Synergy between Experimental and Computational Chemistry Reveals the Mechanism of Decomposition of Nickel-Ketene Complexes

Article abstract of DOI:10.1021/jacs.6b08897

A series of (dppf)Ni(ketene) complexes were synthesized and fully characterized. In the solid state, the complexes possess η²-(C,O) coordination of the ketene in an overall planar configuration. They display similar structure in solution, except in some cases, the η²-(C,C) coordination mode is also detected. A combination of kinetic analysis and DFT calculations reveals the complexes undergo thermal decomposition by isomerization from η²-(C,O) to η²-(C,C) followed by scission of the C=C bond, which is usually rate limiting and results in an intermediate carbonyl carbene complex. Subsequent rearrangement of the carbene ligand is rate limiting for electron poor and sterically large ketenes, and results in a carbonyl alkene complex. The alkene readily dissociates, affording alkenes and (dppf)Ni(CO)₂. Computational modeling of the decarbonylation pathway with partial phosphine dissociation reveals the barrier is reduced significantly, explaining the instability of ketene complexes with monodentate phosphines.

Full text of DOI:10.1021/jacs.6b08897

Subscriber access provided by La Trobe University Library
Article
Synergy Between Experimental and Computational Chemistry Re-
veals the Mechanism of Decomposition of Nickel–Ketene Complexes
Nicholas D. Staudaher, Atta M. Arif, and Janis Louie
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.6b08897 • Publication Date (Web): 30 Sep 2016
Downloaded from http://pubs.acs.org on October 1, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 10
Journal of the American Chemical Society
Synergy Between Experimental and Computational Chemistry Re-
veals the Mechanism of Decomposition of Nickel–Ketene Complexes
1
2
3
4
5
6
7
Nicholas D. Staudaher, Atta M. Arif, and Janis Louie*
Department of Chemistry, the University of Utah, 315 S 1400 E, Salt Lake City, Utah 84112.
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ABSTRACT: A series of (dppf)Ni(ketene) complexes were synthesized and fully characterized. In the solid state, the com-
plexes possess η2-(C,O) coordination of the ketene in an overall planar configuration. They display similar structure in
solution, except in some cases, the η2-(C,C) coordination mode is also detected. A combination of kinetic analysis and
DFT calculations reveals the complexes undergo thermal decomposition by isomerization from η2-(C,O) to η2-(C,C) fol-
lowed by scission of the C=C bond, which is usually rate limiting and results in an intermediate carbonyl carbene com-
plex. Subsequent rearrangement of the carbene ligand is rate limiting for electron poor and sterically large ketenes, and
results in a carbonyl alkene complex. The alkene readily dissociates, affording alkenes and (dppf)Ni(CO)2. Computational
modeling of the decarbonylation pathway with partial phosphine dissociation reveals the barrier is reduced significantly,
explaining the instability of ketene complexes with monodentate phosphines.
Introduction
The decomposition of Ni–ketene complexes to Ni–CO
complexes is particularly intriguing considering the lack
of Ni–CO intermediates in many reaction mechanisms
that likely involve Ni–ketene complexes. That is, for-
mation of Ni–CO would prevent product formation in
these reactions. Thus, it may be possible that Ni–CO
complexes form reversibly in some cases or, more likely,
Ni–ketene complexes with higher stability have rates of
reactions that are faster than rates of decomposition to
Ni–CO. Clearly, a deeper understanding regarding the
conversion of Ni–ketene complexes to Ni–CO complexes
could aid in the development of new Ni–mediated chem-
istries of ketenes. Herein we report a detailed evaluation
of the structure and mechanism of decomposition of
(dppf)Ni(ketene) complexes.
Ni–Ketene complexes are highly probable reaction in-
termediates in the carbonylation1 of diazo compounds,2
Danheiser Benzannulations,3 [2+2+2] cycloadditions of
diynes and ketenes,4 and, importantly, the Fisher–
Tropsch reaction.5 As such, some effort has gone into de-
veloping methods for synthesizing Ni–ketene complexes
and, furthermore, understanding their fundamental reac-
tivity. Unfortunately, effective and general syntheses to
these complexes are still lacking. For example, only a
handful of Ni–ketene complexes (e.g., complexes 1-7) have
been successfully prepared to date (Figure 1) and complex
76 has not been fully characterized. Furthermore, many of
these complexes are unstable at ambient conditions
thereby hampering detailed reactivity studies.
To further complicate matters, Ni–ketene complexes
tend to decompose to form catalytically incompetent Ni–
CO complexes. Miyashita and co-workers showed that
the decomposition of complex 1a, which is stable at -30
˚C, at room temperature7 affords (PPh3)2Ni(CO)2 as well
as styrene and but-2-ene-2,3-diyldibenzene as organic by-
products (Figure 1). They propose the decomposition pro-
ceeds by an isomerization from η2-(C,O) to η2-(C,C) fol-
lowed by scission of the C=C bond, forming a carbonyl
carbene complex. In contrast to the instability of com-
plexes 1-3, compounds 4-68 are stable at room tempera-
ture. Although the thermal decomposition of complexes
59 and 610 were not evaluated, complex 4 decomposes over
24 h at 80 ˚C.
Despite the lack of information on Ni–ketene complex-
es, Ir– and Rh–ketene complexes have been studied in
much more depth and lend insight into the decomposi-
tion of Ni–ketene complexes. In fact, coordination
through both the C=O11 and C=C12 bonds of a ketene have
been observed in Ir complexes. Furthermore, the scission
of the C=C bond to form an Ir carbonyl carbene complex
has been shown to be reversible and has been studied in
depth computationally.13 (Figure 1c).
Figure 1: Known Ni–Ketene Complexes and Proposed
Decomposition Pathway
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 10
Results
Synthesis of (dppf)Ni(ketene) Complexes. Our ini-
8ca
8cb
8cc
8cd
8d
iBu
iBu
F
76
67
75
1
2
3
4
5
6
7
8
H
tial efforts focused on the synthesis of a library of Ni–
ketene complexes that are stable at room temperature but
decompose upon heating. Since all known Ni–ketene
complexes have phosphines as supporting ligands, we
began by screening the reaction of a variety of monoden-
tate and bidentate phosphine ligands with Ni(COD)2 and
butyl phenyl ketene (eq 1). Not surprisingly, the use of a
monodentate phosphine such as PPh3 did not afford a
long-lived Ni–ketene complex at room temperature as
determined by 31P{1H} NMR and IR spectroscopy. Indeed,
carbonyl peaks, 2002 and 1944 cm-1, were observed and
are consistent with the formation of decomposition prod-
uct (PPh3)2Ni(CO)2.14
iBu Me
iBu OMe 37
iPr
H
H
80
78
8e
sBu
Crystallography. Solid state structures for a variety of
complexes are shown in Figure 2, and selected bond dis-
tances and lengths are in Table 2. The structure of 8a is
planar about the Ni center, and the ketene is bound
through the C=O bond. The phenyl ring containing C(9)
is nearly in plane with the ketene fragment, with a O(1)-
C(1)-C(2)-C(9) dihedral angle of -4.6(4)˚, indicating the p
orbitals of the ketene and the arene are in conjugation.
The other ketene phenyl ring, which contains C(3), is in a
π-stacking18 interaction with the ring containing C(15) on
the dppf fragment. The centroid-centroid distance is
3.644 Å and the planes are skewed by 7.38˚.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Various bidentate phosphine ligands were also exam-
ined. To our dismay, reactions run with dppe, dppp,
dppb, or dppbenzene with Ni(COD)2 and diphenylketene
also did not afford the desired Ni–ketene complexes. In-
stead, a mixture of LNi(COD) and L2Ni species were ob-
The structure of 8b is, unfortunately, disordered. The
unit cell contains four complexes, with random orienta-
tion of each ketene fragment relative to the unit cell, i.e. a
given unit cell may have all butyl chains down, and an-
other all up, or any combination thereof. As such, 8b ap-
pears to have C2 symmetry about the line containing the
Fe and Ni atoms. Thermal ellipsoids of some carbon at-
oms have therefore been omitted from Figure 2b for clari-
ty, and reliable bond lengths and angles for the metallao-
xirane cannot be obtained. Nevertheless, the X-ray struc-
ture clearly demonstrates that the ketene is C=O bound;
the phenyl ring is oriented away from the dppf fragment
and almost in plane with the C=C bond.
31
served by P{1H} NMR spectroscopy as we have seen with
Xantphos.15 The formation of these species were con-
firmed by omitting ketene from the reactions, which led
to identical 31P{1H} NMR spectra. Furthermore, in the re-
action with dppe, only a broad singlet at 44.6 ppm, which
is consistent with (dppe)2Ni, was observed.16 The for-
mation of bis-chelated species is detrimental to Ni–ketene
complex formation because for every L2Ni species, an
equivalent of Ni(COD)2 is left over which reacts rapidly
with ketene to form nickel carbonyl compounds.8
The structures of 8e, 8ca, 8cb, and 8cc show similar
features in that they are planar, bound through the C=O
fragment of the ketene, and have the aromatic ring in
plane with the ketene. Interestingly, the structures of 8ca,
8cb, and 8cc reveal that the O(1)-C(1)-C(2)-C(7) dihedral
angle increases with more electron rich aromatic rings,
indicating slightly better orbital overlap with electron
poor ketenes. Also, the O(1)-Ni-C(1) angle decreases
slightly from 8ca to 8cc, indicating tighter binding of the
C=O fragment with electron poor ketenes. The fact that
aryl alkyl ketenes have the arene oriented away from the
Ni center indicates that the orbital overlap between the
arene and the ketene is energetically more favorable than
the π-stacking interaction in the structure of 8a. Finally,
the P(1)-Ni (trans to O) bond length is slightly shorter
than the P(2)-Ni (trans to C(1)) bond, revealing a modest
trans effect. The Ni-O bonds are shorter than the Ni-C
bonds, indicating tighter binding of the O atom than the
C.
Spectroscopy. The 31P{1H} NMR spectrum of 8b con-
tains two doublets at 42 and 17 ppm with a coupling con-
stant of 23 Hz, consistent with a planar Ni complex with
cis phosphines and an unsymmetrical π substituent that
does not rotate freely. The aromatic region of the 1H NMR
spectrum is complicated, but clearly shows four signals
between 4.3–3.6 ppm, the cyclopentadienyl region, con-
sistent with a dppf ligand containing mirror symmetry
Gratifyingly, the reaction between dppf and Ni(COD)2
formed (dppf)Ni(COD) quantitatively, observed as a sole
31
singlet at 34 ppm in the P{1H} NMR spectrum; (dppf)2Ni
was not observed.17 Furthermore, when phenyl butyl ke-
tene was added to a solution of Ni(COD)2 and dppf that
was allowed to equilibrate for 60 seconds, clean formation
of Ni–ketene 8b occurred as indicated by the disappear-
ance of the singlet at 34 ppm and the appearance of two
doublets at 42 and 17 ppm, consistent with inequivalent
phosphines of a Ni–ketene complex (vide infra). Butyl
phenyl ketene could be substituted with other alkyl aryl
ketenes (eq 2) under similar reaction conditions to afford
a series of (dppf)Ni(ketene) complexes that were easily
recrystallized from pentane (Table 1).
Table 1. Yields of Ketene Complexes
Complex R1
R2
H
H
Yield (%)
8a
8b
Ph
nBu
85
79
ACS Paragon Plus Environment

Page 3 of 10
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Figure 2. ORTEP diagrams of (dppf)Ni(ketene) complexes. Ellipsoids are set at 30% probability.
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Complex
8a
P1-Ni
P2-Ni
Ni–O1
Ni–C1
O1-C1
C1-C2
O1-Ni–C1
40.59(7)
43.6(4)
P1-Ni–P2
105.60(2)
108.13(8)
105.18(2)
O1-C1-C2-C7
-4.6(4)
-19(1)
2.1557(6)
2.2252(6) 1.866(1)
1.875(2)
1.862(5)
1.878(2)
1.887(2)
1.886(2)
1.878(2)
1.297(3)
1.383(14)
1.292(2)
1.293(2)
.1292(2)
1.291(2)
1.350(3)
1.291(11)
1.352(3)
1.352(2)
1.354(3)
1.350(3)
8b
2.1832(15) 2.1833(15) 1.862(5)
8e
2.1488(7)
2.1551(5)
2.1569(6)
2.1569(6)
2.2151(7)
1.861(2)
1.851(1)
40.42(7)
40.46(6)
40.41(7)
40.28(7)
-7.1(4)
8ca
2.2293(5)
104.86(2) -2.0(3)
8cb
8cc
2.2306(6) 1.855(1)
2.2277(6) 1.871(1)
104.78(2)
106.34(2)
3.0(3)
6.2(3)
Table 2. Selected bond lengths, angles, and torsions of (dppf)Ni(ketene) complexes.
about the plane containing the Fe and P atoms. In addi-
tion, the alkyl region contains signals corresponding to an
n-butyl chain. The 13C{1H} spectrum contains a doublet at
168.7 (J = 48 Hz) corresponding to the ketene O=C=C car-
bon and a doublet of doublets at 79.3 ppm (J = 7, 4 Hz)
consistent with the O=C=C carbon. The IR spectrum of
8b contains only one carbonyl peak at 1624 cm-1, con-
sistent with a C=O coordinated ketene.
stretch at 1597 cm-1 correspond to a (dppf)Ni–ketene
complex where the ketene is C=O coordinated and the
aromatic ring is oriented away from the dppf fragment (A,
Figure 3). This orientation matches what is seen in the
solid state by X-ray crystallography. Based on the similari-
ty in chemical shifts and coupling constants, the doublets
at 40.2 and 17.0 ppm in the 31P{1H} NMR spectrum and the
smallest set of peaks in the 1H NMR spectrum likely corre-
spond to the coordination mode where the ketene is still
C=O coordinated but the aromatic ring is oriented toward
the dppf fragment and the isobutyl chain is distal to the
Ni center (B, Figure 3). Finally, we attribute the singlet
that appears at 23.5 ppm in the 31P{1H} NMR, the second
Interestingly, the spectral data for 8ca are much more
complicated. The 31P{1H} NMR spectrum (Figure 3) dis-
plays two doublets at 40.8 and 17.3 ppm (J = 22 Hz) as well
as two much smaller doublets, each slightly upfield of the
first set of doublets at 40.2 and 17.0 ppm (J = 21 Hz) and a
singlet at 23.5 ppm (Figure 3), indicative of three coordi-
nation modes of the ketene. The 1H NMR spectrum is also
consistent with three structural isomers. The Cp region
displays this behavior more clearly than the aromatic or
alkyl regions. There are three sets of four signals, with
integrations 25:4:2; one each of the largest and second
largest peaks overlap. The IR spectrum shows two car-
bonyl peaks at 1597 and 1743 cm-1. The largest signals in
1
set of peaks in the Cp region of the H NMR, and the car-
bonyl stretch at 1743 cm-1 to the coordination mode where
the ketene is C=C coordinated (C, Figure 3), which is of-
ten the case in Ir and Rh ketene complexes.11-12,19 The
13C{1H} NMR spectrum of complex 8ca lacks the signal to
noise to see all the peaks for the three coordination
modes. Only one carbonyl peak is observed, a ddd at 169.1
ppm (JPC = 41, 7 Hz; JFC = 1 Hz), which we attribute to the
coordination mode observed in the solid state (A). The
O=C=C carbon gives rise to a doublet of doublets at 76.7
the 31P{1H} NMR and H NMR spectra and the carbonyl
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
ppm (JPC = 8, 6 Hz). However, the carbon meta to fluorine
Page 4 of 10
1
2
3
4
5
6
7
8
displays a set of three doublets between 117 and 114 ppm,
one of each corresponding to each coordination mode. To
our knowledge, this is the first example of multiple coor-
dination modes of ketenes observed simultaneously, alt-
hough it is similar to the coordination of α,β-unsaturated
aldehydes and ketones to CpRe(NO)(PPh3)+.20
Table 3. Kinetics of decomposition of (dppf)Ni(ketene) complex-
es.
Entry Complex R1
R2
H
Kobs (x 10-4 /s)
1.0377(5)
0.771(15)
0.997(81)
5.02(26)
3.57(32)
1
8b
nBu
nBu-d9
iBu
2
3
4
5
6
7
8b-d9
8ca
8cb
8cc
8cd
8d
H
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
F
iBu
H
iBu
Me
OMe
H
iBu
2.69(26)
2.40(9)
iPr
8cb
0.00
-0.25
-0.50
-0.75
-1.00
-1.25
-1.50
-1.75
8cc
8cd
y = 2.31σp - 0.006
R2 = 0.9978
Figure 3: 31P{1H} spectrum and coordination modes of 8ca.
Kinetics of Decomposition. The rates of decomposi-
1
tion of 8b-e were measured by H NMR at 100 ˚C in tolu-
ene-d8 (eq 3). The reaction is first order overall and there-
fore first order in ketene complex with half-lives between
0.48–2.5 h, indicating these complexes are quite stable to
thermal decomposition compared to 1a. A 1:1 mixture of
cis- and trans-alkene products was observed in the de-
composition of 8b. Furthermore, the kinetic isotope effect
was investigated by monitoring the decomposition of 8b
and its analogue with a fully deuterated n-butyl chain, 8b-
d9 (Table 3, entries 1 and 2). A kH/kD = 1.35 is observed,
consistent with secondary kinetic isotope effects, and
indicates that in the rate determining step the C-H bond
is not broken and that the O=C=C carbon undergoes a
change in hybridization. Increasing steric bulk of the alkyl
chain from n-butyl to i-butyl (entry 4) decreases the sta-
bility of the complex, but increasing the steric size of the
chain to isopropyl (entry 7) gave a rate constant in-
between the n-butyl and i-butyl. Electronic effects were
also investigated with isobutyl alkyl chains. Increasing the
donating character from R2 = H to Me and to OMe (en-
tries 4-6) increases the stability of the complex, and these
three data points produce a near perfect Hammett plot
with ρ = 2.31 vs σp (Figure 4). This data suggests that nega-
tive charge is built up on the O=C=C carbon of the ketene
in the rate determining step of the reaction. However,
complex 8ca (entry 3) is an outlier of this relationship,
indicating the mechanism of decomposition for this com-
plex is different or has a different rate determining step.
Unfortunately, complex 8e was not kinetically competent.
Significant line broadening was observed within the first
few time points, indicative of paramagnetic product for-
mation.
8ca
-0.3
-0.2
-0.1
0.0
0.1
0.2
σp
Figure 4. Linear free relationship for decomposition of electron
neutral and rich (dppf)Ni(aryl isobutyl ketene) complexes.
Calculation of the Reaction Coordinate. The mech-
anism of decomposition of complex 8b was studied by
DFT calculations to gain further insight into the struc-
tures and energies of the intermediates and transition
states (Figure 5). We found the BP86 functional and a
split basis set consisting of tzvp for Ni, P, and the O=C=C
fragment and svp for all other atoms reproduced the crys-
tal structure of 8cb, and this level of theory was used for
optimizations and frequency calculations in the study of
the decomposition of 8b. BLYP/6-31G* gave an energy of
C=C coordinated complex 9 that was slightly higher than
8b, consistent with our observation of this species spec-
troscopically for 8ca, and the energy of TS2 (which our
experimental data indicates is rate limiting, vide infra)
that aligned well with the experimental activation energy.
Several higher levels of theory were evaluated for single
point energy calculations, and they failed to accurately
predict these energies (see SI for further computational
details and benchmarking). Complex 8b isomerizes from
C=O coordination to C=C coordination (9) through TS1.
The C=C bond is then cleaved, leading to carbonyl car-
bene complex 10 through TS2. The energy of this transi-
tion state, 28.7 kcal/mol agrees with the experimental
activation energy of 28.8 kcal/mol, and this step has the
ACS Paragon Plus Environment

Page 5 of 10
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. Calculated reaction coordinate for the decomposition of 8b. Optimizations were carried out with BP86/tzvp-svp and single point
energies were located with BLYP/6-31G*. Units are in kcal/mol relative to 8b. Numbers in parenthesis are for formation of the trans al-
kene, where formation of the cis alkene is shown.
largest individual barrier (26.8 kcal/mol) along the calcu-
lated pathway. previous report has found that
phosphine arm in complex 9a is only 1.9 kcal/mol. Fur-
thermore, phosphine dissociation from the carbonyl car-
bene complex 10 has a barrier of only 1.0 kcal/mol (to
TS6) and is actually downhill in energy by 4.4 kcal/mol
(to 10a). The transition states for hydride transfer with
partial phosphine dissociation are 35.0 and 34.6 kcal/mol
relative to 8b for the cis and trans alkenes, respectively,
and the conformation of the carbonyl carbene complex
that leads to the cis alkene is 1.5 kcal/mol higher in energy
than that leading to the trans alkene (figure 7).
A
(dtbpe)Ni(η2-(C,C)-O=C=CH2) undergoes C-C scission
with a ΔG‡ of 41.6 kcal/mol and a ΔG of 38.8 kcal/mol.21
These numbers are at odds with our theoretical and ex-
perimental work, albeit (dtbpe)Ni(O=C=CH2) is signifi-
cantly different than our complexes and that work was
benchmarked to the dissociation of CO from Ni(CO)4.
Subsequent to the decarbonylation event, the carbene
rearranges to an alkene via hydrogen transfer in TS3. The
barrier for this step is 22.0 kcal/mol for the cis alkene and
24.7 kcal/mol for the trans alkene. Furthermore, the con-
formation of 10 leading to the trans alkene is 3.2 kcal/mol
higher in energy than that leading to the cis alkene. These
barriers for hydrogen migration are less than the barrier
for decarbonylation by 2.1–4.8 kcal/mol for the trans and
cis alkenes, respectively. Nevertheless, these transition
states represent the second highest individual barrier and
the highest point on the reaction coordinate.
Discussion
Resting State of Multiple Coordination Modes of
the Ketene. The spectral and crystallographic data of our
ketene complexes indicate that the C=O coordination
mode with the ketene’s aryl group oriented away from the
Ni atom is the thermodynamic minimum. In the solid
state, the aromatic group of the ketene is oriented away
from the Ni atom and the alkyl group is toward the
(dppf)Ni fragment. The spectral data of 8ca indicate the
existence of two other coordination modes that have
similar energies to the isomer observed in the solid state.
The first of these is also CO coordinated, with the aryl
and alkyl groups interchanged such that the alkyl group is
oriented away from the Ni atom and the arene is in the π-
stacking interaction with one of the dppf phenyl rings (B,
Figure 3). Our calculations indicate this coordination
mode is 2.6 kcal/mol higher in energy for complex 8b.
The third isomer has the ketene coordinated via the C=C
bond (C, Figure 3). Isomer C gives a singlet in the 31P{1H}
NMR, indicating that the ketene is freely rotating. The
calculated energy of 9 relative to 8b (1.9 kcal/mol) also
corroborates the assignment of our spectral data to this
coordination mode. Furthermore, the calculated barrier
for isomerization of 8b to 9 of 15.7 kcal/mol agrees well
with the experimental barrier of 14 kcal/mol for the isom-
We also investigated carbene dissociation of 10 to 13,
which is strongly uphill in energy and therefore not rele-
vant to the decomposition pathway. Complex 11 then un-
dergoes an exergonic alkene dissociation to form three-
coordinate (dppf)Ni(CO), which is highly reactive and
explains the observation of (dppf)Ni(CO)2 by IR in sam-
ples subjected to kinetic analysis.
In addition, we investigated the possible role of partial
phosphine dissociation in each step. Dissociation of one P
atom from the C=O coordinated complex 8b was found to
be endergonic by 22.2 kcal/mol. Given the lower energy
required to reach TS1, we believe dissociation of a phos-
phine from 8b is irrelevant to the mechanism. In contrast,
phosphine dissociation from C=C coordinated complex 9
may be thermodynamically and kinetically competent
(Fig. 6). The barrier for decarbonylation with one P atom
coordinated (9a) is 20.3 kcal/mol, significantly less than
the experimental activation energy of 28.4 kcal/mol.
However, the barrier for re-association of the tethered
1
erization of 1a based on line broadening in the H spec-
trum at low temperature.7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. Role of partial phosphine dissociation for C=C coordinated ketene complexes, carbonyl carbene complex, and decarbonylation.
Energies are in kcal/mol relative to the energy of 8b.
lies in-between sp3 and sp2, while in complex 10, this car-
bon is a carbene, which is sp2 hybridized. Furthermore,
the value of ρ (2.31) indicates significant buildup of nega-
tive charge on the O=C=C carbon during the rate-
determining step for complexes 8cb-8cd. Mulliken analy-
sis of intermediates associated with decarbonylation and
hydride transfer shows significant buildup of negative
charge on the O=C=C carbon atom during decarbonyla-
tion and not during hydride migration. This carbon atom
has a -0.19 charge in C=C coordinated complex 9, -0.36 in
carbonyl carbene complex 10, and -0.38 in carbonyl al-
kene complex 11. The buildup of charge on this carbon
during decarbonylation can be easily rationalized since it
gains a lone pair. Thus, the combination of KIE and linear
free energy relationship (LFER) analysis, aided by DFT
calculations, indicate decarbonylation is rate-limiting in
most cases. We benchmarked our single point energies
accordingly, and the fit is excellent. Applying the rates
Figure 7. Reaction paths for hydride migration, leading to cis and
trans alkenes. Numbers in parenthesis are for the trans alkene.
Energies are in kcal/mol relative to 8b.
derived from the DFT calculations to the steady state ap-
proximation (eq 4) for an equilibrium between C=O and
C=C coordinated ketene complexes followed by scission
of the C=C bond (eq 5) gives a rate constant of 1.28 x 10-4
Hz, which is in excellent agreement with experimental
rate of 1.034 x 10-4 Hz. This indicates that our computa-
tional model fits the kinetic data very well and that de-
carbonylation is rate limiting for complex 8b. Further-
more, Mulliken analysis reveals that 8b and 9 are essen-
tially Ni(II) species, whereas 10 is nearly Ni(0). Decar-
bonylation is therefore a reductive process with regard to
the Ni center, and we predict that more electron donating
phosphines will stabilize Ni–ketene complexes even more
effectively than dppf.
Evidence for Rate Limiting Decarbonylation. Our
calculations indicate that the decarbonylation step has
the highest barrier for an individual step in the pathway,
and that the hydride migration is the highest point on the
pathway thereby restricting the possible rate-determining
step to these two options. The kinetic isotope effect ob-
served for 8b and 8b-d9 (1.35) rules out C-H bond scission
in the rate determining step, and is consistent with sever-
al β and γ secondary KIEs associated with rehybridization
of the O=C=C carbon during the rate-determining step.
Mayer bond order analysis suggests a change of hybridi-
zation occurs for the O=C=C carbon during decarbonyla-
tion: the difference in C=C bond order in complex 9,
which is 0.948, and the free ketene, which is 1.564, indi-
cates significant back bonding into the C=C bond and
that the hybridization of the O=C=C carbon of complex 9
ACS Paragon Plus Environment

Page 7 of 10
Journal of the American Chemical Society
of accelerated decarbonylation and decelerated hydride
1
2
3
4
5
6
7
8
9
migration to the point where hydride migration, rather
than decarbonylation, is rate-limiting.
Another outlier of the trends associated with ketene
complex decomposition is complex 8d. Though the elec-
tronic character between complexes 8b, 8cb, and 8d are
the same, the steric bulk of the alkyl substituent steadily
increases (i.e., n-butyl < i-butyl < i-propyl) yet the rate of
decomposition both increases then decreases for this se-
ries of complexes. That is, complex 8cb decomposes ~5
fold faster than complex 8b; yet complex 8d, which one
would expect to have an even higher rate of decomposi-
tion than complex 8cb, decomposes ~2 fold slower than
complex 8cb (i.e., rates of decomposition display the fol-
lowing trend: 8b < 8cb > 8d). Along these lines, complex
8e would be predicted to have a slower rate of decompo-
sition than 8cb and 8d. However, kinetic analysis of
complex 8e was unsuccessful; significant amounts of par-
amagnetic Ni species were produced. It is possible that
the increased steric bulk of the ethyl of the s-Bu group,
relative to the methyl of the i-Pr group, is sufficiently
large enough to destabilize complex 9 such that homolyt-
ic cleavage of the Ni-C bond and radical reactions.
Mechanism of Hydrogen Atom Migration. Follow-
ing decarbonylation, the carbene ligand rearranges to an
alkene. To do so, the β hydrogen must migrate to the α
carbon. A close examination of transition state TS3-cis
(Figure 8) shows the migrating hydrogen atom is orthog-
onal to the nodal plane of the π-bond that is formed.
Thus, the hydrogen transfer involves orbital overlap be-
tween the carbene sp2 orbital that contains the lone pair
and the C-H σ* as well as overlap between the filled C-H σ
and empty carbene p orbitals. Consequently, the transfer
involves a hydride rather than a proton or an H•. Howev-
er, from our computational model it is difficult to say if
this step occurs with or without phosphine dissociation.
Application of the Curtin-Hammett principle to the ener-
gies associated with hydride transfer with both phosphine
associated (vide supra) gives a cis:trans ratio of 49:1. A
ratio of 1:2 is predicted with partial phosphine dissocia-
tion, albeit the barriers for hydride migration with one
phosphorus coordinated to Ni are unrealistically high. It
is difficult to be certain where the error in our computa-
tional model is: either the barriers for hydride transfer
with partial phosphine dissociation or the ΔΔG‡ for the
cis and trans alkene isomers without partial phosphine
dissociation have been overestimated.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Our calculations suggest that a deviation in the rate-
limiting step for the decomposition of complex 8d, as well
as 8ca, occurs, albeit for two different reasons. In these
cases, hydrogen migration is rate determining. Inspection
of complex 10 (the carbonyl carbene complex of 8b) (Fig-
ure 9) reveals that this intermediate has the orbital over-
lap necessary for hydride migration (vide supra) in its
lowest energy conformation. The Ni–C-C-H dihedral of
complex 10 is -122.84˚, similar to that of TS3-Cis, which
has a dihedral of -81.33˚, a difference of less than an anti–
gauche isomerization. On the other hand, the carbonyl
carbene complex of 8d has a Ni–C-C-H dihedral of -9.90˚,
indicating that this complex must undergo significant
rotation about the C-C bond to adopt the orbital overlap
necessary for hydride transfer. Furthermore, rotation of
the isopropyl group is significantly more energetically
demanding than for an n- or iso-butyl group. We there-
fore propose that the hydrogen migration is rate limiting
with large alkyl groups (such as those in 8d) as well as
electron withdrawing groups (such as those in 8ca).
Figure 8: TS3-Cis and simplified orbital interactions of carbonyl
carbene complex leading to hydride transfer. Ferrocene and most
H atoms omitted for clarity.
Hydride Transfer is Rate-Limiting in Some Cases.
Complex 8ca, which possesses an electron withdrawing p-
fluorophenyl on the ketene, is a clear outlier of the LFER
(Figure 4). Extrapolation of the LFER gives a rate of de-
carbonylation of 7.06 x 10-4 Hz, significantly faster than
the observed rate. The reason decomposition is much
slower for this complex is likely twofold. First, the elec-
tron withdrawing group allows for greater stabilization of
the buildup of negative charge on the O=C=C carbon dur-
ing decarbonylation thereby enhancing the rate of this
step. Furthermore, the hydride migration step is attenu-
ated by electron withdrawing groups. That is, the nucleo-
philicity of the carbene sp2 orbital containing the lone
pair and the electrophilicity of the empty carbene p or-
bital are both attenuated by the electron withdrawing
group. We therefore believe 8ca defies the LFER because
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 10
were able to synthesize and characterize a library of
1
(dppf)Ni[O=C=C(Ar)(R)] complexes. Crystallographic
data reveals that the ketene is C=O coordinated with the
ketene’s aryl group distal to the Ni atom. Spectroscopic
data reveal that the coordination mode observed in the
solid state predominates in solution; however, C=O coor-
dination with the aryl group oriented toward the (dppf)Ni
fragment and C=C coordination are both observed. A
combined experimental and theoretical approach was
used to elucidate the mechanism of decomposition of
these complexes which involves a rate determining step of
either decarbonylation or hydride transfer. That is, decar-
bonylation was determined to be the rate-determining
step for 8b through kinetic isotope effect and Mayer bond
order analyses. Similarly, a LFER for 8cb-8cd combined
with Mulliken analysis suggested decarbonylation and not
hydride migration was rate limiting for these complexes.
Electron poor and sterically large ketenes decompose
with rates that do not fit the trends for the complexes
with rate limiting decarbonylation. Our computational
model provided a rationale for why the hydride transfer is
attenuated and becomes rate limiting in these cases.
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 9: a: 10, the carbonyl carbene complex of 8b, b: carbonyl
carbene complex of 8d.
Decarbonylation with Partial Phosphine Dissocia-
tion is Relevant for Monodentate Phosphines. De-
spite the lower barrier for the decarbonylation step with
one phosphine associated, partial phosphine dissociation
is clearly not relevant to the decomposition of ketene
complexes with bidentate phosphines. Applying the
steady state approximation rate equation (eq 4) to the
reaction in eq 6, where an equilibrium exists between the
C=C coordinated complex with either one or two phos-
phines associated, followed by decarbonylation gives a
kobs of 1.83 x 10-5 Hz, which is ~5 fold slower than the ex-
perimental rate. Nevertheless, this pathway could be
highly relevant for ketene complexes with monodentate
phosphines, which are known to be quite unstable. When
the phosphine is not tethered to the complex (eq 7) phos-
phine dissociation is entropically favored thereby increas-
ing the population of species B. Furthermore, the steady
state approximation for this pathway (eq 8) is distinct
from the steady state approximation for a bidentate
phosphine, and has the concentration of phosphine in the
denominator. The phosphine concentration is very low,
and the overall rate of decarbonylation with one phos-
phine associated should increase greatly compared to
bidentate phosphines. This explains the inherent instabil-
ity of Ni ketene complexes with monodentate phosphine
ligands.
Importantly, our calculations revealed that the Ni cen-
ter is reduced during decarbonylation. We believe that
more electron rich phosphines should further stabilize
Ni–ketene complexes and are currently investigating this
hypothesis. Furthermore, decarbonylation is more facile
with one phosphine associated, explaining the instability
of known Ni–ketene complexes with monodentate phos-
phines.
This report bridges the gap between the instability of
the known Ni–ketene complexes and their use as inter-
mediates in synthesis. We believe that knowledge of the
mechanism of decomposition of Ni–ketene complexes,
particularly the fact that bidentate phosphines stabilize
these complexes, will lead to improvement of known re-
actions of Ni–ketene complexes and aid in the develop-
ment of further reactions that use these complexes as
intermediates.
ASSOCIATED CONTENT
The Supporting Information is available free of charge on the
ACS Publications website at DOI:
Synthesis and characterization of (dppf)Ni(ketene)
complexes (NMR, and X-ray data), NMR/IR scale ex-
periments with other phosphines, kinetic analyses,
computational data.
AUTHOR INFORMATION
Corresponding Author
* Louie@chem.utah.edu
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENT
We gratefully acknowledge the National Science Foundation
(CHE1213774) and the National Institute of Health
(GM076125) for financial support. We are grateful to Profes-
sor Matthew T. Kieber-Emmons for valuable discussions on
Conclusion
Dppf was discovered to be a uniquely suitable ligand
for the generation of stable Ni–ketene complexes, and we
ACS Paragon Plus Environment

Page 9 of 10
Journal of the American Chemical Society
computational chemistry. We also thank Joseph Lovelace for
(20) Wang, Y.; Agbossou, F.; Dalton, D. M.; Liu, Y.;
1
2
3
4
help synthesizing ketenes. Calculations were performed on
the Extreme Science and Engineering Discovery Environ-
ment (XSEDE), which is supported by National Science
Foundation (ACI-1053575).
Arif, A. M.; Gladysz, J. E. Organometallics 1993, 12, 2699.
(21) Barcs, B.; Kollár, L.; Kégl, T. Organometallics
2012, 31, 8082.
5
6
ABBREVIATIONS
7
8
Dppe, 1,2-bis(diphenylphosphino)ethane
Dppp, 1,3-bis(diphenylphosphino)propane
9
Dppb, 1,4-bis(diphenylphosphino)butane
Dppbenzene, 1,2-bis(diphenylphosphino)benzene
Dppf, 1,1’-bis(diphenylphosphino)ferrocene
Dtbpe, 1,2-bis(di-tert-butylphosphanyl)ethane
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
REFERENCES
(1) Zhang, Z.; Zhang, Y.; Wang, J. ACS Catal. 2011, 1,
1621.
(2) Ruchardt, C.; Schrauzer, G. N. Chem. Ber. 1960,
93, 1840.
(3) (a) Auvinet, A. L.; Harrity, J. P. Angew. Chem., Int.
Ed. 2011, 50, 2769. (b) Huffman, M. A.; Liebeskind, L. S.
J. Am. Chem. Soc. 1991, 113, 2771.
(4) Kumar, P.; Troast, D. M.; Cella, R.; Louie, J. J. Am.
Chem. Soc. 2011, 133, 7719.
(5) Rofer-DePoorter, C. K. Chem. Rev. 1981, 81, 447.
(6) Mindiola, D. J.; Hillhouse, G. L. J. Am. Chem. Soc.
2002, 124, 9976.
(7) Miyashita, A.; Sugai, R.; Yamamoto, J. J.
Organomet. Chem. 1992, 248, 239.
(8) Hoberg, H.; Korff, J. J. Organomet. Chem. 1978,
152, 255.
(9) Hofmann, P.; Perez-Moya, L. A.; Stegelmann, O.;
Riede, J. Organometallics 1992, 11, 1167.
(10) Curley, J. J.; Kitiachvili, K. D.; Waterman, R.;
Hillhouse, G. L. Organometallics 2009, 28, 2568.
(11) Grotjahn, D. B.; Collins, L. S. B.; Wolpert, M.;
Bikzhanova, G. A.; Lo, H. C.; Combs, D.; Hubbard, J. L.
J. Am. Chem. Soc. 2001, 123, 8260.
(12) Grotjahn, D. B.; Bikzhanova, G. A.; Collins, L. S.
B.; Concolino, T.; Lam, K.; Rheingold, A. L. J. Am.
Chem. Soc. 2000, 122, 5222.
(13) Urtel, H.; Bikzhanova, G. A.; Grotjahn, D. B.;
Hofmann, P. Organometallics 2001, 20, 3938.
(14) Ellgen, P. C. Inorg. Chem. 1971, 10, 232.
(15) Staudaher, N. D.; Stolley, R. M.; Louie, J. Chem.
Commun. 2014, 50, 15577.
(16) Le Page, M. D.; Patrick, B. O.; Rettig, S. J.; James,
B. R. Inorg. Chem. Acta. 2015, 431, 276.
(17) Yin, G.; Kalvet, I.; Englert, U.; Schoenebeck, F. J.
Am. Chem. Soc. 2015, 137, 4164.
(18) Sinnokrot, M. O.; Sherrill, C. D. J. Phys. Chem. A
2004, 108, 10200.
(19) (a) Bleuel, E.; Laubender, M.; Weberndorfer, B.;
Werner, H. Angew. Chem., Int. Ed. 1999, 38, 156. (b)
Cordaro, J. G.; van Halbeek, H.; Bergman, R. G. Angew.
Chem., Int. Ed. 2004, 43, 6366. (c) Grotjahn, D. B.;
Bikzhanova, G. A.; Hubbard, J. L. Organometallics 1999,
18, 8614.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment