq in eq 13 were measured to be
approximately unity. The relative bond energies of LnM−R (R
= alkyl, alkenyl, aryl, alkynyl, etc.) were included and varied
over a range of 40 kcal·mol−1 from R = CH2Ph to CCR′.
bond that is also formed/broken is emphasized. In the
branched products, a strong primary methyl C−H bond is
formed (∼100 kcal mol−1), whereas in the linear product a
weak α-cyano or weak benzylic secondary C−H bond is formed
(∼88−92 kcal mol−1). In the substituted derivatives, it is the
formation of the (largely ignored) C−H bond that determines
the product selectivity, not the metal−carbon bond strength!
On this basis, many of the prior statements in the literature
(including our own) regarding metal−carbon bond strengths in
compounds with α-electron-withdrawing substituents need to
be re-evaluated.
Keq
LnM−X + H−Y HooI LnM−Y + H−X
LnM = (DPPE)MePt, Cp*(PMe3)Ru
(13)
Wolczanski studied a number of (t-Bu3SiO)2(t-Bu3SiNH)-
TiR compounds, which displayed a strong correlation of
D(TiR)rel with D(RH) with a slope of 1.1 when the data for R
= Bz, Mes, H, and Ph were removed.36 The observation of a
slope >1 indicates about a 10% increase in Ti−R bond
strengths relative to R−H bond strengths, which implies
thermodynamic control of the product distribution. The data
for toluene and mesitylene lay above the line and expressed
additional stabilization of ∼6−7 kcal/mol, which presumably
has its origin due to the similar effects quantified here.
Eisenstein and Perutz have examined the above titanium and
Tp′Rh systems computationally, where both M−C and C−H
bond strengths can be calculated using DFT. They found
reasonably good correlations for the substrates examined (all
hydrocarbons), with the exception of benzyl and allyl, which lay
above the correlation line.37 In studying the activation of
polyfluorobenzenes with a variety of metal complexes (Zr−Ni),
they found good correlations with slopes in the range 1.9−
3.0.32 Landis has also summarized and compared a number of
systems where some correlation between M−X and X−H (X
includes C) bond strengths has been observed.38
CONCLUSIONS
■
In this work, we have taken kinetic measurements for the
reductive elimination of R−H in a series of H−Rh−CH2X
complexes and determined the relationship between the relative
M−C bond energies and C−H bond energies. The stabilization
effect by α-substitution on Rh-methyl has been quantified and
shows that the substituents aryl, vinyl, alkynyl, alkoxy, CN, and
keto all weaken the Rh−CH2X bond compared to the Rh−
methyl bond. While these metal−carbon bonds are weaker than
in the Rh−methyl complex, they are not as weak as one would
expect on the basis of the relative C−H bond strengths
(methane vs H−CH2X). Hence, the substituent is having a
positive effect on strengthening the metal−carbon bond.
Fluorine or chlorine substitution slightly increases the Rh−
methyl bond strength, and difluoro substitution strengthens the
bond by almost 6 kcal mol−1. The stabilization effect is believed
to result from increased polarization of the M−C bond, adding
more ionic character.
EXPERIMENTAL SECTION
■
Finally, we return to the issues first raised at the beginning of
this manuscript, the control of regioselectivity in olefin
insertion reactions. Consider the insertion of an alkene to
give a linear or branched hydrocarbon as shown in eq 1 or 2. Is
the linear product favored because the primary carbon−metal
bond is stronger than the secondary carbon−metal bond? Yes.
General Procedures. All operations and routine manipulations
were performed under a nitrogen atmosphere, either on a high-vacuum
line using modified Schlenk techniques or in a Vacuum Atmospheres
Corp. Dri-Lab. Acetone, acetonitrile, carbon tetrachloride, bromoform,
methyl tert-butyl ether, and 2-butyne were purchased from Aldrich
Chemical Co. Fluoromethane, difluoromethane, trifluoromethane and
7001
dx.doi.org/10.1021/ja400966y | J. Am. Chem. Soc. 2013, 135, 6994−7004

Journal of the American Chemical Society
Article
α,α,α-trifluoroethane were purchased from Matrix Scientific and used
straight from lecture bottles. Benzene-d6 was purchased from
Cambridge Isotopes. Prior to use it was distilled under vacuum from
a dark purple solution of benzophenone ketyl and stored in an ampule
with a Teflon valve. Acetone was distilled under vacuum from a
solution dried over potassium carbonate. Carbon tetrachloride was
distilled under vacuum from a solution dried over calcium chloride.
Methyl tert-butyl ether was purchased dry and used straight from
bottle with a sure-seal cap. 2-Butyne was distilled under vacuum after
drying over molecular sieves. Preparation of Tp′Rh(CNneopentyl)(η2-
PhNCNneopentyl) (1) has been previously reported.25
Preparation of Tp′Rh(CNneopentyl)(CH2O-t-Bu)H (4). A solution
of 1 (9 mg, 0.013 mmol) dissolved in 0.5 mL of methyl tert-butyl ether
was placed in an NMR tube sealed with a Teflon cap. This sample was
irradiated for 30 min at −20 °C. The solvent was immediately
removed in vacuo at room temperature. The resulting yellow residue
1
was dissolved in C6D6. H NMR (400 MHz, C6D6): δ −14.403 (d,
1JRh−H = 25.2 Hz, 1 H, Rh−H), 0.741 (s, 9 H, C(CH3)3), 1.340 (s, 9
H, O−C(CH3)3), 2.211 (s, 3 H, pz-CH3), 2.232 (s, 3 H, pz-CH3),
2.318 (s, 3 H, pz-CH3), 2.400 (s, 3 H, pz-CH3), 2.664 (s, 3 H, pz-
CH3), 2.721 (s, 3 H, pz-CH3), 2.788 (s, 2 H, NCH2), 4.888 (d, 2JRh−H
= 16.1 Hz, 2 H, RhCH2O), 5.645 (s, 1 H, pz-H), 5.656 (s, 1 H, pz-H),
5.868 (s, 1 H, pz-H). 13C NMR (500 MHz, C6D6): δ 12.68 (s, 2 C, pz-
CH3), 12.91 (s, 1 C, pz-CH3), 14.56 (s, 1 C, pz-CH3), 15.31 (s, 1 C,
All photolysis experiments were performed using a 200 W Hg(Xe)
arc lamp purchased from Oriel, which was fitted with a water-filled IR
filter and a 300 nm low pass filter. Low temperatures were maintained
pz-CH3), 15.64 (s, 1 C, pz-CH3), 26.85(s, 3 C, OC(CH3)3), 28.17 (s,
1
with methanol/N2 in a Pyrex Dewar. All H and 13C NMR spectra
1
3 C, CH2C(CH3)3), 31.74 (s, 1 C, CH2C(CH3)3), 53.88 (d, JRh−C
=
were collected on either a Bruker Avance 400 or Avance 500 MHz
spectrometer. All HSQC experiments were done on an Avance 500
MHz spectrometer. All chemical shifts were reported in ppm (δ)
referenced to the chemical shifts of residual solvent resonances
26.1 Hz, 1 C, RhCH2O), 55.97 (s, 1 C, RhCNCH2), 72.88 (s, 1 C,
OC(CH3)3), 102.26 (s, 1 C, pz-H), 106.25 (s, 1 C, pz-H), 106.37 (s, 1
C, pz-H), 142.96 (s, 1 C, pz-C), 143.22 (s, 1 C, pz-C), 143.56 (s, 1 C,
pz-C), 148.83 (s, 1 C, pz-C), 149.85 (s, 1 C, pz-C), 150.61 (s, 1 C, pz-
C).
1
(C6HD5, δ 7.16 or 128.0). While H chemical shifts are given to 3
decimal places ( 0.4 Hz), these values can vary slightly with
concentration and temperature. 13C shifts are given to 2 decimal
places ( 1 Hz). Elemental analysis was performed by the University of
Rochester using a Perkin-Elmer 2400 series II elemental analyzer in
CHN mode. All kinetic plots and least-squares error analysis were
done using Microsoft Excel.
Preparation of Tp′Rh(CNneopentyl)(CH2CF3)H (5). Method A.
To a yellow solution of 10 mg (0.018 mmol) of Tp′Rh-
(CNneopentyl)(CH3)Cl in 1 mL of THF was added 4 mg (0.018
mmol) of Cp2ZrH2. The suspension was stirred for 30 min and
changed from light yellow to white. The slurry was transferred to a
high pressure NMR tube after filtration through a glass wool plug.
Removal of the volatiles gave a white residue to which was added 0.5
mL C6D12 to give white cloudy suspension after sonication. The tube
was then pressurized with 30 psi of CH3CF3 and shaken carefully at
room temperature. 1H NMR spectroscopic analysis shows 51%
conversion to 5 after standing overnight at room temperature.
Complete conversion was achieved after 2 days. The slurry was then
filtered and a clear colorless solution was obtained, giving a white
crystalline solid after evaporation. The formation of 5 was confirmed
by elemental analysis and NMR analysis in C6D6 (see below).
Remarkably, 5 did not react with excess CCl4. It also has a surprising
resistance to air, lasting over one week.
Preparation of Tp′Rh(CNneopentyl)(CH2C(O)CH3)H (2). A
solution of 1 (6 mg, 0.013 mmol) dissolved in 0.4 mL of acetone
was placed in an NMR tube sealed with a Teflon cap. This sample was
irradiated for 3 min at −20 °C, as the bright yellow solution
photobleached to a pale yellow. The solvent was immediately removed
in vacuo at −20 °C. The resulting pale-yellow residue was dissolved in
1
1
C6D6. H NMR (400 MHz, C6D6): δ −14.778 (d, 1 H, JRh−H = 19.8
Hz, Rh−H), 0.712 (s, 9 H, C(CH3)3), 2.142 (s, 3 H, pz-CH3), 2.169
(s, 3 H, pz-CH3), 2.268 (s, 3 H, pz-CH3), 2.317 (s, 3 H, pz-CH3),
2.469 (s, 3 H, COCH3), 2.530 (s, 3 H, pz-CH3), 2.840 (s, 3 H, pz-
2
2
CH3), 2.871 (dd, JRh−H = 5.3 Hz, JH−H = 3.3 Hz, 1 H, RhCH2CO),
2.993 (dd, JRh−H = 5.7 Hz, JH−H = 3.6 Hz, 1 H, RhCH2CO), 2.920
2
2
Method B. Eight mg (mmol) of 1 was partially dissolved in 0.5 mL
pentane and led to a pale yellow clear solution after irradiation for 30
min at 10 °C. The solution was then transferred to a high pressure
NMR tube and freeze−pump−thaw degassed (3X). Thirty psi of
CH3CF3 was then introduced into the tube, which was shaken
carefully at room temperature. Clean formation of 5 was observed after
1 week with trace amounts of products from activation of carbodiimide
(<5%). 1H NMR (400 MHz, C6D6): δ −14.344 (d, 1JRh−H = 21.3 Hz, 1
H, RhH), 0.612 (s, 9 H, C(CH3)3), 2.134 (s, 3 H, pzCH3), 2.146 (s, 3
H, pzCH3), 2.265 (s, 3 H, pzCH3), 2.323 (s, 3 H, pzCH3), 2.515 (s, 3
2
2
(d, JH−H = 14.0 Hz, 1 H, NCH2), 3.131 (d, JH−H = 14.0 Hz, 1 H,
NCH2), 5.593 (s, 1 H, pz-H), 5.610 (s, 1 H, pz-H), 5.822 (s, 1 H, pz-
H). 13C{1H} NMR (500 MHz, C6D6): δ 12.43 (s, 1 C, pz-CH3), 12.59
(s, 1 C, pz-CH3), 12.78 (s, 1 C, pz-CH3), 14.72 (s, 1 C, pz-CH3),
15.55 (s, 1 C, pz-CH3), 16.18 (s, 1 C, pz-CH3), 21.02 (d, 1JRh−C = 20.6
Hz, 1 C, RhCH2CO), 26.70 (s, 3 C, CH2C(CH3)3), 28.73 (s, 1 C,
CH3), 31.49 (s, 1 C, CH2C(CH3)3), 56.72 (s, 1 C, RhCNCH2), 105.55
(s, 1 C, pz-CH), 106.77 (s, 1 C, pz-CH), 106.78 (s, 1 C, pz-CH),
143.26 (s, 1 C, pz-C), 143.68 (s, 1 C, pz-C), 144.00 (s, 1 C, pz-C),
149.26 (s, 1 C, pz-C), 150.40 (s, 1 C, pz-C), 150.64 (s, 1 C, pz-C),
216.68 (s, 1 C, CO).
4
H, pzCH3), 2.524 (s, 3 H, pzCH3), 2.703 (d, JRh−H = 2.2 Hz, 2 H,
NCH2), 5.530 (s, 1 H, pzH), 5.605 (s, 1 H, pzH), 5.795 (s, 1 H, pzH),
signals for RhCH2 are overlapping with those for pzCH3 based on
Preparation of Tp′Rh(CNneopentyl)(CH2CCCH3)H (3). A
solution of 1 (6 mg, 0.013 mmol) dissolved in 0.4 mL of 2-butyne
was placed in an NMR tube sealed with a Teflon cap. This sample was
irradiated for 3 min at −20 °C or until bright yellow solution
photobleached to pale yellow. The solvent was immediately removed
in vacuo at −20 °C. The resulting yellow residue was dissolved in
1
cross coupling in the H−13C HSQC spectrum. 13C{1H} NMR (500
1
2
MHz, C6D6): δ 8.31 (quint, JRh−C = JF−C = 26.9 Hz, 1 C, RhCH2),
12.44 (s, 1 C, pz-CH3), 12.60 (s, 1 C, pz-CH3), 12.72 (s, 1 C, pz-
CH3), 14.28 (s, 1 C, pz-CH3), 15.46 (s, 1 C, pz-CH3), 15.59 (s, 1 C,
pz-CH3), 26.54 (s, 3 C, C(CH3)3), 31.38 (s, 1 C, C(CH3)3), 56.11 (s,
1 C, RhCNCH2), 106.81 (s, 1 C, pz-H), 106.55 (s, 1 C, pz-H), 105.56
(s, 1 C, pz-H), 135.70 (q,1JF−C = 275.0 Hz, CF3), 143.40 (s, 1 C, pz-
C), 143.61 (s, 1 C, pz-C), 143.87 (s, 1 C, pz-C), 149.34 (s, 1 C, pz-C),
149.60 (s, 1 C, pz-C), 150.55 (s, 1 C, pz-C). 19F NMR (400 MHz,
1
1
C6D6. H NMR (400 MHz, C6D6): δ −14.567 (d, 1 H, JRh−H = 21.9
Hz, Rh−H), 0.783 (s, 9 H, C(CH3)3), 1.742 (br, 3H, CCCH3),
2.181 (s, 3 H, pz-CH3), 2.196 (s, 3 H, pz-CH3), 2.283 (s, 3 H, pz-
CH3), 2.349 (s, 3 H, pz-CH3), 2.502 (m, 1 H, RhCH2CC), 2.584 (s,
3 H, pz-CH3), 2.657 (m, 1 H, RhCH2CC), 2.699 (s, 3 H, pz-CH3),
2.796 (s, 2 H NCH2), 5.600 (s, 1 H, pz-H), 5.626 (s, 1 H, pz-H), 5.785
(s, 1 H, pz-H). 13C{1H} NMR (500 MHz, C6D6): δ −11.88 (d, 1JRh−C
= 22.4 Hz, 1 C, RhCH2), 5.20(s, 1 C, CCCH3), 12.57 (s, 1 C, pz-
CH3), 12.59 (s, 1 C, pz-CH3), 12.78 (s, 1 C, pz-CH3), 14.42 (s, 1 C,
pz-CH3), 15.35 (s, 1 C, pz-CH3), 15.70 (s, 1 C, pz-CH3), 26.58 (s, 3
C, CH2C(CH3)3), 32.15 (s, 1 C, CH2C(CH3)3), 58.58 (s, 1 C,
RhCNCH2), 70.62 (s, 1 C, RhCH2CC), 93.13 (s, 1 C, RhCH2C
C), 105.380 (s, 1 C, pz-H), 106.29 (s, 1 C, pz-H), 106.41 (s, 1 C, pz-
H), 143.09 (s, 1 C, pz-C), 143.14 (s, 1 C, pz-C), 143.39 (s, 1 C, pz-C),
148.98 (s, 1 C, pz-C), 149.67 (s, 1 C, pz-C), 150.93 (s, 1 C, pz-C).
2
3
C6D6): δ 9.278 (dt, JCH −F = 15.4 Hz, JRh−F = 5.4 Hz, 3 F). Anal.
2
Calcd (found) for C23H36BF3N7Rh·THF0.5: C, 48.64 (48.34); H, 6.53
(6.34); N, 15.88 (15.88).
Preparation of Tp′Rh(CNneopentyl)(CH2F)H (6). The method for
preparing 6 was the same as method B for 5, except that CH3F was
used as the gas and a longer exchange time of 1 week was required for
completion. 1H NMR (400 MHz, C6D6): δ −14.221 (dd, 1JRh−H = 24.7
3
Hz, JF−H = 8.6 Hz, 1 H, RhH), 0.705 (s, 9 H, C(CH3)3), 2.163 (s, 3
H, pz-CH3), 2.211 (s, 3 H, pz-CH3), 2.280 (s, 3 H, pz-CH3), 2.378 (s,
3 H, pz-CH3), 2.528 (s, 3 H, pz-CH3), 2.633 (d, 4JRh−H = 6.9 Hz, 2 H,
7002
dx.doi.org/10.1021/ja400966y | J. Am. Chem. Soc. 2013, 135, 6994−7004

Journal of the American Chemical Society
Article
NCH2), 3.039 (s, 3 H, pz-CH3), 5.563 (s, 1 H, pz-H), 5.667 (s, 1 H,
pz-H), 5.807 (s, 1 H, pz-H), signals for RhCH2 are overlapping with
those for pzCH3 (see spectra in SI). 19F NMR (400 MHz, C6D6): δ
26.50 (s, 3 C, CH2C(CH3)3), 31.80 (s, 1 C, CH2C(CH3)3), 56.20 (s, 1
C, RhCNCH2), 74.78 (s, RhCH2CC), 88.91 (s, RhCH2CC), 106.89 (s,
1 C, pz-H), 107.55 (s, 1 C, pz-H), 108.39 (s, 1 C, pz-H), 142.47 (s, 2
C, pz-C), 142.76 (s, 1 C, pz-C), 144.01 (s, 1 C, pz-C), 151.40 (s, 1 C,
pz-C), 151.44 (s, 1 C, pz-C), 153.44 (s, 1 C, pz-C). Anal. Calcd
(found) for C25H38BClN7Rh: C, 51.26 (49.70); H, 6.54 (6.43); N,
16.74 (15.52).
−145.30 (tdd, 3JRhH−F = 8.6 Hz, 2JRh−F = 19.6 Hz, 2JCH −F = 50.1 Hz, 1
2
F). Trace amounts of the activation products of the released
carbodiimide (<5%) were also observed as well as a hydride resonance
1
at −14.16 ppm with JRh−H = 25.2 Hz. No other fluorine resonance
Preparation of Tp′Rh(CNneopentyl)(CH2O-t-Bu)Cl (4-Cl). A
solution of 1 (50 mg, 0.073 mmol) dissolved in 0.5 mL of methyl
tert-butyl ether was placed in an NMR tube sealed with a Teflon cap.
This sample was irradiated for 30 min at −20 °C; 1.0 mL of carbon
tetrachloride was added and the solution stirred under a nitrogen
atmosphere for 1 day. The volatiles were removed under vacuum, and
the yellow solid was purified by chromatography with 5:1 hexane/
THF as the eluent. Yellow crystals were collected (21.0 mg, 46%) by
recrystallization in dichloromethane layered with hexane at room
temperature. 1H NMR (500 MHz, C6D6): δ 0.761 (s, 9 H, C(CH3)3),
1.398 (s, 9 H, O−C(CH3)3), 2.132 (s, 3 H, pz-CH3), 2.153 (s, 3 H, pz-
CH3), 2.244 (s, 3 H, pz-CH3), 2.655 (s, 3 H, pz-CH3), 2.767 (d, 2JH−H
= 10.6 Hz, 2 H, NCH2), 2.814 (s, 3 H, pz-CH3), 2.958 (s, 3 H, pz-
CH3), 5.268 (dd, JRh−H = 0.8 Hz, JH−H = 3.4 Hz, 1 H, RhCH2O),
6.386 (dd, 2JRh−H = 0.4 Hz, 2JH−H = 3.4 Hz, 1 H, RhCH2O), 5.506 (s, 1
H, pz-H), 5.687 (s, 1 H, pz-H), 5.726 (s, 1 H, pz-H). 13C NMR (500
MHz, C6D6): δ 12.30 (s, 1 C, pz-CH3), 12.75 (s, 1 C, pz-CH3), 12.94
(s, 1 C, pz-CH3), 14.49 (s, 1 C, pz-CH3), 14.51 (s, 1 C, pz-CH3),
14.67 (s, 1 C, pz-CH3), 26.84 (s, 3 C, CH2C(CH3)3), 28.44 (s, 3 C,
OC(CH3)3), 31.91 (s, 1 C, CH2C(CH3)3), 56.11 (s, 1 C, RhCNCH2),
57.88 (d, 1JRh−C = 20.6 Hz, 1 C, RhCH2O), 73.94 (s, 1 C, OC(CH3)3),
106.60 (s, 1 C, pz-H), 107.67 (s, 1 C, pz-H), 108.73 (s, 1 C, pz-H),
142.53 (s, 1 C, pz-C), 142.86 (s, 1 C, pz-C), 144.23 (s, 1 C, pz-C),
151.12 (s, 1 C, pz-C), 152.03 (s, 1 C, pz-C), 152.71 (s, 1 C, pz-C).
Anal. Calcd (found) for C26H44N7BClRh·(C6H14)0.25: C, 51.50
(51.55); H, 7.46 (7.53); N, 15.29 (15.41).
Solution and Refinement of Crystal Structures for 2-Cl and
4-Cl. Tp′Rh(CNneopentyl)(CH2C(O)CH3)Cl (2-Cl). A well-formed
crystal with approximate dimensions of 0.18 × 0.14 × 0.04 mm3 was
mounted on a glass fiber and placed on a Bruker SMART APEX II
CCD Platform diffractometer under a cold stream of nitrogen at −173
°C. The lattice constraints were obtained from 53,344 reflections with
values of c between 1.73 and 33.14°. Cell reduction revealed a
monoclinic crystal system. Data were collected in accord with the
parameters in the SI. The space group was assigned as P21/n on the
basis of systematic absences and intensity statistics. The structure was
solved using SIR97 and refined using SHELXS-97. All non-hydrogen
atoms were refined with anisotropic displacement parameters. All
hydrogen atoms were placed in ideal positions and refined as riding
atoms with relative isotropic displacement parameters.
Tp′Rh(CNneopentyl)(CH2O-t-Bu)Cl (4-Cl). A well-formed crystal
with approximate dimensions of 0.20 × 0.16 × 0.05 mm3 was mounted
on a glass fiber and placed on a Bruker SMART APEX II CCD
platform diffractometer under a cold stream of nitrogen at −173 °C.
Final cell constants were calculated from the xyz centroids of 3525
strong reflections from the actual data collection after integration. Cell
reduction revealed a monoclinic crystal system. Data were collected in
accord with the parameters in the SI. The space group was assigned as
Cc on the basis of systematic absences and intensity statistics. The
structure was solved using SHELXS-97 and refined using SHELXL-97.
All non-hydrogen atoms were refined with anisotropic displacement
parameters. All hydrogen atoms were placed in ideal positions and
refined as riding atoms with relative isotropic displacement parameters.
Calculation of C−H and Relative Rh−C Bond Strengths. Bond
dissociation energies were calculated as the reaction:
was observed for this byproduct. This byproduct has t1/2 = 2.0 h in
C6D6 at 66.9 °C, suggesting the possibility of activation of some
substituted alkane, which was later confirmed by NMR and GC−MS
analysis to be dimethyl ether in the commercial CH3F. The structure
for this hydride complex was therefore assigned as Tp′Rh-
(CNneopentyl)(CH2OCH3)H (7).
Preparation of Tp′Rh(CNneopentyl)(CHF2)H (8). The method for
preparing 8 was the same as method B for 5, except that CH2F2 was
used as the gas and a longer exchange time of 1 week was required for
completion. Significant amounts of the activation products of the
released carbodiimide (∼71%) were also observed. 1H NMR (400
3
MHz, C6D6): δ −14.095 (dd, JF−H = 11.7 Hz,1JRh−H = 24.9 Hz, 1 H,
3
RhH). 19F NMR (400 MHz, C6D6): δ −11.480 (dddd, JRhH−F = 11.7
2
2
2
2
2
Hz, JRh−F = 16.1 Hz, J
= 54.1 Hz, JF−F = 246.3 Hz, 1 F),
CH2−F
−17.238 (ddd, 2JRh−F = 6.9 Hz, 2JCH −F = 54.1 Hz, 2JF−F = 246.3 Hz, 1
2
F).
Preparation of Tp′Rh(CNneopentyl)(CH2C(O)CH3)Cl (2-Cl). A
solution of 1 (50 mg, 0.073 mmol) dissolved in 1.0 mL of acetone was
placed in an NMR tube sealed with a Teflon cap. This sample was
irradiated for 20 min at −20 °C. The solvent was immediately
removed in vacuo at room temperature. 1.0 mL of carbon tetrachloride
was added and the solution stirred under a nitrogen atmosphere for 1
day. The volatiles were again removed and the yellow solid purified by
chromatography with 5:1 hexane:THF as the eluent. Yellow crystals
were collected (19.2 mg, 45%) by recrystallization in dichloromethane
layered with hexane at room temperature. 1H NMR (500 MHz,
C6D6): δ 0.773 (s, 9 H, C(CH3)3), 2.059 (s, 3 H, pz-CH3), 2.144 (s, 3
H, pz-CH3), 2.151 (s, 3 H, pz-CH3), 2.751 (s, 3 H, pz-CH3), 2.798 (s,
3 H, pz-CH3), 2.956 (s, 3 H, pz-CH3), 2.682 (s, 3 H, COCH3), 2.896
4
4
(d, JRh−H = 14.2 Hz, 1 H, NCH2), 3.202 (d, JRh−H = 13.8 Hz, 1 H,
NCH2), 3.564 (dd, 2JRh−H = 6.8 Hz, 2JH−H = 2.9 Hz, 1 H, RhCH2CO),
3.975 (dd, 2JRh−H = 6.8 Hz, 2JH−H = 2.9 Hz, 1 H, RhCH2CO), 5.559 (s,
1 H, pz-H), 5.584 (s, 1 H, pz-H), 5.623 (s, 1 H, pz-H). 13C{1H} NMR
(500 MHz, C6D6): δ 12.30 (s, 1 C, pz-CH3), 12.59 (s, 1 C, pz-CH3),
12.80 (s, 1 C, pz-CH3), 14.87 (s, 1 C, pz-CH3), 15.13 (s, 1 C, pz-
1
CH3), 15.61 (s, 1 C, pz-CH3), 21.11 (d, JRh−C = 19.9 Hz, 1 C,
RhCH2CO), 26.78 (s, 3 C, CH2C(CH3)3), 31.76 (s, 1 C,
CH2C(CH3)3), 34.72 (s, 1 C, CH3), 56.88 (s, 1 C, RhCNCH2),
108.83 (s, 1 C, pz-CH), 109.02 (s, 1 C, pz-CH), 108.19 (s, 1 C, pz-
CH), 143.05 (s, 2 C, pz-C), 144.57 (s, 1 C, pz-C), 152.95 (s, 1 C, pz-
C), 152.03 (s, 1 C, pz-C), 151.75 (s, 1 C, pz-C), 218.17 (s, 1 C, CO).
Anal. Calcd (found) for C24H38N7BOClRh·(C6H14)0.25: C, 50.10
(50.27); H, 6.84 (6.68); N, 16.04 (16.07).
Preparation of Tp′Rh(CNneopentyl)(CH2CCCH3)Cl (3-Cl). A
solution of 1 (50 mg, 0.073 mmol) dissolved in 1.0 mL of 2-butyne
was placed in an NMR tube sealed with a Teflon cap. This sample was
irradiated for 20 min at −20 °C; 1.0 mL of carbon tetrachloride was
added, and the solution stirred under a nitrogen atmosphere for 1 day.
The volatiles were removed under vacuum, and the yellow solid was
purified by chromatography with 5:1 hexane/THF as the eluent. Red-
orange crystals were collected (36.8 mg, 86%) following recrystalliza-
tion from dichloromethane layered with hexane at room temperature.
1H NMR (500 MHz, C6D6): δ 0.691 (s, 9 H, C(CH3)3), 1.294 (t, 3 H,
CCCH3), 2.107 (s, 3 H, pz-CH3), 2.147 (s, 3 H, pz-CH3), 2.223 (s,
3 H, pz-CH3), 2.563 (s, 3 H, pz-CH3), 2.669 (d, JRh−H = 2.6 Hz, 1 H,
NCH2), 2.782 (s, 3 H, pz-CH3), 3.070 (s, 3 H, pz-CH3), 3.763 (dq,
JRh−H = 13.5 Hz, JH−H = 2.7 Hz, 1 H, RhCH2CC), 4.038 (dq, 2JRh−H
= 13.5 Hz, 2JH−H = 2.9 Hz, 1 H, RhCH2CC), 5.573 (s, 1 H, pz-H),
5.660 (s, 1 H, pz-H), 5.711 (s, 1 H, pz-H). 13C{1H} NMR (500 MHz,
C6D6): δ −3.21 (d, 1JRh−C = 19.0 Hz, RhCH2), 4.92 (s, CH3), 12.27 (s,
1 C, pz-CH3), 12.74 (s, 1 C, pz-CH3), 12.97 (s, 1 C, pz-CH3), 14.41
(s, 1 C, pz-CH3), 14.72 (s, 1 C, pz-CH3), 14.82 (s, 1 C, pz-CH3),
Tp′Rh(CNMe)(R)(H) → Tp′Rh(CNMe)(H) + R
The only simplification used during calculations was replacing
neopentyl isocyanide with methyl isocyanide. We have previously
shown this simplification to have no discernible effect (see ref 23 for
details on the choice of method). Gas-phase structures were calculated
using unrestricted DFT with the M06-2x functional. Calculations were
done using the Gaussian09 package. Light atoms (H through F) were
7003
dx.doi.org/10.1021/ja400966y | J. Am. Chem. Soc. 2013, 135, 6994−7004

Journal of the American Chemical Society
Article
modeled with the 6-31G** basis set. Heavy atoms (Rh, Cl) were
modeled with ECP pseudopotentials of the Stuttgart group and the
basis sets further augmented with d or f polarization functions that
have been optimized by Frenking (Rh α = 1.350; Cl α = 0.640).
Geometries were optimized without constraints. Frequency calcu-
lations were done to check for local minima. Free energies were
calculated at 298.15 K and 1 atm.
(19) (a) Jones, W. D.; Hessell, E. T. J. Am. Chem. Soc. 1993, 115, 554.
(b) Jones, W. D.; Wick, D. D. Organometallics 1999, 18, 495.
(20) (a) Vetter, A. J.; Jones, W. D. Polyhedron 2004, 23, 413.
(b) Vetter, A. J.; Rieth, R. D.; Brennessel, W. W.; Jones, W. D. J. Am.
Chem. Soc. 2009, 131, 10742.
(21) Vetter, A. J.; Rieth, R. D.; Jones, W. D. Proc. Natl. Acad. Sci
U.S.A. 2007, 104, 6957.
(22) Evans, M. E.; Li, T.; Vetter, A. J.; Rieth, R. D.; Jones, W. D. J.
Org. Chem. 2009, 74, 6907.
X-ray crystallographic data have been deposited as CCDC
deposition nos. 919108 and 919109.
(23) Choi, G.; Morris, J.; Brennessel, W. W.; Jones, W. D. J. Am.
Chem. Soc. 2012, 134, 9276.
ASSOCIATED CONTENT
■
(24) Evans, M. E.; Burke, C. L.; Yaibuathes, S.; Clot, E.; Eisenstein,
O.; Jones, W. D. J. Am. Chem. Soc. 2009, 131, 13464.
(25) Hessell, E. T.; Jones, W. D. Organometallics 1992, 11, 1496.
(26) Wick, D. D.; Northcutt, T. O.; Lachicotte, R. J.; Jones, W. D.
Organometallics 1998, 17, 4484.
S
* Supporting Information
Tables of NMR data, kinetic data, X-ray crystallographic data as
CIF files, coordinates and energies for calculated complexes, a
summary of the calculation procedure. This material is available
free of charge via the Internet at http://pubs.acs.org.
(27) Vetter, A. J.; Flaschenriem, C.; Jones, W. D. J. Am. Chem. Soc.
2005, 127, 12315.
(28) Jones, W. D.; Hessell, E. T. J. Am. Chem. Soc. 1992, 114, 6087.
(29) For a discussion of this statement, see: Wilkinson, G. Pure Appl.
Chem. 1972, 30, 627.
(30) Note that earlier published calculations (refs 23 24) did not take
into account that the ΔG⧧ for reductive elimination of the substrate
RH was measured at a different temperature than the ΔG⧧ for
reductive elimination of benzene (298 K).
AUTHOR INFORMATION
■
Corresponding Author
jones@chem.rochester.edu
Notes
The authors declare no competing financial interest.
(31) Siegbahn, P. E. M. J. Phys. Chem. 1995, 99, 12723.
ACKNOWLEDGMENTS
́
(32) (a) Clot, E.; Besora, M.; Maseras, F.; Megret, C.; Eisenstein, O.;
Oelckers, B.; Perutz, R. N. Chem. Commun. 2003, 490. (b) Clot, E.;
Oelckers, B.; Klahn, A. H.; Eisenstein, O.; Perutz, R. N. Dalton Trans.
2003, 4065. (c) Clot, E.; Megret, C.; Eisenstein, O.; Perutz, R. N. J.
́
Am. Chem. Soc. 2009, 131, 7817.
(33) Siegbahn, P. E. M. J. Am. Chem. Soc. 1994, 116, 7722.
(34) Sakaki, S.; Biswas, B.; Sugimoto, M. Organometallics 1998, 17,
1278.
■
We acknowledge the U.S. Department of Energy (Grant DE-
FG02-86ER13569) for their support of this work. We
acknowledge the Center for Integrated Research Computing
at the University of Rochester for providing the necessary
computing systems and personnel to enable the research
presented in this manuscript.
(35) Bryndza, H. E.; Fong, L. K.; Paciello, R. A.; Tam, W.; Bercaw, J.
E. J. Am. Chem. Soc. 1987, 109, 1444.
(36) (a) Bennett, J. L.; Wolczanski, P. T. J. Am. Chem. Soc. 1994, 116,
2179. (b) Bennett, J. L.; Wolczanski, P. T. J. Am. Chem. Soc. 1997, 119,
10696.
(37) Clot, E.; Megret, C.; Eisenstein, O.; Perutz, R. N. J. Am. Chem.
́
Soc. 2006, 128, 8350.
(38) Uddin, J.; Morales, C. M.; Maynard, J. H.; Landis, C R.
Organometallics 2006, 25, 5566.
REFERENCES
(1) Pruett, R. L. Adv. Organomet. Chem. 1979, 14, 1.
■
(2) Bini, L.; Muller, C.; Vogt, D. ChemCatChem. 2010, 2, 590.
̈
(3) Schwartz, J.; Labinger, J. A. Angew. Chem., Int. Ed. Engl. 1976, 15,
333.
(4) Tamaki, A.; Kochi, J. K. J. Chem. Soc., Chem. Commun. 1973, 423.
(5) Reger, D. L.; Culbertson, E. C. Inorg. Chem. 1977, 16, 3104.
(6) Reger, D. L.; Garza, D. G.; Lebioda, L. Organometallics 1992, 11,
4285.
(7) Reger, D. L.; Garza, D. G.; Baxter, J. C. Organometallics 1990, 9,
873.
(8) (a) Janowicz, A. H.; Bergman, R. G. J. Am. Chem. Soc. 1983, 105,
3929. (b) Buchanan, J. M.; Stryker, J. M.; Bergman, R. G. J. Am. Chem.
Soc. 1986, 108, 1537.
(9) Northcutt, T. O.; Wick, D. D.; Vetter, A. J.; Jones, W. D. J. Am.
Chem. Soc. 2001, 123, 7257.
(10) Chirik, P. J.; Day, M. W.; Labinger, J. A.; Bercaw, J. E. J. Am.
Chem. Soc. 1999, 121, 10308.
(11) (a) Craig, P. J.; Green, M. J. Chem. Soc. (A) 1968, 1978.
(b) Wojcicki, A. Adv. Organomet. Chem. 1973, 11, 87.
(12) Hartwig, J. F. Selected Reactions of Metal-Alkyl Complexes. In
Organotransition Metal Chemistry: From Bonding to Catalysis;
University Science Books: Sausalito, 2010; Chapter 3, (3.2.1.4), p 90.
(13) Reger, D. L.; McElligott, P. J. J. Organomet. Chem. 1981, 216,
C12.
(14) Harvey, J. N. Organometallics 2001, 20, 4887.
(15) Vela, J.; Vaddadi, S.; Cundari, T. R.; Smith, J. M.; Gregory, E. A.;
Lachicotte, R. J.; Flashenriem, C. J.; Holland, P. L. Organometallics
2004, 23, 5226.
(16) Aeby, A.; Consiglio, G. Dalton Trans. 1999, 655.
(17) Brookhart, M.; Rix, F. C.; Desimone, J. M.; Barborak, J. C. J. Am.
Chem. Soc. 1992, 114, 5894.
(18) Pellecchia, C.; Pappalardo, D.; Oliva, L.; Zambelli, A. J. Am.
Chem. Soc. 1995, 117, 6593.
7004
dx.doi.org/10.1021/ja400966y | J. Am. Chem. Soc. 2013, 135, 6994−7004