Kinetics, Thermodynamics, and Effect of BPh3 on Competitive C-C and C-H Bond Activation Reactions in the Interconversion of Allyl Cyanide by [Ni(dippe)]

Article abstract of DOI:10.1021/ja037002e

Reaction of [(dippe)Ni(μ-H)]₂ with allyl cyanide at low temperature quantitatively generates the η²-olefin complex (dippe)Ni(CH₂=CHCH₂CN) (1). At ambient temperature or above, the olefin complex is converted to a mixture of C-CN cleavage product (dippe)Ni(η³-allyl)(CN) (3) and the olefin-isomerization products (dippe)Ni(η²-crotonitrile) (cis- and trans-2), which form via C-H activation. The latter are the exclusive products at longer reaction times, indicating that C-CN cleavage is reversible and the crotononitrile complexes 2 are more thermodynamically stable than η ³-allyl species 3. The kinetics of this reaction have been followed as a function of temperature, and rate constants have been extracted by modeling of the reaction. The rate constants for C-CN bond formation (the reverse of C-CN cleavage) show a stronger temperature dependence than those for C-CN and C-H activation, making the observed distribution of C-H versus C-CN cleavage products strongly temperature-dependent. The activation parameters for the C-CN formation step are also quite distinct from those of the C-CN and C-H cleavage steps (larger ΔH? and positive ΔS?). Addition of the Lewis acid BPh₃ to 1 at low temperature yields exclusively the C-CN activation product (dippe)Ni(η ³-allyl)(CNBPh₃) (4). Independently prepared (dippe)Ni(crotononitrile-BPh₃) (cis- and trans-7) does not interconvert with 4, indicating that 4 is the kinetic product of the BPh ₃-mediated reaction. On standing in solution at ambient temperature, 4 decomposes slowly to complex 5, with structure [(dippe)Ni(η ³-allyl)(N≡C-BPh₃), while addition of a second equivalent of BPh₃ immediately produces [(dippe)Ni(η ³-allyl)]⁺[Ph₃BC≡NBPh₃] ^- (6). Comparison of the barriers to π-σ allyl interconversion (determined via dynamic ¹H NMR spectroscopy) for all of the η³-allyl complexes reveals that axial cyanide ligands facilitate π-σ interconversion by moving into the P₂Ni square plane when the allyl group is σ-bound.

Full text of DOI:10.1021/ja037002e

Published on Web 02/25/2004
Kinetics, Thermodynamics, and Effect of BPh3 on Competitive
C-C and C-H Bond Activation Reactions in the
Interconversion of Allyl Cyanide by [Ni(dippe)]
Nicole M. Brunkan,† Donna M. Brestensky,‡ and William D. Jones*,†
Contribution from the Department of Chemistry, UniVersity of Rochester,
Rochester, New York 14627, and Chemistry Department, Box BG,
St. BonaVenture UniVersity, St. BonaVenture, New York 14778
Received July 1, 2003; E-mail: jones@chem.rochester.edu
Abstract: Reaction of [(dippe)Ni(µ-H)]2 with allyl cyanide at low temperature quantitatively generates the
η2-olefin complex (dippe)Ni(CH2dCHCH2CN) (1). At ambient temperature or above, the olefin complex is
converted to a mixture of C-CN cleavage product (dippe)Ni(η3-allyl)(CN) (3) and the olefin-isomerization
products (dippe)Ni(η2-crotonitrile) (cis- and trans-2), which form via C-H activation. The latter are the
exclusive products at longer reaction times, indicating that C-CN cleavage is reversible and the crotononitrile
complexes 2 are more thermodynamically stable than η3-allyl species 3. The kinetics of this reaction have
been followed as a function of temperature, and rate constants have been extracted by modeling of the
reaction. The rate constants for C-CN bond formation (the reverse of C-CN cleavage) show a stronger
temperature dependence than those for C-CN and C-H activation, making the observed distribution of
C-H versus C-CN cleavage products strongly temperature-dependent. The activation parameters for the
C-CN formation step are also quite distinct from those of the C-CN and C-H cleavage steps (larger ∆Hq
and positive ∆Sq). Addition of the Lewis acid BPh3 to 1 at low temperature yields exclusively the C-CN
activation product (dippe)Ni(η3-allyl)(CNBPh3) (4). Independently prepared (dippe)Ni(crotononitrile-BPh3)
(cis- and trans-7) does not interconvert with 4, indicating that 4 is the kinetic product of the BPh3-mediated
reaction. On standing in solution at ambient temperature, 4 decomposes slowly to complex 5, with structure
[(dippe)Ni(η3-allyl)(NtC-BPh3), while addition of a second equivalent of BPh3 immediately produces [(dippe)-
Ni(η3-allyl)]+[Ph3BCtNBPh3]- (6). Comparison of the barriers to π-σ allyl interconversion (determined via
dynamic 1H NMR spectroscopy) for all of the η3-allyl complexes reveals that axial cyanide ligands facilitate
π-σ interconversion by moving into the P2Ni square plane when the allyl group is σ-bound.
Introduction
ovnikov byproduct 2-methyl-3-butenenitrile (2M3BN) in a 2:1
ratio. The unwanted 2M3BN is then isomerized, using the same
Reactions that make and break carbon-carbon σ bonds are
essential to organic synthesis on both the laboratory and the
industrial scale. Although often accomplished by organic
reagents, these transformations may also be effected by orga-
nometallic reagents that cleave C-C σ bonds via insertion of
the transition metal center into the bond.1 The most widely
recognized industrial process involving C-C bond activation
by a transition metal complex is DuPont’s synthesis of adi-
ponitrile (ADN) via hydrocyanation of butadiene, which is
catalyzed by triaryl phosphite-Ni(0) complexes.2
Ni catalyst, to 3PN through a C-CN bond activation route
involving a relatively stable, spectroscopically observable Ni(II)
π-allyl cyanide intermediate (eq 1).2 Further isomerization of
3PN to the terminal cyanoolefin 4-pentenenitrile in the presence
of a Ni(0) catalyst and Lewis acid cocatalyst, followed by a
second anti-Markovnikov addition of HCN to the olefin, yields
ADN. Although the mechanisms of the ADN process and related
Ni-catalyzed hydrocyanation reactions have been investigated
extensively by DuPont3 and others,4 isolation and full charac-
terization of the organometallic Ni complexes involved in the
C-C bond activation (2M3BN-3PN isomerization) step have
not been reported in the literature.
The commercial ADN process begins with Ni-catalyzed
addition of HCN to butadiene, forming the desired anti-
Markovnikov product 3-pentenenitrile (3PN) and the Mark-
(3) (a) Tolman, C. A.; McKinney, R. J.; Seidel, W. C.; Druliner, J. D.; Stevens,
W. R. AdV. Catal. 1985, 33, 1-47. (b) Seidel, W. C.; Tolman, C. A. Ann.
N.Y. Acad. Sci. 1983, 201-221.
† University of Rochester.
‡ St. Bonaventure University.
(4) (a) Goertz, W.; Kamer, P. C. J.; van Leeuwen, P. W. N. M.; Vogt, D.
Chem.-Eur. J. 2001, 7, 1514-1618. (b) Goertz, W.; Keim, W.; Vogt, D.;
Englert, U.; Boele, M. D. K.; van der Veen, L. A.; Kamer, P. C. J.; van
Leeuwen, P. W. N. M. J. Chem. Soc., Dalton Trans. 1998, 2981-2988.
(c) Ba¨ckvall, J. E.; Andell, O. S. Organometallics 1986, 5, 2350-2355.
(d) Taylor, B. W.; Swift, H. E. J. Catal. 1972, 26, 254-260. (e) Keim,
W.; Behr, A.; Lu¨hr, H.-O.; Weisser, J. J. Catal. 1982, 78, 209-216. (f)
Campi, E. M.; Elmes, P. S.; Jackson, W. R.; Lovel, C. G.; Probert, M. K.
S. Aust. J. Chem. 1987, 40, 1053-1061.
(1) For reviews of C-C bond activation by organometallic complexes, see:
(a) Murakami, M.; Ito, Y. In Topics in Organometallic Chemistry; Murai,
S., Ed.; Springer-Verlag: New York, 1999; pp 97-129. (b) Rybtchinski,
B.; Milstein, M. Angew. Chem., Int. Ed. 1999, 38, 870-883. (c) Perthuisot,
C.; Edelbach, B. L.; Zubris, D. L.; Simhai, N.; Iverson, C. N.; Mu¨ller, C.;
Satoh, T.; Jones, W. D. J. Mol. Catal. A 2002, 189, 157-168.
(2) McKinney, R. J. In Homogeneous Catalysis; Parshall, G. W., Ed.; Wiley:
New York, 1992; pp 42-50.
9
10.1021/ja037002e CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 3627-3641
3627

A R T I C L E S
Brunkan et al.
a very dark red THF-d8 solution of the hydride dimer at -78
°C, evolution of H2 gas and a color change to yellow-brown
1
were observed. 31P{1H} and H NMR spectra of the reaction
mixture at -30 °C showed quantitative conversion of the starting
material to the Ni(0) η2-olefin complex (dippe)Ni(allyl cyanide)
(1), producing a pair of doublets at 72.1 and 66.9 ppm with
2JP-P ) 60 Hz in the 31P NMR spectrum (eq 2).
C-CN bond cleavage of nitrile substrates other than 2M3BN
by organometallic species has been studied,5 as has the reverse
reaction of C-CN reductive elimination from transition metal
complexes.6 We recently reported reversible C-CN bond
activation of substituted benzonitriles, cyanopyridines, and
cyanoquinones by the electron-rich (dippe)Ni(0) fragment (dippe
) 1,2-bis(diisopropylphosphino)ethane), which is generated
from the hydride dimer [(dippe)Ni(µ-H)]2 upon loss of H2.7 η2-
Coordination of the ArCN cyano group to Ni(0) yields isolable,
fully characterized (dippe)Ni(η2-NtCAr) complexes, which
undergo reversible C-CN oxidative addition reactions to give
isolable Ni(II) (dippe)Ni(Ar)(CN) products.
Because of their relevance to the ADN process, reactions of
allylic cyanide substrates with [(dippe)Ni(µ-H)]2 were then
examined. Herein, we describe the kinetically competitive C-H
and C-CN bond activation reactions that allyl cyanide, the
simplest of this substrate class, undergoes upon coordination
to (dippe)Ni(0) (work with the more complicated substrate
2M3BN is ongoing). Reaction of allyl cyanide with (dippe)Ni-
(0) in the presence of the Lewis acid BPh3 was also examined,
as literature reports indicate that Lewis acids facilitate C-CN
cleavage in the DuPont phosphite-Ni system.3,8 The availability
of structural and spectroscopic information for nearly all of the
organometallic species involved in the interconversion of allyl
cyanide by (dippe)Ni(0), in both the presence and the absence
of the Lewis acid BPh3, provides a unique opportunity to directly
compare C-H and C-C activation processes, and the effect of
Lewis acids on them, in a single well-defined system. Kinetic
data for the interconversion of these species at different
temperatures also provide insight into the mechanisms of the
fundamentally important and industrially relevant C-H cleav-
age, C-CN cleavage, and C-CN bond formation reactions
observed.
Characterization of the η2-Olefin Complex 1. While
product 1 could not be isolated because it isomerized rapidly
at temperatures above -30 °C (vide infra), it was identified as
an η2-olefin complex on the basis of its NMR data. The 31P
NMR spectrum of 1 is similar to that of the previously reported
(dippe)Ni(0) η2-olefin complex of 3-cyanoquinoline (A; δ 65.9,
d; δ 60.9, d, 2JP-P ) 54 Hz), which was characterized by NMR
spectroscopy and X-ray crystallography.7b The small P-P
coupling constant is not consistent with η2-CtN coordination,
which in (dippe)Ni(ArCN) complexes produces 2JP-P between
63 Hz (electron-deficient ArCN) and 72 Hz (electron-rich
ArCN).7b Because the allyl group of allyl cyanide is more
electron-donating than even the most electron-rich aryl group
of the nitriles previously investigated, a P-P coupling constant
greater than 72 Hz is expected for a Ni(0) η2-CN allyl cyanide
complex, in contrast to the observed data for 1.10 The small
difference in chemical shift between the 31P doublets of 1 (∆δ
5.2 ppm) also resembles A rather than the η2-CN-bonded
(dippe)Ni(ArCN) complexes (∆δ ≈ 12 ppm).7b
The 1H NMR spectrum of 1 is consistent with η2-CdC
coordination of allyl cyanide to Ni. Although one of the
multiplets for the olefin protons could not be distinguished from
those of the four magnetically inequivalent methine protons of
the dippe ligand due to extensive overlap, all resonances in the
complex appear upfield of 2.77 ppm,11 whereas the olefin
protons of free allyl cyanide resonate between 5.2 and 5.8 ppm.
Eight distinct resonances (each a doublet of doublets due to
Results and Discussion
Reaction of [(dippe)Ni(µ-H)]2 with Allyl Cyanide at -78
°C. The hydride dimer [(dippe)Ni(µ-H)]2 acts as a source of
the highly reactive fragment (dippe)Ni(0), as it readily loses
H2 upon exposure to a variety of neutral, often weakly
coordinating ligands and produces (dippe)Ni(0)-ligand com-
plexes.9 Upon mixing allyl cyanide (1 equiv per Ni center) with
i
(5) (a) Miller, J. A. Tetrahedron Lett. 2001, 42, 6991. (b) Taw, F. L.; White,
P. S.; Bergman, R. G.; Brookhart, M. J. Am. Chem. Soc. 2002, 124, 4192-
4193. (c) Churchill, D.; Shin, J. H.; Hascall, T.; Hahn, J. M.; Bridgewater,
B. M.; Parkin, G. Organometallics 1999, 18, 2403.
(6) (a) Yamamoto, T.; Yamaguchi, I.; Abla, M. J. Organomet. Chem. 2003,
671, 179-182. (b) Huang, J.; Haar, C. M.; Nolan, S. P.; Marcone, J. E.;
Moloy, K. G. Organometallics 1999, 18, 297-299. (c) Marcone, J. E.;
Moloy, K. G. J. Am. Chem. Soc. 1998, 120, 8527-8528. (d) Favero, G.;
Gaddi, M.; Morvillo, A.; Turco, A. J. Organomet. Chem. 1978, 149, 395-
400.
(7) (a) Garcia, J. J.; Jones, W. D. Organometallics 2000, 19, 5544-5545. (b)
Garcia, J. J.; Brunkan, N. M.; Jones, W. D. J. Am. Chem. Soc. 2002, 124,
9547-9555.
(8) Tolman, C. A.; Seidel, W. C.; Druliner, J. D.; Domaille, P. J. Organome-
tallics 1984, 3, 33-38.
coupling to P and the Pr methine proton) were observed for
the eight methyl groups of the dippe ligand, indicating that
rotation around the Ni-olefin bond does not occur at -30 °C.
In the -30 °C 13C{1H} NMR spectrum of 1, the olefin carbons
of allyl cyanide are shifted far upfield from their positions in
the free ligand (128.4 and 119.1 ppm) to 38.5 and 33.5 ppm,
consistent with η2-CdC coordination to Ni(0). Both appear as
doublets of doublets due to coupling with the trans and cis
phosphorus atoms of 1 (Table 1). The methylene carbon of allyl
(9) For examples, see: (a) Edelbach, B. L.; Vicic, D. A.; Lachicotte, R. J.;
Jones, W. D. Organometallics 1998, 17, 4784. (b) Edelbach, B. L.; Vicic,
D. A.; Lachicotte, R. J.; Jones, W. D. Organometallics 1999, 18, 4040. (c)
Mu¨ller, C.; Iverson, C. N.; Lachicotte, R. J.; Jones, W. D. J. Am. Chem.
Soc. 2001, 123, 9718-9719.
(10) We have since observed several η2-CtN complexes of allylic nitriles by
31P and 1H NMR spectroscopy at low temperature in THF-d8 solution; all
exhibit P-P coupling constants greater than 72 Hz. For example, 2JP-P
74 Hz for (dippe)Ni(η2-CN-3-pentenenitrile).
)
(11) See Supporting Information for the 1H NMR spectrum.
9
3628 J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004

Competitive C−C and C−H Bond Activation Reactions
Table 1. Selected NMR and IR Data for 2-7
A R T I C L E S
31P{1H} NMR
1H NMR
H-X (Hz)
13C{1H} NMR
JC-P (Hz)
IR
c
m
δ (ppm)
2JP-P (Hz)
m
δ (ppm)
<2.77
J
ass.
m
δ (ppm)
ass.
ν )
CN (cm-1
1
d
d
72.1
66.9
60
m
all H
d
121.5
38.5
33.5
26.4
125.5
43.1
26.4
124.6
40.1
26.2
154.8
89.7
53.7
156.4
93.1
55.5
148.2
111.5
62.79
139.3
117.1
64.7
a
6.7
25.6, 2.2
20.9a
17.0, 4.0
4.6
27.3
3.3
5.6, 2.5
27.2
3.5
10.3
1.6
9.0
12.1
a
8.2
CN
dd
dd
dd
t
d
d
dd
d
d
dCH2
dC(H)CH2CN
CH2CN
trans-2
d
d
70.8
67.8
53
54
m
td
<2.51
1.25
all H
CH3
13CN
2181
2181
2096
2150
2182
2245
2244
2244
7.2, 1.6
1.6
dC(H)CH3
dC(H)CN
13CN
dC(H)CH3
dC(H)CN
13CN
cis-2
d
d
70.2
67.0
m
td
<2.42
1.39
all H
CH3
7.4, 1.6
3
s
s
s
s
80.8
84.2
79.5
83.0
5t
d
4.73
2.97
9.0, 2.4
9.0
Ha
Hb/c
t
t
t
t
t
t
allyl CH
allyl CH2
13CNBPh3
allyl CH
allyl CH2
13CN
allyl CH
allyl CH2
B13CNB
allyl CH
allyl CH2
CNBPh3
dC(H)CH3
dC(H)CNB
CNBPh3
dC(H)CH3
dC(H)CNB
4
5t
d
4.89
3.15
9.2, 2.4
7.2
Ha
Hb/c
5
m
5.00
4.22
2.54
5.00
4.40
2.43
<2.71
1.39
Ha
Hb
Hc
Ha
Hb
Hc
dCH
CH3
br s
s
d
br
s
dd
3
13, 3
(b)
7.2
14.4
14.0
6
m
d
d
12.4, 2.0
trans-7
cis-7
d
d
71.8
69.1
44
44
m
td
5.6, 2.4
d
d
s
d
d
41.4
26.6
122.9
39.1
26.4
30.4
3.8
d
d
71.8
68.7
m
td
<2.55
1.39
dCH
CH3
5.6
29.2
3.6
a Too small to measure. b Seven-line multiplet at room temperature; quintet at higher temperatures. c For 3 and 4, “5t” ) quintet of triplets.
Scheme 1
cyanide also gives a doublet of doublets at 26.4 ppm, while the
cyano carbon is a doublet (JC-P ) 6.7 Hz) at 121.5 ppm, slightly
downfield of the CN of free allyl cyanide (118.0 ppm). Similar
13C NMR shifts have previously been noted on coordination of
acrylonitrile to P2Ni or P2Pt complexes (vide infra).12
Competitive C-H and C-C Bond Activation in 1. When
the THF-d8 solution of 1 was warmed to 20 °C or above, several
new species rapidly grew in at the expense of 1. Two sets of
doublets, the major resonances observed by 31P NMR spectros-
copy, appeared at 70.8 and 67.8 ppm with 2JP-P ) 53 Hz, and
2
at 70.2 and 67.0 ppm with JP-P ) 54 Hz, consistent with the
formation of two new Ni(0) η2-olefin complexes. Independent
synthesis and full characterization of these species (vide infra)
revealed that they are η2-olefin complexes of trans- and cis-
cis-2 is formed during the reaction, and the ratio of trans-2:
cis-2 remains constant throughout (1.35:1 ratio at 40 °C; Figure
1).
crotononitrile (trans-2 and cis-2), formed from 1 by isomer-
ization of the double bond of allyl cyanide into the position
conjugated with the cyano group (Scheme 1). Crotononitrile is
thermodynamically more stable than allyl cyanide by 4-5 kcal
mol-1 and coordinates more strongly to Ni(0) because the
electron-withdrawing CN substituent is bound directly to the
olefin, providing a driving force for this transformation.13
Conversion of 1 to 2 is proposed to occur via the C-H bond
activation mechanism shown in Scheme 1. This pathway
probably involves a reactive Ni(II) hydride intermediate,
although no hydrides were observed by 1H NMR spectroscopy
during the reaction, even at -80 °C. Slightly more trans- than
A third species (3) was also observed in the 31P NMR
spectrum of the reaction as a singlet at 80.8 ppm. This resonance
grew in rapidly, reached a maximum that varied with temper-
ature, and then decreased slowly until it disappeared, leaving 2
as the sole product of the reaction. Figure 1 shows the
distribution of species over time (measured via integration of a
series of 31P NMR spectra) for a typical reaction at 40 °C.14
1H NMR spectrum of the reaction, taken when 3 was at a
maximum, included a quintet of triplets at 4.73 ppm (JH-H
A
)
9.0 Hz, JH-P ) 2.4 Hz, 1H) and a doublet at 2.97 ppm (JH-H
) 9.0 Hz, 4H) that were assigned to 3. These resonances are
typical of a Ni(II) π-allyl complex whose allyl syn and anti
(12) (a) Tolman, C. A.; English, A. D.; Manzer, L. E. Inorg. Chem. 1975, 14,
2353-2356. (b) Pellizer, G.; Lenarda, M.; Ganzerla, R.; Graziani, M. Gazz.
Chim. Ital. 1986, 116, 155-161.
(13) Afeefy, H. Y.; Liebman, J. F.; Stein, S. E. Neutral Thermochemical Data.
In NIST Chemistry WebBook, NIST Standard Reference Database Number
69; Linstrom, P. J., Mallard, W. G., Eds.; National Institute of Standards
and Technology: Gaithersburg, MD 20899 (http://webbook.nist.gov), March
2003.
(14) A similar distribution of species plot (obtained by IR spectroscopy) for
the [P(OEt)3]4Ni-catalyzed addition of HCN to butadiene at 35 °C shows
that the intermediate P2Ni(π-1-methallyl)CN behaves much like 3, except
that it grows in more slowly (reaching a maximum after 5 h of reaction
time) and disappears more rapidly than 3.3
9
J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004 3629

A R T I C L E S
Brunkan et al.
(phosphite) L2Ni(0) fragment [L ) P(O-o-tol)3].12a,17 The methyl
groups of trans- and cis-crotononitrile were assigned as triplets
of doublets (due to coupling with P and dC(H)CH3) at 1.25
and 1.39 ppm, respectively. As in 1, eight distinct doublets of
doublets were observed for the eight methyl groups of the dippe
ligand in each complex, indicating that the crotononitrile ligands
do not rotate rapidly at room temperature.
A yellow crystal suitable for X-ray diffraction yielded the
structure of the major isomer trans-2 (Figure 2). Selected bond
lengths and angles for the complex are given in Table 2. The
olefin carbons C2 and C3 lie approximately in the P2Ni square
plane and are nearly equidistant from Ni at 1.959(2) and 1.949(2)
Å, respectively. The olefin in 2 binds much more symmetrically
to Ni(0) than acrylonitrile (ACN) does in the complex [P(O-
o-tol)3]2Ni(ACN), for which Ni-C distances that differ by 0.10
Å were measured (shorter bond to the CN-substituted carbon;
2.016(10) vs 1.911(12) Å).18 The olefin bond length C2-C3 in
2 is comparable to that of the acrylonitrile complex (1.441(4)
vs 1.46(2) Å), while the CtN bond distance is slightly shorter
(1.151(4) vs 1.20(2) Å).
Figure 1. Distribution of species 1 (9), trans-2 ([), cis-2 (2), and 3 (b)
for the reaction shown in Scheme 1 at 40 °C.
protons are equilibrated on the NMR time scale by rapid π-σ
interconversion of the allyl ligand.15 Thus, 3 was identified as
the Ni(II) C-CN bond activation product (dippe)Ni(π-allyl)-
CN (Scheme 1). Independent synthesis and full characterization
of (dippe)Ni(π-allyl)CN (reported elsewhere,16 although some
data are reproduced in Tables 1 and 2) yielded spectroscopic
data identical to that observed for 3 during the rearrangement
of 1, unambiguously confirming its identity. X-ray diffraction
revealed that 3 has a square pyramidal structure with the cyano
ligand in the apical position, and dynamic 1H NMR experiments
demonstrated that conversion of the allyl ligand from π- to
σ-bound (∆Gq ) 9.3 kcal mol-1 at -68 °C) allows the cyano
ligand to move into the Ni(II) square plane, interconverting “CN-
up” and “CN-down” conformers of 3 and rendering all four iPr
groups of the dippe ligand equivalent (see Supporting Informa-
tion).16
Reaction of [(dippe)Ni(µ-H)]2 with a stoichiometric amount
of a commercially available 1:2 mixture of trans- and cis-
crotononitrile afforded trans- and cis-2 in a 1:2 ratio, with NMR
spectra identical to those of the species formed during rear-
rangement of 1 (eq 3). Observation of an isomer ratio (which
does not change upon heating at 55 °C for several days) different
from the one obtained by isomerization of 1 indicates that
trans-2 and cis-2 do not interconvert under the reaction
conditions. In other words, the C-H bond activation pathway
in Scheme 1 is effectively irreversible, unlike the C-C bond
activation pathway. Thus, the ∼1.35:1 trans-2:cis-2 ratio
produced by allyl cyanide isomerization reflects a small kinetic
preference for activation of one of the diastereotopic -CH2CN
protons of allyl cyanide over the other one, producing more
trans- than cis-crotononitrile. The complexes 2 were also
prepared in a 1:2 ratio on a larger scale by reaction of (dippe)-
Ni(COD) with crotononitrile, and nearly pure (97%) cis-2 was
isolated by fractional recrystallization of the 1:2 product mixture.
Although red crystals of 3 are indefinitely stable under inert
atmosphere, a solution of 3 in THF-d8 was completely converted
to trans- and cis-2 (1.13:1 ratio) upon standing at room
temperature for 5 d, just as 3 is converted to 2 during
rearrangement of 1 (Figure 1). This transformation requires that
the C-C bond cleavage reaction is reversible and that 2 is more
thermodynamically stable than 3 (Scheme 1).
Characterization of the Crotononitrile Complexes 2. The
31P NMR spectra of trans- and cis-2 (Table 1) are consistent
with their formulation as crotononitrile complexes. The decrease
2
in JP-P for these species as compared to 1, which indicates
stronger binding of the olefin to Ni, is particularly convincing
evidence that the CdC bond in 2 is conjugated with the cyano
group (as in the η2-CdC 3-cyanoquinoline complex A, which
has almost exactly the same P-P coupling constant as 2). The
IR spectra of 2 (pure cis or a cis-trans mixture) in THF
contain a single CtN stretching band at 2181 cm-1, 40 cm-1
lower than νCN ) 2221 cm-1 for free crotononitrile. This
decrease in νCN reflects the weakening of the CtN bond by
electron donation from the d10 Ni(0) center into the CtN π*
orbital. The decrease in νCN for crotononitrile upon binding to
the (dippe)Ni(0) fragment is larger than the 32-33 cm-1 change
observed on coordination of acrylonitrile to (Ph3P)2Ni(0) or
[P(O-o-tol)3]2Ni(0) (νCN ) 2195 and 2194 cm-1, respectively)17
and the 26-28 cm-1 decrease observed on binding of trans- or
cis-crotononitrile to (tBuNC)2Ni(0) (νCN ) 2195 and 2193
cm-1),19 suggesting that (dippe)Ni(0) is more electron-donating
1
room-temperature H NMR spectrum obtained for the final
products of the reaction, a mixture of trans- and cis-2, also
supports this assignment. The protons on the Ni(0)-coordinated
double bonds could not be distinguished from the methine
protons of dippe, but all resonances of the complexes appear
upfield of 2.51 ppm,11 while the olefin resonances for free
crotononitrile appear between 5.4 and 6.8 ppm. Similarly large
upfield shifts (1.75-4.01 ppm) for the protons of acrylonitrile
have previously been reported upon coordination to a bis-
(15) Wilke, G.; Bogdanovic´, B.; Hardt, P.; Heimbach, P.; Keim, W.; Kro¨ner,
M.; Oberkirch, W.; Tanaka, K.; Steinru¨cke, E.; Walter, D.; Zimmerman,
H. Angew. Chem., Int. Ed. Engl. 1966, 5, 151-266.
(17) Tolman, C. A.; Seidel, W. C. J. Am. Chem. Soc. 1974, 96, 2774-2780.
(18) Guggenberger, L. J. Inorg. Chem. 1973, 12, 499-508.
(19) Ittel, S. D. Inorg. Chem. 1977, 16, 2589-2597.
(16) Brunkan, N. M.; Jones, W. D. J. Organomet. Chem. 2003, 683, 77-82.
9
3630 J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004

Competitive C−C and C−H Bond Activation Reactions
A R T I C L E S
Table 2. Selected Bond Distances (Å) and Angles (deg) for 2-7
trans-2
cis-7
3a
4
6
1.645(8),b 1.677(9)c
N-B1
C1-B2
C1tN
C2dC3
C3dC4
Ni-C1
Ni-C2
Ni-C3
Ni-C4
Ni-P1
Ni-P2
1.612(5)
1.592(2)
1.549(10), 1.548(10)
1.167(9), 1.138(9)
1.390(4)
1.151(4)
1.441(4)
1.151(4)
1.442(5)
1.146(4), 1.149(4)
1.399(4), 1.401(4)
1.402(4), 1.392(4)
1.994(3), 1.983(3)
2.084(3), 2.078(3)
1.997(3), 1.990(3)
2.090(3), 2.098(3)
2.1916(8), 2.1847(8)
2.1894(9), 2.1899(8)
1.1517(19)
1.399(2)
1.391(2)
1.9835(15)
2.0992(16)
2.0080(16)
2.0823(16)
2.2002(4)
2.1899(4)
1.366(4)
1.959(2)
1.949(2)
1.961(4)
1.933(4)
2.023(2)
2.001(2)
2.025(3)
2.1564(7)
2.1608(7)
2.1434(7)
2.1674(7)
2.1465(11)
2.1750(10)
P1-Ni-P2
C2-C3-C4
C2-Ni-C3
Ni-C1-C2
C2-C1tN
C1tN-B1
NtC1-B2
Ni-C1tN
91.64(2)
93.33(4)
88.72(3), 89.06(3)
117.2(3), 117.6(3)
89.103(16)
118.78(16)
91.00(3)
120.6(3)
43.28(11)
111.18(18)
179.4(3)
43.46(14)
111.6(3)
178.7(4)
174.4(3)
175.12(15)
171.05(14)
176.6(11), 175.3(12)
177.7(13), 175.6(14)
177.7(3), 176.5(3)
a Two molecules in the unit cell. b B2-C1-N1-B1 isomer. c B1-C1-N1-B2 isomer.
Scheme 2
dC(H)13CN carbons of trans- and cis-2, in accordance with
the 25.9 ppm shift of dC(H)CN in the acrylonitrile complex.21
Figure 2. ORTEP drawing of trans-2 showing 30% probability ellipsoids.
and a better π-back-bonder than the Ni(0) centers supported by
aryl-phosphine, phosphite, or isocyanide ligands.
Kinetic Distribution of C-H versus C-C Bond Activation
Products and Their Variation with Temperature. Although
C-C bond activation in 1 is thermodynamically less favorable
than C-H bond activation, the C-C and C-H bond cleavage
reactions in Scheme 1 are kinetically competitive (Figure 1).
The kinetic distribution of C-H activation products 2 versus
C-C activation product 3, defined as the ratio (sum of rate
constants for formation of 2)/(rate constant for formation of 3),
was determined at five different temperatures between 20 and
50 °C as described below. Variation of the rate constants with
temperature yielded the activation parameters ∆Hq and ∆Sq for
C-H and C-C bond activation of 1, as well as C-C reductive
elimination from 3.
A clean 1.13:1 mixture of trans- and cis-13CN-labeled 2 was
obtained by allowing a THF-d8 solution of 13CN-labeled 316 to
stand at room temperature for 112 h (eq 4). The -20 °C 13C{1H}
NMR spectrum of this mixture contained a triplet (JP-C ) 4.6
Hz) for the labeled 13CN of trans-2 at 125.5 ppm and a doublet
of doublets (trans- and cis-P coupling of 5.6 and 2.5 Hz,
respectively) for the labeled 13CN of cis-2 at 124.6 ppm (Table
1).20 These chemical shifts are similar to δCN of 1 and almost
identical to the CN shift of 125.5 observed for Pt(0)-coordinated
acrylonitrile in the complex (Ph3P)2Pt(ACN).12b Like the olefin
1
protons in the H NMR spectra of 2, the olefin carbons of
Values for the rate constants k1 (1 f 3), k-1 (3 f 1), k2 (1
f trans-2), and k3 (1 f cis-2), defined in Scheme 2, were
determined at 40 °C by simulation of the concentration versus
time data shown in Figure 1. The program KINSIM provided a
rough fit to the data, and then a least-squares fit was calculated
crotononitrile are also shifted far upfield upon coordination to
(dippe)Ni(0). The dC(H)CH3 carbons of trans- and cis-2 give
doublets at 43.1 and 40.1 ppm, respectively, in good agreement
with the 42.7 ppm shift observed for dCH2 in [P(O-o-tol)3]2Ni-
(ACN).12a Doublets at 26.4 and 26.2 ppm were assigned to the
(21) Note that, in addition to the small cis C-P coupling constant observed for
these resonances, 20-30 Hz coupling to the trans-P is also expected,
suggesting that the observed doublets each represent one-half of a set of
doublets of doublets, whose other halves are obscured by the THF-d8 solvent
resonance at 25.4 ppm.
(20) Resonances for trans- and cis-2 were assigned by comparison of the 13C{1H}
NMR spectrum of the mixture to the 13C{1H} NMR spectrum of unlabeled,
97% cis-2.
9
J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004 3631

A R T I C L E S
Brunkan et al.
Table 3. Rate Constants for C-H and C-C Activation of 1 at
Various Temperaturesa
temp
(°C)
k1 × 103
(min-1
k-1 × 103
(min-1
k2 × 103
(min-1
k3 × 103
(min-1
[trans-2]/ (k2 + k3)/
)
)
)
)
k2/k3 [cis-2]
k1
20
30
35
9.6(1.2) 0.6(0.2) 15.2(1.3) 10.5(1.0) 1.45 1.48
34.7(2.2) 3.6(0.5) 35.4(1.8) 26.1(1.4) 1.36 1.38
71.5(2.6) 10.5(0.8) 47.0(1.4) 35.8(1.2) 1.31 1.33
2.67
1.77
1.16
1.20
1.01
40 113.4(3.5) 22.0(1.3) 78.2(1.9) 58.0(1.5) 1.35 1.37
50 268(13) 89.1(6.4) 157.5(4.8) 112.9(3.6) 1.40 1.40
a Values in parentheses represent 2σ uncertainty (95% confidence limit).
using the program FITSIM.22 This least-squares fit is represented
by the lines drawn in Figure 1 (points represent experimental
data), and the values k1-k3 are listed in Table 3. Distribution
of species plots such as the one in Figure 1 were also created
for reactions at 20, 30, 35, and 50 °C23 and were fit in the same
manner to provide the rest of the rate constants in Table 3. The
relative rates of formation of trans- versus cis-2 (k2/k3; Table
3, column 6) were consistent with the final ratio of trans- to
cis-2 observed at each temperature (Table 3, column 7),
providing a good check of the validity of the calculated rate
constants. The kinetic distributions of C-H activation products
2 versus C-C activation product 3 were also calculated from
(k2 + k3)/k1 at each temperature (Table 3, column 8). Over the
temperature range investigated, C-C activation becomes in-
creasingly kinetically competitive with C-H activation as the
reaction temperature is raised, until at 50 °C the two processes
occur at almost the same rates ((k2 + k3)/k1 ≈ 1).24
Figure 3. Eyring plot for rate constants k1 ([), k2 (b), k3 (9), and k-1
(2).
Table 4. Activation Parameters for C-H and C-C Activation of 1a
q
q
q
∆H
∆S
∆G at 40 °C
(kcal/mol)
(eu)
(kcal/mol)
k1
18.0(0.3)
27.5(0.5)
13.4(0.2)
13.2(0.2)
-14.8(0.8)
11.7(1.6)
-29.9(0.6)
-31.2(0.7)
22.71
23.75
22.95
23.14
k-1
k2
k3
a Uncertainties in parentheses are σ values calculated using a weighted
nonlinear least-squares analysis.26
the complex is more thermodynamically stable at this temper-
ature (k-1 is very small). At elevated temperatures, 3 forms much
more rapidly and higher concentrations are observed initially
(kinetic effect), but the thermodynamic instability of 3 at these
temperatures causes it to disappear completely in <3 h at 50
°C.
C-C cleavage product 3, however, becomes much less
thermodynamically competitive with C-H activation products
2 as the temperature is raised, because the 1 h 3 equilibrium
shifts increasingly to the left at higher temperatures. In other
words, kinetic and thermodynamic influences on the concentra-
tion of 3 in the reaction solution oppose one another. This
situation arises because the rate constants k vary with temper-
ature to different degrees, as is apparent from the Eyring plot
in Figure 3. k-1 is far more sensitive to changes in temperature
than k1 (making Keq for the 1 h 3 equilibrium markedly
temperature-dependent),25 which in turn varies more with
temperature than the C-H activation constants k2 and k3. Thus,
at room temperature, C-C bond activation occurs more slowly
than C-H bond cleavage, producing less C-C cleavage product
3 than at elevated temperatures (kinetic effect); yet, relatively
high concentrations of 3 remain in solution for days because
The activation parameters for C-C bond cleavage (k1), C-C
bond formation (k-1), and C-H bond cleavage (k2, k3) of allyl
cyanide by (dippe)Ni(0) were calculated from the Eyring plot
(Figure 3, Table 4) using a weighted nonlinear least-squares
analysis.26 The enthalpic cost ∆Hq of C-H bond cleavage
(identical within experimental error for trans- and cis-2 reac-
tions) is less than that of C-C cleavage by about 5 kcal mol-1
.
However, the enthalpic cost ∆Hq of C-C bond cleavage is in
turn far less than that of C-C bond formation, by nearly 10
kcal mol-1. Apparently, the rate-determining step of the C-C
bond-forming process involves significantly more bond-breaking
(metal-allyl and metal-cyanide) than the rate-determining step
for C-C bond cleavage. The entropy of activation ∆Sq for
formation of C-H activation products 2 from 1 is negative and
unexpectedly large (∼ -30 eu), suggesting that degrees of
freedom are significantly reduced in the transition state for
isomerization of Ni-bound allyl cyanide to Ni-bound crotono-
nitrile. By comparison, ∆Sq for C-C cleavage is only one-half
as large, but still negative. In contrast to both of these, ∆Sq for
C-C bond formation is positive, suggesting a less-ordered
transition state.
(22) KINSIM Version 3.3, 1983, was written by B. A. Barshop, Department of
Biological Chemistry, Washington University Medical School, St. Louis,
MO 63110. See: Barshop, B. A.; Wrenn, R. F.; Frieden, C. Anal. Biochem.
1983, 130, 134-145. FITSIM Version 1.63, 1987, was developed by C.
T. Zimmerle. See: Zimmerle, C. T.; Patane, K.; Frieden, C. Biochemistry
1987, 26, 6545-6552. For reviews of kinetic regression analysis methods,
see: (a) Mannervick, B. Methods Enzymol. 1982, 63, 103-139. (b)
Motulsky, H. J.; Ransnas, L. A. FASEB J. 1987, 1, 363-374.
(23) Experiments at temperatures outside this range proved impractical, as
reactions above 50 °C became too rapid to monitor by 31P NMR
spectroscopy and reactions below 20 °C were too slow (>48 h reaction
time) to leave in the NMR probe continuously. Also, at low temperatures,
only small amounts of 3 were formed (<5% at 0 °C), making accurate
integration difficult. Note that slightly poorer fits to the data are obtained
at lower temperatures; see Supporting Information.
Scheme 3 shows mechanistic paths that account for these
differences. From η2-allylcyanide complex 1, π-coordination of
the nitrile to give an η4-intermediate would place the allylic
C-H bond in just the proper position for C-H activation,
directly giving the π-allyl hydride intermediate from this very
(24) The trend toward less C-C activation at low temperatures is also supported
by qualitative data collected at 0 °C (the maximum amount of 3 observed
during the reaction was <5%) and -30 °C (no C-C bond cleavage was
observed).
(25) By contrast, raising the temperature from 35 to 50 °C during the
[P(OEt)3]4Ni-catalyzed addition of HCN to butadiene increases the rates
for both formation and decay of intermediate P2Ni(1-methyl-π-allyl)CN
by the same amount (6 times faster).3a Note that the intermediate forms
via a different route than 3, however, so its formation rate constant is not
expected to be analogous to k1.
(26) The nonlinear least squares program LSM was used to calculate the
activation parameters from the rate constants and their errors. The program
is available free of charge from the author, J. A. Mamczur, Institute of
Physics, Rzeszow University of Technology, Al. Powstancow Warszawy
6, 35-959 Rzeszow, Poland, http://www.prz.rzeszow.pl/∼janand/.
9
3632 J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004

Competitive C−C and C−H Bond Activation Reactions
A R T I C L E S
Scheme 3
ing Lewis acid was proposed to lower the barrier to C-CN
cleavage, accounting for the rate acceleration observed.3a We
investigated the reaction of 1 with the Lewis acid BPh3 and
found that this Lewis acid similarly facilitates C-C cleavage
and completely suppresses C-H activation.
When 1 equiv of crystalline BPh3 was added to a solution of
1 in THF-d8 at -78 °C under inert atmosphere, the Lewis acid
remained largely undissolved. However, a 31P NMR spectrum
of the reaction mixture collected at -50 °C revealed resonances
for both 1 (73%) and a new singlet at 84.2 ppm (27%). In the
1H NMR spectrum, the BPh3 resonances integrated for about
0.25 equiv with respect to dippe and were shifted upfield from
their positions in the free Lewis acid. The BPh3 was then
dissolved by warming the NMR tube to -30 °C and shaking.
A -30 °C 31P NMR spectrum collected immediately afterward
showed quantitative conversion of 1 to the species displaying a
singlet at 84.2 ppm, which was assigned as the Lewis acid
adduct of C-C cleavage product 3, (dippe)Ni(π-allyl)(CtN-
BPh3) (4) (eq 6).
ordered transition state (B). Alternatively, release of the olefin
portion of the substrate to give an η2-nitrile complex could then
lead to cleavage of the C-C bond to give a square planar Ni(II)
σ-allyl cyanide complex that then closes to give π-allyl product
3. This pathway is a direct analogy to the established mechanism
of cleavage of C-CN bonds in aryl nitriles by [(dippe)Ni0],7
which occurs from the η2-CN complex. This transition state (C)
is also ordered, but less so than the transition state for C-H
cleavage (B) because there is a freely dangling vinyl group, so
a smaller negative ∆Sq is expected. By the principle of
microscopic reversibility, if this pathway is indeed operative
for C-CN cleavage, then the same pathway must be followed
for C-CN bond formation. That is, the π-allyl complex 3 must
first rearrange to the σ-allyl cyanide complex and then undergo
reductive coupling of the C-C bond via C. This transition state,
with the freely dangling vinyl group, would have less order than
the π-allyl complex 3, and therefore can accommodate the
positive ∆Sq for the C-CN formation step of the reaction. The
large positive ∆Hq of this step is also consistent with the π-allyl
to σ-allyl rearrangement before the rate-determining step.
A cyclic (η2-olefin)-Ni-CN transition state similar to the one
drawn in Scheme 3 has previously been proposed for Lewis
acid-assisted C-CN cleavage of Ni(0)-bound allyl cyanide.3a
Table 4 also contains a list of ∆Gq values for the system,
calculated at 40 °C. All four values lie within a 1 kcal mol-1
range between the barriers for C-C bond cleavage (22.71 kcal
mol-1) and C-C bond formation (23.75 kcal mol-1), demon-
strating that all reaction pathways available to 1 are similar in
energy.
Characterization of the Ni(II) π-Allyl Cyanide-BPh3
1
Adduct 4. Fluxional behavior was observed in the H NMR
spectrum of 4 at -30 °C (vide infra), but on warming the
solution to 20 °C all 1H resonances became sharp. No free BPh3
appeared in the spectrum; instead, all BPh3 resonances were
shifted upfield by 0.3 (para-H) to 0.5 (ortho-H) ppm, consistent
with coordination of the Lewis acid to the Ni-bound cyano group
of 4. Resonances for a π-allyl ligand undergoing rapid π-π
allyl interconversion, similar to those observed for 3 but shifted
downfield by 0.16-0.18 ppm, also appeared in the spectrum.
A quintet of triplets at 4.89 ppm (JH-H ) 9.2 Hz, JH-P ) 2.4
Hz, 1H) was assigned to the methine proton of the π-allyl ligand,
while the syn and anti protons together yielded a slightly
broadened doublet at 3.15 ppm (JH-H ) 7.2 Hz, 4H). Two sets
of doublets of doublets, shifted slightly upfield (<0.1 ppm) from
their positions in the spectrum of 3, were observed for the methyl
groups on the dippe ligand.
The identity of 4 was confirmed by independent synthesis
(eq 7). Addition of THF solvent at -78 °C to a mixture of 3
and 1 equiv of BPh3, followed by slow warming to -20 °C,
produced a homogeneous orange solution of 4. Evaporation of
the solvent and recrystallization of the residue from toluene at
-30 °C afforded orange crystals suitable for X-ray diffraction.
The structure of 4 is shown in Figure 4, and selected bond
lengths and angles for the complex are given in Table 2. Like
3, species 4 is square pyramidal with the CtN-BPh3 ligand
in the axial position.16 The Ni-C-N-B linkage is nearly linear,
as has been reported for structurally characterized M-CNBPh3
C-C Activation of 1 with BPh3. As described above,
attempts to increase the concentration of C-C relative to C-H
activation products by adjusting the reaction temperature of 1
met with limited success. Therefore, a strategy involving Lewis
acids was applied to shift the reactivity of 1 toward C-C
cleavage. Previously, DuPont researchers reported that, although
C-C activation of allyl cyanide by [P(O-o-tol)3]3Ni at room
temperature is very slow, reactions containing Lewis acids such
as AlCl3 and ZnCl2 proceed instantaneously on mixing to give
the π-allyl Ni(II)CN-Lewis acid adducts shown in eq 5.8
Stabilization of the transition state C by the electron-withdraw-
9
J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004 3633

A R T I C L E S
Brunkan et al.
Figure 4. ORTEP drawing of 4 showing 30% probability ellipsoids.
complexes of Fe,27 Ru,28 Mn,29 and Rh30 (M-C-N and
C-N-B angles range from 169° to 179°). The N-B bond
length of 1.592(2) Å lies within the range documented for these
complexes (1.575-1.608 Å), as does the CtN bond length of
1.1517(19) Å (range ) 1.135-1.163 Å). The CtN bond of 4
is slightly longer than that of 3, as was also reported for (Ph3P)3-
RhCNBPh3 as compared to (Ph3P)3RhCN,30 although this trend
is not consistent with the νCN shift observed for 4 versus 3 in
the IR (vide infra). As in 3, the π-allyl group of 4 is
symmetrically bonded to Ni, with Ni-C and allyl C-C bond
lengths nearly identical to those of 3 (Table 2).
Figure 5. Dynamic 1H NMR spectra of 4 in THF-d8 (* ) residual solvent).
ens the CtN bond.27a,31b This effect should also shorten the
CtN bond, but no decrease in C-N bond distance was observed
by X-ray crystallography.
13CN-labeled 4 was prepared from 13CN-labeled 3, and its
13C{1H} NMR spectrum was collected at -20 °C. A triplet
resonance (2JC-P ) 12.1 Hz) at 156.4 ppm, similar to the 13CN
resonance of 3 but shifted downfield by 1.6 ppm, was observed
for the labeled cyano carbon, consistent with the 1.8 ppm
downfield shift noted for the CN of trans-[FeH(CN)(dppe)2]
on coordination of BPh3.27a Both allyl carbon resonances of 4
are shifted downfield of their analogues in 3, just as the allyl
The IR spectrum of 4 in THF contains a CtN stretching
band at 2150 cm-1, 54 cm-1 higher in energy than νCN for 3
(2096 cm-1). Similar-magnitude increases in νCN have been
reported on coordination of BPh3 to the Ni-bound cyano group
of [P(O-o-tol)3]3Ni(H)(CN) (56 cm-1 shift, from 2128 to 2184
1
proton resonances of 4 are shifted downfield in the H NMR
spectrum (Table 1).
1
Dynamic H NMR spectroscopy of 4 in THF-d8 (Figure 5)
cm-1 3b,8
and on coordination of ZnCl2 to [P(O-o-tol)3]2Ni(1-
)
revealed that π-σ allyl interconversion is less facile in this
complex than in the non-BPh3 analogue 3. Like 3, complex 4
undergoes rapid π-σ allyl interconversion at room temperature,
equilibrating the syn (Hb) and anti (Hc) protons of the allyl
group. As the temperature is lowered, the Hb/Hc doublet at 3.15
ppm broadens and separates into a pair of doublets at 3.51 (Hb)
and 2.76 ppm (Hc). These protons coalesce at a much higher
temperature in 4 (-16 °C) than in 3 (-68 °C), however, and
the barrier to σ-π allyl interconversion calculated for 4 at this
temperature (∆Gq ) 11.7 kcal mol-1) is ∼2.5 kcal mol-1 larger
than ∆Gq for 3 at -68 °C.16 Decreasing the temperature of 4
also results in separation of the dippe methine resonance at 2.20
ppm into two multiplets, and separation of the two dippe methyl
resonances, each a doublet of doublets at 25 °C, into four
(although three of these are coincident at 1.1 ppm; Figure 5).
These data indicate that 4 possesses only a “vertical” mirror
plane of symmetry (containing Ni-C-N-B and the allyl
methine carbon, in the plane of the paper for 4 in eq 8) at the
low-temperature limit and that the axial position of the cyanide
methyl-π-allyl)CN (42 cm-1 shift, from 2168 to 2210 cm-1).3a
Increases of 46-72 cm-1 in the CtN stretching frequency are
also well-documented on coordination of BPh3 to M-CN (M
) Fe, Ru, Mn, Pd)31 and are attributed to the fact that the N
lone pair on the metal-bound cyano group is antibonding in
nature. Coordination of a Lewis acid to this lone pair removes
electron density from CN antibonding orbitals, which strength-
(27) (a) Almeida, S. S. P. R.; da Silva, M. F. C. G.; da Silva, J. J. R. F.; Pombeiro,
A. J. L. J. Chem. Soc., Dalton Trans. 1999, 467-472. (b) Laing, M.; Kruger,
G.; Du Preez, A. L. J. Organomet. Chem. 1974, 82, C40-C42. (c)
Ginderow, D. Acta Crystallogr., Sect. B 1980, 36, 1950-1951. (d) Amrhein,
P. I.; Lough, A. J.; Morris, R. H. Inorg. Chem. 1996, 35, 4523-4525.
(28) Rocchini, E.; Rigo, P.; Mezzetti, A.; Stephan, T.; Morris, R. H.; Lough,
A. J.; Forde, C. E.; Fong, T. P.; Drouin, S. D. J. Chem. Soc., Dalton Trans.
2000, 3591-3602.
(29) Bellamy, D.; Connelly, N. G.; Hicks, O. M.; Orpen, A. G. J. Chem. Soc.,
Dalton Trans. 1999, 3185-3190.
(30) Fernandes, M. A.; Circu, V.; Weber, R.; Varnali, T.; Carlton, L. J. Chem.
Crystallogr. 2002, 32, 273-278.
(31) (a) Ermi, J.; Gyo¨ri, B.; Bakos, A.; Czira, G. J. Organomet. Chem. 1976,
112, 325-331. (b) Pankowski, M.; Bigorgne, M. J. Organomet. Chem.
1983, 251, 333-338. (c) Haines, R. J.; Du Preez, A. L. J. Organomet.
Chem. 1975, 84, 357-367. See also refs 6b, 27a, 28, and 29.
9
3634 J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004

Competitive C−C and C−H Bond Activation Reactions
A R T I C L E S
i
ligand renders the two Pr groups on each P inequivalent.32 At
Scheme 4
the high-temperature limit, however, an additional “horizontal”
plane of symmetry coincident with the P-Ni-P plane, proposed
to arise from rapid interconversion of “CNBPh3 up” and
“CNBPh3 down” conformers of 4 via movement of the CNBPh3
group into the P-Ni-P plane when the allyl group is σ-bound,
must be present (eq 8). This σ-π allyl interconversion mech-
anism is analogous to that proposed for 316 except for the slightly
higher barrier to interconversion in 4, which reflects the
electronic effects of BPh3 coordination to CN.
the appearance of separate, sharp doublet resonances at 4.40
(JH-H ) 7.2 Hz) and 2.43 (JH-H ) 14.4 Hz) ppm for the syn
and anti protons of the allyl group clearly indicated that σ-π
allyl interconversion does not occur at ambient temperature in
this new species (6), unlike in 5. Two new sets of BPh3
resonances, different from the single set observed for 5, also
appeared in the aromatic region of the spectrum. Slow diffusion
of pentane into the reaction mixture yielded a yellow crystal
suitable for X-ray diffraction, which revealed that 6 consists of
a (dippe)Ni(π-allyl) cation paired with the very bulky, nonco-
ordinating anion [Ph3BCtNBPh3]-. The identity of 6 was
confirmed by adding 2 equiv of BPh3 to a red THF solution of
3 (eq 10). The reaction solution turned pale yellow immediately
on mixing, and NMR spectra identical to those of 6 were
observed for the reaction product.
Abstraction of Cyanide from 4 by BPh3. On standing at
room temperature in THF-d8 solution, isolated orange crystals
of 4 were slowly converted to a new species (5) that appeared
as a singlet at 79.5 ppm in the 31P NMR spectrum of the reaction
mixture (eq 9). After 42 days, 31P NMR spectroscopy showed
that a 1:1 mixture of 4 and 5 had been obtained. In the 1H NMR
spectrum, slightly broadened resonances for 4 were accompanied
by broad resonances of 5 for the syn and anti protons of a π-allyl
group undergoing π-σ allyl interconversion at a rate similar
to the time scale of the 1H NMR experiment (estimated ∆Gq )
14.8 kcal mol-1 at +60 °C; see Supporting Information). After
90 days, a 1:2.8 mixture of 4 and 5 was attained.
Species 5 was not isolated, because it was always produced
as part of an equilibrium mixture with 4, but is assigned as the
complex (dippe)Ni(π-allyl)(NtC-BPh3) in which the cyanide
has reversed its coordination geometry. A plausible mechanism
for conversion of 4 to 5 involves dissociation of the [Ph3BNtC]-
ligand from the Ni center of 4, followed by migration of BPh3
from N to C to give the carbon-bound Lewis acid adduct
[Ph3BCtN]-, followed by reattachment to the metal center.
Migration of BPh3 between the C and N termini of a cyanide
ligand has previously been reported to convert M-NtCBPh3
to its linkage isomer M-CtNBPh3 (M ) Pt,34 Rh,35 Ru31c) at
elevated temperatures. In these cases, movement of BPh3 from
C to N is thermodynamically preferred because it allows cyanide
coordination to the metal via a stronger M-C rather than a
weaker M-N σ bond; in the present case, movement in the
opposite direction suggests that the M-N complex 5 is more
thermodynamically stable than the M-C complex 4. The N-B
bond of CtNBPh3 has also been reported to be significantly
more labile than the C-B bond of NtCBPh3,35 which may
kinetically favor conversion of the N-B to the C-B isomer.
Formation of 6 in the presence of 2 equiv of BPh3 presumably
occurs by trapping [Ph3BNtC]- with a second equivalent of
Lewis acid as soon as it dissociates from Ni. The rapid rate of
The same species 5 was also observed on addition of a 1:1
mixture of allyl cyanide and BPh3 (1 equiv of each per Ni center)
in THF-d8 solution to [(dippe)Ni(µ-H)]2 at room temperature.
A 2:1 mixture of 4 and 5 was observed by 31P NMR
spectroscopy, and broad resonances for the fluxional species 5
1
again appeared in the H NMR spectrum. Upon heating at 55
°C, 5 grew in at the expense of 4 until a 1:4 equilibrium ratio
of 4 to 5 was reached. A second equivalent (per Ni center) of
BPh3 was then added to the reaction in an attempt to shift the
4 h 5 equilibrium toward 5 (Scheme 4). A 31P NMR spectrum
of the reaction mixture showed, however, that 4 was completely
converted to a species displaying a singlet at 83.0 rather than
79.5 ppm. In the 1H NMR spectrum, a set of sharp resonances
that coincide with the resonances of the Ni(II) cation [(dippe)-
Ni(π-allyl)]+,33 the precursor used to prepare 3,16 were observed
instead of the broad resonances associated with 5. In particular,
i
(32) Note that the diastereotopic methyl groups of each Pr unit are always
magnetically inequivalent, so at least two types of methyl groups will be
observed in even the most symmetric (dippe)Ni complex. Also, the low-
temperature NMR data do not allow differentiation between fast or slow
rotation of the π-allyl group, although low-temperature data for 6 (vide
infra) suggest that π-allyl rotation does not occur.
(34) Manzer, L. E.; Anton, M. F. Inorg. Chem. 1977, 16, 1229-1231.
(33) Tenorio, M. J.; Puerta, M. C.; Salcedo, I.; Valerga, P. J. Chem. Soc., Dalton
Trans. 2001, 653-657. See also ref 16 for important 1H NMR data.
(35) Carlton, L.; Weber, R. Inorg. Chem. 1996, 35, 5843-5850. See also ref
30.
9
J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004 3635

A R T I C L E S
Brunkan et al.
Figure 7. ORTEP drawing of 6 showing 30% probability ellipsoids. Note
that C1 and N1 are disordered and the B-C-N-B unit is refined in both
orientations using the Shelx SAME instruction.
Figure 6. 11B{1H} and 13C{1H} NMR spectra of 4, 5, and 6.
Scheme 5
boron resonances upon coordination of B(C6F5)3 to the nitro-
gen as opposed to the carbon of a cyanide ligand.36
Characterization of the [(Ph3B)2CN]- Anion and Ni(II)
π-Allyl Cation of 6. The X-ray structure of 6 is shown in Figure
7. The bulky [Ph3BCtNBPh3]- counterion of 6 was previously
prepared by Lippard and was characterized by X-ray crystal-
lography in conjunction with a Mo cation.37 In Lippard’s
structure, disorder of the anion across a center of symmetry in
the crystal precluded resolution of the cyano C and N atoms,
so composite half-carbon, half-nitrogen atoms were used in the
refinement, yielding an average B-N/B-C bond distance of
1.603 Å. Similar disorder was encountered in the anion of 6,
which was modeled as a 50/50 mixture of B1-C-N-B2 and
B1-N-C-B2 isomers. Note that the average of these four
B-N and B-C bond distances is 1.605 Å, in excellent
agreement with the value obtained by Lippard.37 The structure
of [(C6F5)3BCtNB(C6F5)3]-, an even more weakly coordinating
anion developed for metallocene polymerization catalysts, has
also been determined.38 Surprisingly, the B-C distance in the
perfluorinated anion is about 0.3 Å longer than the B-C
distances of 6, despite the electron-withdrawing influence of
the fluoroaryl groups, while the B-N distance is shorter than
in 6. The B-N bond distance of 4 is also shorter than the
average B-N distance in 6 (Table 2).
The structure of the (dippe)Ni(π-allyl) cation of 6 is similar
to that of the closely related analogue [(dippe)Ni(π-2-methyl-
allyl)]+.33 Also, shorter Ni-C and Ni-P bond distances are
observed for 6 than for 3 or 4, in accord with the increased
charge on the metal center (Table 2). The IR spectrum of 6 in
THF contains a CtN stretching band at 2245 cm-1, in good
agreement with the value of 2255 cm-1 reported for Lippard’s
Mo salt (KBr pellet).37
The 13CN-labeled analogue of 6 was prepared from 13CN-
labeled 3, and a room-temperature 13C{1H} NMR spectrum
revealed a large, broad (56 Hz width at half-height) resonance
at 139.3 ppm for the labeled 13CN of 6. The resonance sharpens
6 formation from 3 thus implies that BPh3 migration, rather
than dissociation of [Ph3BNtC]-, is the rate-determining step
in the slow transformation of 4 to 5 (Scheme 5). Finally, the
lower barrier to σ-π allyl interconversion observed for 5 relative
to 6 suggests that the [Ph3BCtN]- ligand of 5 interacts more
weakly with the Ni(II) center than the [Ph3BNtC]- ligand.
Further evidence for the formulation of 5 as the N-bound
isomer (dippe)Ni(π-allyl)(NtC-BPh3) comes from comparison
of the 11B{1H} and 13C{1H} NMR spectra of 4, 5, and 6 (Figure
6). The initial adduct 4 formed from reaction of 1 or 3 with
BPh3 shows a broad (ν1/2 ≈ 3 ppm) downfield singlet in the
11B NMR spectrum due to coordination of boron to the
quadrupolar nitrogen of the cyanide ligand. The 13C NMR
spectrum shows a sharp triplet (J ) 12.1 Hz) for the carbon
attached to nickel coupling to two equivalent phosphorus nuclei.
Isomer 5, however, shows a sharp singlet (ν1/2 ≈ 1 ppm) shifted
to higher field in the 11B NMR spectrum, as the boron is now
bound to carbon. The 13C spectrum now shows a broad
resonance (ν1/2 ≈ 0.5 ppm) for the cyano carbon bound to
quadrupolar boron. As a comparison, salt 6 shows both broad
and narrow singlets in the 11B NMR spectrum, because two
kinds of boron are present, one in each environment, and a
slightly broadened singlet in the 13C NMR spectrum. A recent
publication by Green et al. shows several examples of metal-
cyanide complexes that display downfield shifts and broadened
(36) Vei, I. C.; Pascu, S. I.; Green, M. L. H.; Green, J. C.; Schilling, R. E.;
Anderson, G. D. W.; Rees, L. H. Dalton Trans. 2003, 2550-2557.
(37) Giandomenico, C. M.; Dewan, J. C.; Lippard, S. J. J. Am. Chem. Soc. 1981,
103, 1407-1412.
(38) (a) Lancaster, S. J.; Walker, D. A.; Thornton-Pett, M.; Bochmann, M. Chem.
Commun. 1999, 1533-1534. (b) Zhou, J.; Lancaster, S. J.; Walker, D. A.;
Beck, S.; Thornton-Pett, M.; Bochmann, M. J. Am. Chem. Soc. 2001, 123,
223-237.
9
3636 J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004

Competitive C−C and C−H Bond Activation Reactions
A R T I C L E S
barrier to σ-π allyl interconversion for 4 is higher than for 3
but lower than for 6 implies that the CNBPh3 ligand of 4 is
less strongly coordinated to Ni than the CN ligand of 3, and
therefore less good at stabilizing the square planar σ-allyl isomer
(eq 8). In fact, weakening of the Pt-C bond of a Pt(CN)
complex upon coordination of BPh3 to the cyano group’s
nitrogen atom has previously been documented, lending support
to this analysis.39 Observation of a barrier to σ-π allyl
interconversion for 5 higher than for 4, yet lower than for 6,
likewise suggests that the [Ph3BCtN]- ligand of 5 interacts
weakly with the Ni center, more than in the completely
noncoordinating [Ph3BCtNBPh3]- counterion of 6, although
less than the [Ph3BNtC]- ligand of 4.
Preparation and Decomposition of the Ni(0) Crotononi-
trile-BPh3 Adducts 7. As described above, C-C activation
product 4 is the only species observed when 1 equiv of BPh3
reacts with 1; no traces of the BPh3 adduct of 1 or BPh3-
coordinated counterparts of the C-H activation products 2 were
detected (eq 6). To test whether 4 is the thermodynamic or
kinetic product of the reaction of 1 with BPh3, the BPh3-
coordinated crotononitrile complexes trans- and cis-7 were
prepared by addition of BPh3 to 2 (eq 11). A 31P NMR spectrum
of the reaction products consisted of two sets of doublets, at
71.8 and 69.1 ppm with 2JP-P ) 44 Hz for trans-7, and at 71.8
2
and 68.7 ppm with JP-P ) 44 Hz for cis-7. Isolated yellow
1
crystals (vide infra) of trans- and cis-7 were heated to 55 °C in
THF-d8 solution to investigate whether 7 equilibrates with 4,
but no conversion to 4 was observed (eq 11); at 80-100 °C,
traces of 4 were detected along with many unidentified
decomposition products. These results demonstrate that 4 and
7 do not interconvert under the reaction conditions for eq 6
(-30 °C), making 4 the kinetic product of the reaction of 1
with BPh3. Thus, C-CN cleavage of Ni(0)-coordinated allyl
cyanide is both faster than C-H bond activation and effectively
irreversible in the presence of BPh3, in marked contrast to the
non-BPh3 reaction.
Figure 8. Dynamic H NMR spectroscopy of 6 in THF-d8 (* ) residual
solvent).
and shifts upfield by ∼4 ppm at -50 °C, consistent with
quadrupolar line broadening. Resonances for the methine and
terminal allyl carbons were identified at 117.1 and 64.7 ppm,
significantly downfield of their counterparts in 3 and 4.
1
Dynamic H NMR spectroscopy of 6 in THF-d8 (Figure 8)
revealed that π-σ allyl interconversion is slow for this complex
at room temperature, unlike for 3 and 4. However, spectra of 4
and 6 at the low-temperature limit are similar in that both contain
i
i
two Pr methine and four Pr methyl resonances (overlapping
one another in 6), indicating that the two iPr groups on each P
of dippe are inequivalent (neither complex possesses a “hori-
zontal” mirror plane of symmetry, vide supra). Because no apical
ligand is present in 6 to lower the symmetry of the complex,
the“horizontal” mirror plane must be absent because the π-allyl
group does not rotate at this temperature. As π-σ allyl
interconversion becomes rapid at higher temperatures, the
methine protons coalesce into a single resonance and the methyl
groups coalesce into two nearly coincident doublets of doublets,
consistent with the behavior observed for 3 and 4. Doublets at
4.40 and 2.43 ppm for the syn (Hb) and anti (Hc) protons of 6
also broaden and coalesce (at 75 °C) into a single resonance at
about 3.38 ppm, although the high-temperature limit could not
be reached due to solvent boiling point limitations. Calculation
of the barrier for π-σ allyl interconversion in 6 yielded ∆Gq
) 15.9 kcal mol-1, significantly higher than in 4 (11.7 kcal
mol-1) or 3 (9.3 kcal mol-1), and slightly higher than in 5 (14.8
kcal mol-1).
Characterization of the Ni(0) Crotononitrile-BPh3 Ad-
ducts 7. In the 31P NMR spectrum of 7, the 10 Hz decrease in
P-P coupling constant as compared to 2 (Table 1) suggests
that the Lewis acid-coordinated olefin of 7 binds more strongly
to Ni(0) than does the olefin of 2, perhaps because it can accept
1
more π-electron donation from Ni. The room-temperature H
NMR spectrum of 7 is similar to that of 2, with the addition of
BPh3 resonances shifted upfield from their positions in the free
Lewis acid, consistent with BPh3 coordination to crotononitrile.
As in spectra of 2, resonances for the olefin protons of
crotononitrile could not be distinguished from those of the four
inequivalent methine protons of dippe, but all dippe and
The difficulty of π-σ allyl isomerization in 6 as compared
to 3 or 4 indicates that movement of the cyano ligand into the
P2Ni square plane during the interconversion speeds up the
process considerably, perhaps through a mechanism in which
the CN ligand effectively displaces one end of the π-allyl group
from Ni. Consistent with this interpretation, the fact that the
(39) Manzer, L. E.; Parshall, G. W. Inorg. Chem. 1976, 15, 3114-3115.
9
J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004 3637

A R T I C L E S
Brunkan et al.
quency as in 2 for both trans- and cis-7.40 By contrast,
resonances for both olefin carbons and the cyano carbon of
Lewis acid adduct 4 are shifted downfield from their positions
in the non-BPh3 analogue 3.
Summary and Conclusions
The reaction of [(dippe)Ni(µ-H)]2 with allyl cyanide at low
temperature cleanly produces the η2-olefin complex (dippe)Ni-
(CH2dCHCH2CN) (1). At ambient temperature or above, 1 is
converted to a mixture of C-H cleavage products (dippe)Ni-
(η2-crotonitrile) (cis-2 and trans-2) and the C-CN cleavage
product (dippe)Ni(π-allyl)CN (3). At longer reaction times, 3
is completely converted to 2, indicating that C-C cleavage is
reversible and the Ni-olefin complexes 2 are more thermody-
namically stable than the Ni(π-allyl) cyanide 3.
The rates for the competitive C-H cleavage, C-C cleavage,
and C-C bond formation reactions that occur in this system
were determined as a function of temperature and showed that
the rate constant for the C-CN bond formation step exhibited
a much stronger temperature dependence than the rate constants
for C-H and C-C cleavage. C-C activation becomes increas-
ingly kinetically competitive with C-H activation at elevated
temperatures, producing more 3. The activation parameters for
the cleavage and bond formation steps revealed that ∆Hq is
much larger for C-CN formation than for C-C or C-H
cleavage, while ∆Sq for C-CN coupling is positive rather than
negative.
The reaction of 1 with BPh3 at low temperature yields the
C-C activation product (dippe)Ni(π-allyl)(CNBPh3) (4) rapidly
and exclusively, demonstrating that C-C activation of allyl
cyanide is both faster than C-H activation and irreversible in
the presence of the Lewis acid. This complex shows linkage
isomerism of the cyano group bridging between the nickel and
boron centers. Additional Lewis acid completely removes the
cyanide as [Ph3B-CN-BPh3]-. The (dippe)Ni(crotononitrile)-
BPh3 adducts 7, analogues of the C-H activation products 2,
were not observed in the reaction of allyl cyanide with (dippe)-
Ni(0) in the presence of BPh3, but were prepared independently
and characterized by X-ray crystallography, IR, and NMR
spectroscopy. Conversion of 7 to the C-C cleavage product 4
was not observed, even at high temperatures, indicating that 4
is the kinetic product of the Lewis acid-assisted reaction of allyl
cyanide with Ni(0).
Figure 9. ORTEP drawing of cis-7 showing 30% probability ellipsoids.
crotononitrile resonances appear upfield of 2.71 ppm (Table 1),11
consistent with η2-olefin binding of the nitrile to Ni. Eight
distinct resonances for the eight methyl groups of dippe were
1
observed in the H NMR spectrum, indicating that the Lewis
acid-coordinated olefin of 7 (like the olefin of 2) does not rotate
at room temperature.
Recrystallization from THF/pentane yielded a crystal of cis-7
whose X-ray structure was determined (Figure 9). Selected bond
lengths and angles for the complex are given in Table 2. The
structure of cis-7 is square planar, with all of the atoms in the
C4-C3-C2-C1-N-B unit lying approximately in a plane
perpendicular to the P2Ni plane. The C2-C1-N-B linkage is
nearly linear, although the bulky BPh3 group bends slightly away
from the iPr groups of the dippe ligand. The N-B bond in cis-7
is longer than the corresponding bond in 4, but shorter than the
N-B distance in 6. Surprisingly, coordination of BPh3 to
crotononitrile does not alter the CtN or CdC bond lengths of
the ligand. The only difference between the (dippe)Ni(croto-
nonitrile) cores of 2 and 7 is that the olefin is bound slightly
less symmetrically to Ni(0) in cis-7 than in trans-2, with Ni-
C1 and Ni-C2 bond distances of 1.961(4) Å for the carbon
bound to CNBPh3 and 1.933(4) Å for the carbon bound to the
methyl group. Otherwise, coordination of BPh3 to the crotono-
nitrile ligand of 2 has remarkably little effect on the structure
of the complex.
The IR spectrum of 7 (pure cis-, pure trans-, or a mixture)
contains a single CtN stretching band at 2244 cm-1, 63 cm-1
higher in energy than that of 2 (Table 1). This increase upon
Lewis acid coordination to Ni-bound crotononitrile is greater
than the 54 cm-1 increase observed when BPh3 coordinates to
the Ni-bound cyano group of 3, but smaller than the 74 cm-1
increase observed on coordination of BPh3 to free crotononitrile
(νCN shifts from 2221 to 2295 cm-1). Comparison of the CtN
stretching frequency of 7 to that of free crotononitrile also
reveals that the two substituents bound to crotononitrile in 7,
the π-electron-withdrawing Lewis acid BPh3 and the π-electron-
donating Lewis base (dippe)Ni(0), exert opposite but roughly
equal electronic effects on cyano bond strength, leaving Ni
complex 7 with a CtN bond slightly stronger than that of the
free nitrile.
The barriers to π-σ allyl interconversion (determined from
dynamic 1H NMR experiments) for the π-allyl complexes
studied lie in the order 3 < 4 < 5 < 6. The presence of an
axial ligand in the coordination sphere of Ni clearly facilitates
π-σ allyl interconversion, as it moves into the P2Ni plane when
the allyl ligand is σ-bound and stabilizes the σ-allyl conformer.
Experimental Section
General Methods. All reactions were performed under N2 in a
Vacuum Atmospheres glovebox or using standard Schlenk techniques.
Allyl cyanide and crotononitrile (2:1 cis:trans) were purchased from
Aldrich, degassed, purified by vacuum transfer, and stored in the
glovebox over molecular sieves. BPh3 was purchased from Aldrich,
recrystallized from pentane or hexanes at -30 °C under inert
atmosphere, dried in vacuo overnight, and stored in the glovebox.
In the 13C{1H} NMR spectrum of cis-7, the carbon resonance
of the BPh3-coordinated cyano group is shifted 1.7 ppm upfield
from 13CN of cis-2 (Table 1). The dC(H)CH3 carbon also shifts
upfield by 1.0 ppm for cis-7 and 1.7 ppm for trans-7, while the
dC(H)CNBPh3 carbon apparently resonates at the same fre-
(40) However, the chemical shifts for dC(H)CNBPh3 given in Table 1 are
probably not accurate because one-half of this resonance lies under the
THF-d8 resonance at 25.4 ppm; see ref 21.
9
3638 J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004

Competitive C−C and C−H Bond Activation Reactions
A R T I C L E S
(COD)2Ni and Ph3PdO were purchased from Aldrich and used as
received. The ligand dippe41 and the complexes [(dippe)Ni(µ-H)]241 and
(dippe)Ni(π-allyl)CN (3)16 were prepared by literature protocols. All
solvents, including THF-d8, were distilled or vacuum transferred from
Na/benzophenone and stored in the glovebox.
Routine room-temperature 31P{1H} and 1H NMR spectra were
recorded on a Bruker AMX400 instrument; low- and high-temperature
31P{1H}, 1H, and 13C{1H}NMR spectra were recorded on a Bruker
Avance400 spectrometer. Chemical shifts are given in ppm and
referenced to residual solvent peaks (1H and 13C NMR) or to an external
standard (85% H3PO4, 31P NMR). Complete spectra data are reported
in the Supporting Information. IR spectra were recorded in THF solution
on a Mattson Instruments 6020 GALAXY Series FT spectrophotometer.
Crystallographic data were collected using a Siemens SMART diffrac-
tometer with CCD detection. Elemental analysis was performed by
Complete Analysis Laboratories, Inc., Parsippany, NJ.
(dippe)Ni(η2-CdC-Allyl cyanide) (1). A small stir bar and a very
dark red solution of [(dippe)Ni(µ-H)]2 (15.6 mg, 24.2 µmol) in THF-
d8 were carefully added to a special NMR tube (made in-house, see
Supporting Information) equipped with a sidearm and Schlenk adapter,
while allyl cyanide (4.0 µL, 49.7 µmol) was added to the sidearm using
a 10 µL syringe. The Ni solution and the allyl cyanide were cooled in
separate acetone/dry ice baths, the tube was evacuated to degas both
liquids, and the allyl cyanide was vacuum transferred into the Ni
solution, which turned yellow-brown. The cold bath around the NMR
tube was removed briefly (for <10 s) so that the reaction solution could
be mixed by moving the stir bar up and down in the tube with a magnet.
The reaction solution was degassed to remove H2 (vigorous bubbling
occurs), and was then stirred and degassed several more times until no
more H2 gas was evolved. The stir bar was drawn up into the sidearm
with the magnet, and the NMR tube was flame sealed under vacuum
beneath the sidearm. The tube was transferred directly from the -78
°C bath to the -30 °C NMR probe.
(dippe)Ni(η2-CdC-Crotononitrile) (2). A solution of dippe (191
mg, 0.73 mmol) in 15 mL of THF was added to crystalline yellow
(COD)2Ni (201 mg, 0.73 mmol), forming a yellow solution that was
stirred for 30 min. The solvent was removed in vacuo, and the product
(dippe)Ni(COD) was dried for 30 min to remove free COD. A solution
of crotononitrile (75 mg, 1.12 mmol; 2:1 cis:trans) in 15 mL of THF
was added to the reaction flask, giving a brownish-yellow solution that
was stirred for 30 min. The solvent was removed in vacuo, and the
residue was dried for 30 min; the 31P NMR spectrum of an aliquot
showed that 6% (dippe)Ni(COD) remained. A solution of another 50
mg (0.74 mmol) of crotononitrile in 15 mL of THF was added to the
reaction flask and stirred for 30 min. After removal of the solvent and
drying for 30 min in vacuo, the 31P NMR spectrum of the yellow-
brown oily residue showed complete conversion to 2 (58% cis:42%
trans). Some of the residue was dissolved in (TMS)2O and allowed to
stand at -30 °C, yielding 114 mg of yellow microcrystals (42% cis:
58% trans-2 by 31P NMR). The rest of the residue was dissolved in
THF; slow diffusion of pentane into this solution at -30 °C yielded
47 mg of large, pale yellow crystals (97% cis:3% trans-2 by 31P NMR;
57% total yield). IR (THF) νCN ) 2181 cm-1. Anal. Calcd for C18H37P2-
NNi: C, 55.70; H, 9.61; N, 3.61. Found: C, 55.64; H, 10.09; N, 3.61.
Schlenk adapter, which was evacuated and cooled in a liquid N2 bath.
Next, 30 mL of dry, degassed THF was vacuum transferred into the
reaction flask, and the reaction mixture was thawed in an acetone/dry
ice bath and stirred. Neither reagent dissolved at -78 °C, so the bath
was gradually warmed by removal of dry ice and addition of warm
acetone until a homogeneous orange solution was obtained at -20 °C.
The cold bath was removed, and the solvent was evaporated in vacuo
as the mixture warmed gradually to room temperature. The orange
powder obtained was recrystallized from toluene at -30 °C, yielding
red-orange crystals that were dried in vacuo (22.0 mg, 54% yield). IR
(THF) νCN ) 2150 cm-1. Anal. Calcd for C36H52P2NBNi: C, 68.61;
H, 8.32; N, 2.22. Found: C, 68.60; H, 8.57; N, 2.24.
13CN-Labeled 4. First, 0.6 mL of dry, degassed THF-d8 was vacuum
transferred from a Na/benzophenone still into an evacuated, liquid N2-
cooled NMR tube containing red solid (dippe)Ni(π-allyl) 13CN (25.8
mg, 66.3 µmol) and white solid BPh3 (14.4 mg, 59.5 µmol). The
reaction mixture was then thawed in an acetone/dry ice bath. The NMR
tube was cooled thoroughly, and was then removed briefly from the
bath and shaken several times until a homogeneous orange solution
was obtained. The sample was warmed to room temperature, and a
31P{1H} NMR spectrum taken immediately showed 77% 13CN-4, 12%
byproduct 13CN-6, and 11% dippe oxide impurity. After removal of
the solvent in vacuo, the resulting orange powder was recrystallized
from toluene at -30 °C, yielding red-orange crystals that were dried
in vacuo and found to be 97% pure by 31P NMR (3% 13CN-6).
[(dippe)Ni(π-Allyl)]+[Ph3BCNBPh3]- (6). A solution of BPh3 (50.1
mg, 207 µmol) in 4-5 mL of THF was added to red solid (dippe)Ni-
(π-allyl)CN (3; 40.2 mg, 104 µmol). Upon mixing, the reaction mixture
turned from wine-red to pale yellow. Fluffy pale yellow solid
precipitated immediately upon addition of 5 mL of pentane to the
reaction mixture. The product was isolated by filtration and dried in
vacuo (75.0 mg, 83% yield). Anal. Calcd for C54H67P2NB2Ni: C, 74.35;
H, 7.74; N, 1.61. Found: C, 74.36; H, 7.73; N, 1.61.
13CN-Labeled 6. This was prepared in the same way as the
nonlabeled analogue from 19.9 mg of (dippe)Ni(π-allyl)13CN (51.1
µmol) and 24.7 mg of BPh3 (102 µmol); 32.6 mg of product was
obtained (73% yield).
(dippe)Ni(η2-CdC-Crotononitrile-BPh3): 97% cis-7. A yellow
solution of 97% cis-2 (18.2 mg, 46.9 µmol) in THF-d8 was added to
solid BPh3 (11.4 mg, 47.1 µmol), producing a yellow solution.
Quantitative conversion to cis-7 was observed by 31P NMR. After
evaporation of the solvent, the sample was redissolved in toluene and
a couple drops of THF; slow diffusion of pentane into this solution at
-30 °C produced pale yellow microcrystals (9.2 mg, 31% yield). IR
(THF) νCN ) 2244 cm-1. Anal. Calcd for C36H52P2NBNi: C, 68.61;
H, 8.32; N, 2.22. Found: C, 68.63; H, 8.55; N, 2.31.
(dippe)Ni(η2-CdC-Crotononitrile-BPh3): 94% trans-7. This
was prepared in the same way as 97% cis-2 using 59.0 mg of 2 (152
µmol, 42% cis:58% trans) and 36.4 mg of BPh3 (150 µmol).
Recrystallization from THF/pentane at -30 °C produced 40.9 mg of
yellow microcrystals (70% cis:30% trans by 31P NMR), while a second
crop obtained from the mother liquor yielded 15.0 mg of yellow needles
that were 94% trans-2 by 31P NMR (58% overall yield).
1
Dynamic H NMR of 4 and 5. About 20 mg of orange crystalline
13CN-Labeled 2. Upon standing at room temperature for 112 h, a
red solution of (dippe)Ni(π-allyl)13CN16 in THF-d8 slowly turned
yellow. 31P{1H} NMR spectroscopy revealed that the sample still
contained 3% of the starting material along with the desired 13CN-2
(53% cis:47% trans), so it was heated to 55 °C for 1 h, producing a
solution with <1% (dippe)Ni(π-allyl)13CN.
(dippe)Ni(π-Allyl)(CNBPh3) (4). Red solid (dippe)Ni(π-allyl)CN
(3; 26.0 mg, 67.0 µmol) and white solid BPh3 (15.7 mg, 64.8 µmol)
were combined in a 100 mL round-bottom flask equipped with a
4 was added to a J. Young NMR tube, which was cooled in a liquid
N2 bath. Next, 0.5 mL of THF-d8 was vacuum transferred from a Na/
benzophenone still into the NMR tube, and the solution was thawed in
an acetone/dry ice bath. The NMR tube was thoroughly cooled, and
was then removed briefly from the bath and shaken until a homogeneous
orange solution was obtained. 31P{1H} and 1H NMR spectra were
recorded at 5 °C intervals between -70 and 50 °C. All 31P{1H} NMR
spectra consisted of a sharp singlet, which shifted progressively upfield
as the temperature was raised [3.8 ppm difference between δ(-70 °C)
1
and δ(50 °C)], while the H NMR spectra varied with temperature as
shown in Figure 5. Additional 1H NMR spectra were collected at -16
and -17 °C to precisely identify the temperature at which the syn and
(41) Supporting Information from: Vicic, D. A.; Jones, W. D. J. Am. Chem.
Soc. 1997, 119, 10855-10856.
9
J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004 3639

A R T I C L E S
Brunkan et al.
anti protons on the terminal carbon of the π-allyl group coalesce (-16
°C). After several months, 4 converted to 5, and another set of VT
NMR spectra were recorded (see Supporting Information).
the experimental data in Microsoft Excel (Figure 1; smooth lines; see
also Supporting Information).
Formation of 4 from 1 and BPh3 at Low Temperature. A small
stir bar and a very dark red solution of [(dippe)Ni(µ-H)]2 (15.9 mg,
24.7 µmol) in THF-d8 were added to a special NMR tube equipped
with a sidearm and Schlenk adapter (see Supporting Information), and
solid BPh3 (11.9 mg, 49.1 µmol) was added to the sidearm. A THF-d8
solution of allyl cyanide (5.0 µL, 62.2 µmol) was vacuum distilled
into the tube, turning the Ni solution yellow-brown. The reaction
solution was mixed and degassed as described above for 1, until no
more H2 was evolved. The solution was then frozen in liquid N2, and
the NMR tube was removed from the Schlenk line and tilted so that
the BPh3 in the sidearm fell into the tube. The reaction mixture was
thawed in an acetone/dry ice bath, but only some of the solid dissolved
at -78 °C, even after mixing with the stir bar. The stir bar was drawn
up into the sidearm with the magnet, and the NMR tube was flame
sealed under vacuum below the sidearm. The tube was transferred
directly from the -78 °C bath to the -50 °C NMR probe. 31P and 1H
NMR spectra were collected at -50 and -30 °C; the sample was then
removed from the probe, shaken once so that the BPh3 dissolved, and
returned to the probe. Further spectra were collected at -30, -10, and
20 °C, as described in the text.
Dynamic 1H NMR of 6. About 20 mg of pale yellow 6 was
dissolved in 0.5 mL of THF-d8 and was added to a J. Young NMR
1
tube. 31P{1H} and H NMR spectra were recorded at 10 °C intervals
between -20 and 30 °C, at 5 °C intervals between 30 and 100 °C, and
at 110 and 120 °C. All 31P{1H} NMR spectra consisted of a singlet
that shifted progressively upfield as the temperature was raised [1.1
ppm difference between δ(-20 °C) and δ(120 °C)], while the 1H NMR
spectra varied with temperature as shown in Figure 8.
Distribution of Species Plots (Figure 1). Complex 1 was generated
at low temperature in a special flame-sealed NMR tube as described
above (see Supporting Information), except that Ph3PdO was also added
to the reaction as an internal standard, dissolved in THF-d8 along with
1
[(dippe)Ni(µ-H)]2. 31P and H NMR spectra were recorded at -30 °C
to verify the purity of 1. The NMR tube was removed from the probe
and returned to an acetone/dry ice bath, and the probe was warmed to
the reaction temperature (20, 30, 35, 40, or 50 °C). The magnet was
preshimmed at this new temperature using a “dummy” sample. The
NMR tube was transferred directly from the dry ice/acetone bath into
a water bath at the reaction temperature (the moment of transfer is
considered the reaction start time, t ) 0), where it was allowed to
equilibrate for 10-15 s. The tube was transferred to the NMR probe,
the shims were adjusted quickly, and a variable-delay kinetics program
was initiated. The program collected 31P NMR spectra over 3 (50 °C
reaction) to 36 h (20 °C reaction). Integration of these spectra afforded
the percents distribution of species 1-3, which were plotted in
Microsoft Excel as shown in Figure 1.
Formation of 5 and 6 from [(dippe)Ni(µ-H)]2, Allyl Cyanide, and
BPh3 at Room Temperature. A solution of allyl cyanide (3.1 µL,
38.5 µmol) and BPh3 (9.4 mg, 38.8 µmol) in THF-d8 was added to
solid [(dippe)Ni(µ-H)]2 (12.1 mg, 18.8 µmol). H2 gas was immediately
evolved, and the solution turned dark yellow-brown. A 31P NMR
spectrum taken after 20 min showed 66% 4 and 34% 5. After the
mixture was allowed to stand at room temperature for 18 h, was heated
at 55 °C for 9 h, and was heated at 80 °C for 16 h, 5 had increased to
84% at the expense of 4. 5/6 increased to 91% on addition of another
∼5 mg of BPh3; yet another ∼5 mg of BPh3 resulted in complete
conversion to 6. Recrystallization from THF/pentane at -30 °C yielded
a yellow crystal of 6 for X-ray diffraction.
Two minor impurities, a broad singlet at ∼78 ppm (1-2%) and a
sharp singlet at 90.9 ppm (e5%), were also observed in the 31P NMR
spectrum of the reaction but are not included in Figure 1. The former
was identified as (dippe)Ni(π-allyl)Cl by comparison of a mass
spectrum of the reaction mixture to the mass spectrum of an authentic
sample, generated by reaction of [(dippe)Ni(µ-H)]2 with allyl chloride.
The chloride ion needed to generate this side product is an impurity in
the starting material [(dippe)Ni(µ-H)]2 (prepared from (dippe)NiCl2 and
superhydride).41 The δ 90.9 species disappeared from the reaction
mixture concomitant with formation of pale yellow crystals in the NMR
tube, which were identified by X-ray diffraction as the decomposition
product (dippe)Ni(CN)2.42 This product has also been observed during
C-CN cleavage reactions of (dippe)Ni(ArCN) complexes.7
Experimental Details of X-ray Structural Determinations. Single
crystals of 2, 4, 5, 6, and cis-7 were mounted under Paratone-8277 on
glass fibers and immediately placed in a cold nitrogen stream at -80
°C on the X-ray diffractometer. The X-ray intensity data were collected
on a standard Siemens SMART CCD Area Detector System equipped
with a normal focus molybdenum-target X-ray tube operated at 2.0
kW (50 kV, 40 mA). A total of 1321 frames of data (1.3 hemispheres)
were collected using a narrow frame method with scan widths of 0.3°
in ω and exposure times of 30 s/frame using a detector-to-crystal
distance of 5.09 cm (maximum 2θ angle of 56.54°) for all of the
crystals. The total data collection time was approximately 13 h. Frames
were integrated with the Siemens SAINT program to 0.75 Å for all of
the data sets. The unit cell parameters for all of the crystals were based
upon the least-squares refinement of three-dimensional centroids of
>5000 reflections.43 Data were corrected for absorption using the
program SADABS.44 Space group assignments were made on the basis
of systematic absences and intensity statistics by using the XPREP
program (Siemens, SHELXTL 5.04). The structures were solved by
using direct methods and refined by full-matrix least-squares on F 2.45
For all of the structures, the non-hydrogen atoms were refined with
anisotropic thermal parameters (except 6, as noted below) and
hydrogens were included in idealized positions giving data:parameter
ratios greater than 10:1. There was nothing unusual about the solution
or refinement of any of the structures, with the exception of 6. The
CN group in the anion was modeled as being disordered 50/50 (a slight
Determination of Rate Constants via Simulation of Distribution
of Species Plots. The mechanism shown in Scheme 2 and the
experimental data for one reaction temperature were input into the
program KINSIM.22 Values for the rate constants k1, k-1, k2, and k3
were entered manually, a simulation was performed, and the calculated
concentration versus time curves were superimposed on the experi-
mental data. The goodness of fit was evaluated by eye, and the rate
constant values were adjusted until a reasonable fit of calculated to
experimental data was obtained. The output from KINSIM was then
input into the program FITSIM,22 which used a nonlinear regression
analysis method to vary the rate constants until the best fit to the
experimental data was obtained. The output of FITSIM included the
rate constants and their standard deviations σ (Table 3), as well as
calculated concentration versus time data, which was plotted along with
(42) Because (dippe)Ni(CN)2 precipitates from solution, integration of the 31P
NMR resonance at 90.9 ppm does not measure the full extent of
decomposition to this species. Addition of an internal standard (Ph3PdO)
to the reaction mixture, however, showed that e5% of the total amount of
dippe present at the beginning of the reaction had disappeared from solution
by the end, providing an upper limit of ∼10% for the amount of (dippe)-
Ni(CN)2 formed in the reaction. Also, note that less decomposition was
observed at low temperatures as compared to that observed at high
temperatures.
(43) It has been noted that the integration program SAINT produces cell constant
errors that are unreasonably small, because systematic error is not included.
More reasonable errors might be estimated at 10× the listed value.
(44) The SADABS program is based on the method of Blessing, see: Blessing,
R. H. Acta Crystallogr., Sect. A 1995, 51, 33-38.
2
(45) Using the SHELX95 package, R1 ) (∑||Fo| - |Fc||)/∑|Fo|, wR2 ) {∑[w(Fo
- Fc2)2]/∑[w(Fo2)2]}1/2, where w ) 1/[σ2(Fo2) + (a‚P)2 + b‚P] and P )
[f ‚(maximum of 0 or Fo2) + (1 - f)‚Fc2].
9
3640 J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004

Competitive C−C and C−H Bond Activation Reactions
A R T I C L E S
difference in thermal parameters was noted if a single orientation was
chosen, see Supporting Information Figure S-13), using the Shelx
SAME instruction to keep the two B-N-C geometries similar (see
Supporting Information p S-48). The C and N atoms were kept isotropic
in the refinement. Positional parameters for all atoms, anisotropic
thermal parameters, all bond lengths and angles, as well as fixed
hydrogen positional parameters are given in the Supporting Information
for all of the structures.
and N atoms by X-ray diffraction, in this case refinement of
the structure as the Ni-N-C-B isomer 5 clearly gives a better
model for the bridging atoms than refinement as the Ni-C-
N-B isomer 4 (see Supporting Information).
Supporting Information Available: Complete NMR data for
1, 2, 3, 4, 6, and 7; a diagram of the special apparatus for the
1
preparation of 1 in an NMR tube; H NMR spectra of cis-2,
Acknowledgment. Acknowledgment is made to the U.S.
Department of Energy, grant #FG02-86ER13569, for their
support of this work, and to the Donors of the American
Chemical Society Petroleum Research Fund, for partial support
of this research. We thank Christine Flashenriem for obtaining
the structure of compound 5.
trans-2, cis-7, and trans-7; VT NMR data for 3, 4, 5, and 6;
distribution of species plots (similar to Figure 1) showing
experimental and calculated data for the C-H and C-CN
activation of 1 (Scheme 1) at 20, 30, 35, 40, and 50 °C; X-ray
crystallographic data for 2, 4, 5, 6, and cis-7, including tables
of positional parameters for all atoms, anisotropic thermal
parameters, all bond lengths and angles, and fixed hydrogen
positional parameters (PDF and CIF). This material is available
free of charge via the Internet at http://pubs.acs.org.
Note Added in Proof. In addition to the NMR evidence for
the orientation of the CN group in 5, the X-ray structure of a
single crystal obtained from a toluene solution of 5 is consistent
with this assignment. The structure is isomorphic and isostruc-
tural with that of 4. While it can be difficult to distinguish C
JA037002E
9
J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004 3641