(Diphosphine)Nickel-catalyzed negishi cross-coupling: An experimental and theoretical study

Article abstract of DOI:10.1002/chem.201500192

The use of a strongly donating "(bis-dialkylphosphine)Ni" fragment promotes the catalytic coupling of a large range of ArCl and ArZnCl derivatives under mild conditions. Stoichiometric mechanistic investigations and DFT calculations prove that a Ni⁰/Ni^II cycle is operative in this system.

Full text of DOI:10.1002/chem.201500192

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
DOI: 10.1002/chem.201500192
Communication
&
Cross-Coupling Reactions
“(Diphosphine)Nickel”-Catalyzed Negishi Cross-Coupling:
An Experimental and Theoretical Study
Emmanuel Nicolas,[a] Alexia Ohleier,[b] Florian D’Accriscio,[b] Anne-Frꢀdꢀrique Pꢀcharman,[b]
Matthieu Demange,[a] Philippe Ribagnac,[a] Jorge Ballester,[a] Corinne Gosmini,[a] and
Nicolas Mꢀzailles*[b]
more complex than the related Pd one. Kochi et al. showed in
Abstract: The use of a strongly donating “(bis-dialkylphos-
phine)Ni” fragment promotes the catalytic coupling of
a large range of ArCl and ArZnCl derivatives under mild
conditions. Stoichiometric mechanistic investigations and
DFT calculations prove that a Ni0/NiII cycle is operative in
this system.
the late 1970s in seminal contributions that in the case of reac-
tions of ArBr and Ni complexes featuring monophosphine li-
gands (such as PPh3), a single electron transfer (SET) process
could occur between the NiII complex generated by oxidative
addition of ArBr, resulting in the formation of a NiI com-
plex.[22,23] This SET is kinetically favorable over the transmetala-
tion/reductive elimination sequence, and hence the overall cat-
alytic process then involves NiI/NiIII species. In a 1986 study
dealing with the coupling of ArCl using Ni catalysts and reduc-
ing agents, Colon and Kelsey proposed a slightly different
mechanism in which the NiII complex is reduced by the reduc-
ing agent and not the ArCl.[24] Interestingly, ligand effects were
also studied and they concluded that triarylphosphines were
the best ligands, with trialkylphosphines resulting in much
slower reaction and poorer yield, and bidentate diarylphos-
phines being inefficient.
Over the past 40 years, Pd-catalyzed cross-coupling reactions
have been extensively studied.[1] This is especially true for
Csp2ÀCsp2 bond forming processes, because unsymmetrical
biaryl compounds are found in attractive targets ranging from
liquid crystals to pharmaceuticals.[2] Among them, the Negishi
reaction, involving organozinc derivatives, has been studied to
a lesser extent than the Suzuki–Miyaura reaction, involving or-
ganoboron species. The most active catalyst systems have
been devised from the early 2000s, and feature a single bulky
phosphine-like[3–5] or N-heterocyclic carbene (NHC)[6–8] ligands
bound to the Pd center.[9–12] It is, however, still desirable to de-
velop the chemistry of related Ni complexes because of the
much lower cost. Negishi et al. reported such arylÀaryl cou-
pling using a catalytic Ni0 precursor, [Ni(PPh3)4], generated in
situ from commercially available [Ni(acac)2], PPh3, and diisobu-
tylaluminum hydride (DIBALH) as early as 1977.[13] In terms of
catalytic efficiency, until recently, somewhat high catalytic load-
ings (ca. 5%) were necessary for efficient coupling processes.[14]
Ligand optimizations have resulted in significant lowering of
the Ni precatalyst amounts (typically ca. 1%, but down to
0.01% in the most favorable cases).[15–21] In terms of mecha-
nism, the possibility for Ni complexes to reach all oxidation
states ranging from Ni0 to NiIII makes the understanding of the
cross-coupling reaction involving a Ni catalyst undoubtedly
In the present contribution, we report an efficient Ni-cata-
lyzed Negishi cross-coupling process between R1C6H4Cl and
R2C6H4ZnCl derivatives under mild conditions, using the
chosen Ni0 precatalyst [(dcpp)Ni(h2-toluene)] (dcpp=1,3-bis(di-
cyclohexylphosphino)propane; Scheme 1).[25] We show here
Scheme 1. The nickel-catalyzed process studied.
the successful use of strongly donating bis-dialkylphosphine
ligand, well known to stabilize unsaturated Ni0 complexes
unlike the bis-diarylphosphines. Preliminary mechanistic inves-
tigations as well as a full computational analysis of the com-
plete catalytic cycle support a Ni0/NiII cycle.
The coupling with the cheapest and most accessible, but
still challenging, derivatives R1C6H4Cl was studied. Most satisfy-
ingly, the reaction between 4-chlorotoluene and phenylzinc
chloride yielded the expected biphenyl in an excellent yield of
93% within 6 h (entry 1, Table 1; 1% cat., 97% GC yield, to-
gether with 3% of the homocoupling product PhÀPh). The
quantity of Ni complex can be reduced to 0.2%, but the reac-
tion required 19 h (turnover frequency (TOF): 26 hÀ1) to be
complete (entry 2). When the quantity is lowered down to
0.01%, the yield was 21% after 58 h (TOF: 36 hÀ1) and 47%
after 138 h (TOF: 34 hÀ1), showing that no catalyst decomposi-
[a] Dr. E. Nicolas, Dr. M. Demange, Dr. P. Ribagnac, Dr. J. Ballester,
Dr. C. Gosmini
Laboratoire de Chimie Molꢀculaire
UMR 9168 CNRS, Ecole Polytechnique
91128 Palaiseau (France)
[b] A. Ohleier, F. D’Accriscio, Dr. A.-F. Pꢀcharman, Dr. N. Mꢀzailles
Laboratoire Hꢀtꢀrochimie Fondamentale et Appliquꢀe
UMR 5069 CNRS, Universitꢀ Paul Sabatier
118 Route de Narbonne, 31062 Toulouse (France)
E-mail: mezailles@chimie.ups-tlse.fr
Homepage: lhfa.cnrs.fr
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201500192.
Chem. Eur. J. 2015, 21, 1 – 6
1
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Communication
pling processes have been shown to involve a NiI/NiIII cycle
with many ligand systems.[14,27] Nonetheless, the use of [(BI-
NAP)Ni(h2-NC-Ph)] (BINAP=2,2’-bis(diphenylphosphino)-1,1’-bi-
naphthyl) in the catalytic coupling of ArCl with amines was re-
cently reported and shown to occur via a Ni0/NiII pathway.[28]
Several experiments, stoichiometric as well as catalytic, were
carried out to further prove the involvement of the Ni0/NiII
cycle using the strongly donating bidentate phosphine dcpp,
as well as to identify likely intermediates in the catalytic pro-
cess (Scheme 2).
Table 1. Scope of the coupling reaction.
Entry R1PhCl
R2PhZnCl Cat. [%] t [h] Yield [%]
R2:
(GC conversion [%])
1
4-CH3PhCl
H
1
6
19
58
93 (97)
93 (98)
(21)
2
3
4
5
6
7
8
9[a]
10
11
12
13
14
15
16
17[b]
18[c]
4-CH3PhCl
4-CH3PhCl
4-CH3PhCl
4-CF3PhCl
4-CO2MePhCl
PhCl
4-OMePhCl
2,6-(Me)2PhCl
2-ClPy
2,6-Cl2Py
2-Cl-3-NH2Py
4-CF3PhCl
4-CO2MePhCl 4-OMe
PhCl
H
H
H
H
H
H
H
H
0.2
0.01
0.01
0.2
0.2
0.2
0.2
1
1
1
1
0.2
0.2
0.2
0.2
1
138 (47)
1
6
16
32
7
2
6
4
0.5
1.5
4
8
2
94 (98)
92 (96)
98 (100)
90 (93)
75 (81)
96 (100)
58 (73)
90 (100)
91 (97)
89 (94)
90 (97)
88 (95)
(97)
Firstly, the stoichiometric addition of chlorobenzene to
[(dcpp)Ni(h2-toluene)] at 408C resulted in the quantitative for-
mation of the [(dcpp)Ni(Ph)Cl] complex, II, within minutes. This
complex is square planar as shown by its diamagnetism; two
sets of doublets are observed in the 31P{1H} NMR spectrum at
H
H
H
4-OMe
2
5.8 and 19.0 ppm, JP,P =48.4 Hz, see the Supporting Informa-
4-OMe
4-OMe
H
H
H
tion). The analogous complex, III, was obtained through the re-
action between complex I and 4-CH3PhCl. The oxidative addi-
tion of ArCl to the Ni0 fragment is therefore a low energy pro-
cess. Then, the possibility of a redox process between NiII com-
plex II and ArCl was studied. Despite heating for days at 808C,
no evolution of II was observed (Scheme 2, Eq. (1)), which
shows that the strong coordination of the bidentate phos-
phine prevents reduction to a NiI species, observed in the case
of monophosphines. On the other hand, addition of one equiv-
alent of PhZnCl to complex III resulted in the quantitative for-
4-CH3PhCl
4-CH3PhCl
4-CH3PhCl
1
1
6
72
(95)
(7)
19[d,e] 4-CH3PhCl
Reaction conditions: R1ArCl (1 mmol, 1 equiv), R2ArZnCl (1.1 mmol,
1.1 equiv), cat. [(dcpp)Ni(tol)], THF (2 mL), 608C. The difference between
the conversion measured by GC and the yield is accounted for by the for-
mation of the homocoupled product, PhÀPh. [a] 1.8 equiv PhZnCl used.
[b] [(dcpp)NiCl2]
as
catalyst.
[c] [(dcpp)Ni(Ph)Cl]
as
catalyst.
[d] [(dcpp)Ni(m2-Cl)2Ni(dcpp)] as catalyst. [e] PhÀPh is also formed in 33%
yield.
tion occurred. In a second set of experiments, the
electronic nature of the para substituent in R1C6H4Cl
was modulated. As classically observed, increasing
the electron density of the aryl moiety resulted in
a slower reaction (compare entries 2, 7, and 8). On
the other hand, increasing the electron density of the
zinc partner resulted in faster reaction (compare en-
tries 5–7 to 13–15). Nevertheless, because of the sta-
bility of the complex over time, near quantitative
yields of coupled products were obtained in every
case. Importantly, more challenging substrates, ortho-
substituted as well as heteroaryl or even amino-con-
Scheme 2. Mechanistic investigations.
taining derivatives, proved successful in the coupling
process (entries 9–12).
In summary, the representative examples provided
here show that the new catalytic process, based on a well-de-
fined Ni0 complex, allows the formation of biaryl compounds
under mild conditions with very low catalyst loading. It was
then verified that similar results were obtained by using more
readily manipulated NiII precursors, namely [(dcpp)NiCl2] or
[(dcpp)Ni(Ph)Cl] (entries 17 and 18). Most importantly, isolated
NiI precursor [(dcpp)Ni(m2-Cl)2Ni(dcpp)][26] not only performed
sluggishly (40% conversion after 72 h with 1% cat.), but also
less selectively (7% of the desired 4-CH3PhÀPh compound
versus 33% of the homocoupling product PhÀPh), proving
that the NiI/NiIII cycle is much less efficient in this reaction
(entry 19). These latter experiments are in contrast with some
literature precedence. Indeed, Ni-catalyzed Negishi cross-cou-
mation of the expected 4-CH3PhÀPh, without any formation of
the homocoupling product PhÀPh (Scheme 2, Eq. (2)). The for-
mation of the latter compound is indicative of the intermedi-
ate formation of a NiI species (see above). To further test the
involvement of a NiI/NiIII cycle, the known dimeric NiI complex
[(dcpp)Ni(m2-Cl)2Ni(dcpp)], V, was reacted with a stoichiometric
amount of PhZnCl at room temperature. A mixture of complex
[(dcpp)Ni(Ph)Cl], II, and a second species, VI, characterized by
a singlet in 31P{1H} NMR spectrum at 56 ppm, was formed
within minutes as shown by 31P{1H} NMR spectroscopy
(Scheme 2, Eq. (3)). This mixture rapidly evolved as complex II
reacted further with PhZnCl to form the homocoupled product
PhÀPh, preventing further characterization of compound VI.
&
&
Chem. Eur. J. 2015, 21, 1 – 6
www.chemeurj.org
2
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Communication
Nonetheless, this latter experiment is crucial for mechanistic
understanding as it shows that the present “(dcpp)Ni” system
readily evolves from NiI species to mixtures of diamagnetic
complexes (NiII and/or Ni0 complexes) in the presence of
PhZnCl. This reaction provides a rationale for the formation of
the homocoupled product of PhZnCl. Together, these experi-
ments are in full accord with the catalytic reaction presented
above (Table 1, entry 19), explaining the formation of mixtures
of homo- and heterocoupling products when NiI complex V
was used in catalytic amounts. In summary, these stoichiomet-
ric experiments prove that all three steps of the Ni0/NiII cycle
are facile at room temperature: oxidative addition of PhCl to
Ni0; transmetalation at Ni(Ar)(Cl) with Ar’ZnCl; and reductive
elimination from Ni(Ar)(Ar’). These experiments also show that
NiI complexes can be involved in cross-coupling, but they then
lead to a poor selectivity of heterocoupling versus homocou-
pling, favoring the homocoupling product. Given the low yield
of homocoupled product in the catalytic experiments, the in-
volvement of NiI in the present catalytic process can be safely
ruled out. In an additional effort to identify the resting state of
the catalyst, we monitored the reaction of 4-CF3PhCl with
PhZnCl catalyzed by 1% of [(dcpp)Ni(h2-toluene)] at 608C by
31P{1H} NMR spectroscopy. The only species that was observed
during the whole reaction (after 30%, 70%, and ca. full conver-
sion) was characterized by two doublets at 5.9 and 19.7 ppm,
2JP,P =56.0 Hz. This complex has almost identical spectroscopic
features to [(dcpp)Ni(4-CF3-Ph)Cl], that is, two doublets at 5.9
2
and 19.3 ppm, JP,P =53.0 Hz, in C6D6 (see the Supporting Infor-
mation. We, thus, propose that it corresponds to either com-
plex D or the heterobimetallic adduct [(dcpp)Ni(4-CF3-Ph)(m-
Cl)Zn(Cl)(Ph)(THF)] E (Scheme 5, see below). Importantly, in
contrast to Hartwig’s recent report, wherein the [(BINAP)Ni(h2-
NC-Ph)] partly evolves during the course of catalysis to [(BI-
NAP)2Ni], the bis-dialkylphosphine ligand dcpp prevents de-
composition of the catalytic species.[28]
As a final investigation of the mechanism, a full DFT study
has been performed on the catalytic system (for details see the
Supporting Information). First, the entry into the catalytic cycle
was computed. The results implied displacement of the tolu-
ene ligand from the 16-electron complex (h2 coordination) by
the ArCl substrate. Both dissociative and associative mecha-
nisms were envisaged (Scheme 3). Dissociation of toluene to
form B is endergonic, but the energy required is low
(18.9 kcalmolÀ1). The associative mechanism involves
an 18-electron transition state, TSAB2. It features both
incoming PhCl and toluene ligands as h1 coordinated
species (see the Supporting Information). Evolution
to complex C occurs via the PhCl h1-coordinated in-
termediate complex B2. Overall, the two mechanisms
are close in energy (less than 4 kcalmolÀ1 difference),
but more importantly the activation energy required
to enter the catalytic cycle is very low (18.9 kcal
molÀ1).
In a second stage, the catalytic process was com-
puted and the reaction profile is depicted in
Scheme 4. The previous step forms C, which is elec-
tron rich enough to perform an oxidative addition in
the carbon–chloride bond leading to NiII complex D,
which has been experimentally observed. The transi-
Scheme 3. Reaction pathways for the coordination of chlorobenzene.
Scheme 4. Reaction pathway for one catalytic cycle starting from complex C.
Chem. Eur. J. 2015, 21, 1 – 6
www.chemeurj.org
3
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Communication
tion state situated 12.9 kcalmolÀ1 above C is very low despite
elimination of ZnCl2(THF)2 leads to G, lying 12.6 kcalmolÀ1
the classical difficulty in performing oxidative addition with CÀ lower than F. The whole transmetalation step is rather exer-
Cl bonds, proving the electron-rich nature of the (dcpp)Ni frag-
ment, B. This reaction is quite exergonic, liberating 30.8 kcal
molÀ1 from C. Once the CÀCl bond has been activated, the
next steps consist of the approach and reaction with the orga-
nozinc reagent. The coordination of the arylzinc chloride to D
gonic, with an overall energy gain from D to G of 9.0 kcal
molÀ1. Finally, reductive elimination of biphenyl from G is facile
as the transition state TSG-H lies 13.2 kcalmolÀ1 higher than G.
Interestingly, the biphenyl remains coordinated through two
electrons of one aromatic ring, in a similar way than in com-
plex A, to form H (see the Supporting Information). Overall,
each catalytic cycle is strongly exergonic, by 52.5 kcalmolÀ1
,
and the transmetalation step possesses the highest barrier
within the cycle (21.8 kcalmolÀ1).
In a third stage, the product to reagent exchange that con-
nects the biphenyl complex H to C was computed. This pro-
cess allows a new catalytic cycle to occur. Here also, the disso-
ciative and associative mechanisms were computed and the
reaction profile is depicted in Scheme 5. Note that, for clarity,
the energy of H is set at 0. The dissociative process involves
the intermediacy of the 14-electron complex B, is calculated to
be 22.9 kcalmolÀ1 higher in energy than H and leads to C, lo-
cated 1.8 kcalmolÀ1 lower than H, making the exchange exer-
gonic.
In the associative process, a transition state, TSHB2, was locat-
ed at 29.7 kcalmolÀ1 above H. In TSHB2, the biphenyl and the
PhCl ligands are coordinated to the Ni center in an h1 fashion
(see the Supporting Information). This transient complex
evolves to the h1 complex B2 on the way to C. The dissociative
Figure 1. Computed structures of E (left), TSE-F (top), and F (right).
leads to heterobimetallic complex E (Figure 1), which
features a tetrahedral zinc derivative having for li-
gands one phenyl, one chloride, one THF molecule,
and the chlorine (bridging) atom from D behaving as
an L-type ligand. This coordination is again exergon-
ic, E being 9.0 kcalmolÀ1 lower in energy than D,
a fact that is directly linked to the strength of the
ZnÀCl bond. A transition state for the key transmeta-
lation step could be located on the potential energy
surface: TSE-F (Figure 1) is situated 21.8 kcalmolÀ1
higher in energy than E. In TSE-F, the NiÀCl and ZnÀCl
bonds are 0.10 ꢂ longer and 0.16 ꢂ shorter than in E
respectively, showing the Cl transfer. In parallel, the
PhÀZn bond is only marginally elongated (0.07 ꢂ)
and the PhÀNi bond is rather long (2.58 ꢂ). Most in-
Scheme 5. Reaction pathways for the product to reagent exchange: regeneration of h2
ArCl complex C.
terestingly, a very short NiÀZn intermetallic distance
of 2.59 ꢂ is calculated, much shorter than the sum of
the van der Waals radii (3.02 ꢂ), and in fact only mar-
ginally longer than the sum of the covalent radii (2.52 ꢂ), sug-
gesting metallophilic interaction in the transition state. Related
transition states featuring such short interactions have been
calculated for Pd–Zn[29] or Pd–Au/Sn–Au species,[30] whereas
they are absent in the Pd–Sn[30] transmetalation processes, and
were linked to lower activation energies. The transition state
TSE-F leads to F, which lies 14.6 kcalmolÀ1 higher than E. Com-
plex F is an interesting heterobimetallic complex with one
bridging phenyl ring between the Ni and the Zn centers
(Figure 1). The NiÀC (2.04 ꢂ) and ZnÀC (2.20 ꢂ) bond lengths
indicate that the phenyl is almost fully transferred to Ni. This
bridging Ph moiety is thus clearly a weaker ligand than THF, as
mechanism is calculated to be 6.8 kcalmolÀ1 lower than the as-
sociative one, making the former process favorable. Overall,
the regeneration of C and liberation of the biphenyl product is
a favorable and facile process, thereby allowing further catalyt-
ic cycles to proceed.
In conclusion, we present here an efficient nickel-based cata-
lytic system to perform a Negishi-type cross-coupling of aryl
chlorides with equimolar arylzinc chlorides. It is shown here for
the first time that strongly electron donating diphosphine li-
gands are in fact excellent ligands for this nickel-catalyzed pro-
cess. Mechanistic investigations prove the involvement of Ni0
and NiII intermediates, confirmed by the DFT analysis of the
&
&
Chem. Eur. J. 2015, 21, 1 – 6
www.chemeurj.org
4
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!