Nickel-on-charcoal-catalyzed aromatic aminations and Kumada couplings: Mechanistic and synthetic aspects

Article abstract of DOI:10.1021/jo020297e

Protocols for aromatic aminations and Kumada couplings catalyzed by 'heterogeneous' nickel-on-charcoal (Ni/C) have been revised, making them simpler and more time efficient. For both types of reactions, reduction of the catalyst precursor Ni(II)/C using n-BuLi prior to addition of a substrate can be avoided. Instead, in amination reactions, the amine in combination with lithium tert-butoxide was found to convert Ni(II)/C to active Ni(0). For Kumada couplings, direct reduction of Ni(II)/C by the Grignard reagent is easily achieved. Reactions run in the presence of triarylphosphine ligands of varying substitution patterns and with differing electronic properties provided insight into the mechanism of these nickel-catalyzed transformations. Ligandless Kumada couplings were facile with aryl Grignards, which may be a consequence of π-complexation of nickel by the aryl group in the reagent. Larger scale reactions of both types of couplings have been successfully performed, suggesting that Ni/C-based processes can be scaled-up as needed.

Full text of DOI:10.1021/jo020297e

Nick el-on -Ch a r coa l-Ca ta lyzed Ar om a tic Am in a tion s a n d Ku m a d a
Cou p lin gs: Mech a n istic a n d Syn th etic Asp ects
Stefan Tasler and Bruce H. Lipshutz*
Department of Chemistry & Biochemistry, University of California,
Santa Barbara, California 93106-9510
lipshutz@chem.ucsb.edu
Received April 29, 2002
Protocols for aromatic aminations and Kumada couplings catalyzed by ‘heterogeneous’ nickel-on-
charcoal (Ni/C) have been revised, making them simpler and more time efficient. For both types of
reactions, reduction of the catalyst precursor Ni(II)/C using n-BuLi prior to addition of a substrate
can be avoided. Instead, in amination reactions, the amine in combination with lithium tert-butoxide
was found to convert Ni(II)/C to active Ni(0). For Kumada couplings, direct reduction of Ni(II)/C
by the Grignard reagent is easily achieved. Reactions run in the presence of triarylphosphine ligands
of varying substitution patterns and with differing electronic properties provided insight into the
mechanism of these nickel-catalyzed transformations. Ligandless Kumada couplings were facile
with aryl Grignards, which may be a consequence of π-complexation of nickel by the aryl group in
the reagent. Larger scale reactions of both types of couplings have been successfully performed,
suggesting that Ni/C-based processes can be scaled-up as needed.
In tr od u ction
special ligands had to be developed to allow for the
corresponding conversions via Pd(0) catalysis.8,15,16 An
aromatic amination catalyzed by nickel was first reported
in 1950,17 and the method was further extended by
Cramer et al. in 1975.18 The real breakthrough for
transition metal-catalyzed amination reactions, however,
was achieved during the past decade by extensive work
from the Buchwald and Hartwig groups.4-6,11,12
Aromatic aminations and Kumada-type couplings rep-
resent powerful tools for the formation of C-N and C-C
bonds, respectively. While Kumada developed a coupling
strategy for aryl halides and Grignard reagents based
on nickel catalysis,1-3 aromatic aminations are mainly
performed using palladium catalysts.4-6 Generally,
nickel(0) complexes display greater reactivity toward
oxidative addition with aryl halides relative to
palladium(0).7,8 Thus, transformations based on catalytic
Ni(0) were readily performed on aryl chlorides,1,9-14 while
The high reactivity of nickel toward inexpensive and
readily available aryl chlorides, combined with the
(10) (a) Shirakawa, E.; Yamasaki, K.; Hiyama, T. J . Chem. Soc.,
Perkin Trans. 1 1997, 2449. (b) Saito, S.; Oh-tani, S.; Miyaura, N. J .
Org. Chem. 1997, 62, 8024. (c) Indolese, A. F. Tetrahedron Lett. 1997,
38, 3513. (d) Iyoda, M.; Otsuka, H.; Sato, K.; Nisato, N.; Oda, M. Bull.
Chem. Soc. J pn. 1990, 63, 80, and references therein.
(11) For recent nickel-catalyzed aromatic aminations, see: Wolfe,
J . P.; Buchwald, S. L. J . Am. Chem. Soc. 1997, 119, 6054.12
(12) Desmarets, C.; Schneider, R.; Fort, Y. J . Org. Chem. 2002, 67,
3029; and references therein.
* Phone: +805-893-2521; Fax: +805-893-8265.
(1) (a) Tamao, K.; Sumitani, K.; Kiso, Y.; Zembayashi, M.; Fujioka,
A.; Kodama, S.-i.; Nakajima, I.; Minato, A.; Kumada, M. Bull. Chem.
Soc. J pn. 1976, 49, 1958. (b) Kumada, M. Pure Appl. Chem. 1980, 52,
669.
(2) Corriu, R. J . P.; Masse, J . P. J . Chem. Soc., Chem. Commun.
1972, 144.
(13) Lipshutz, B. H. Adv. Synth. Catal. 2001, 343, 313.
(3) For recent nickel-catalyzed Kumada couplings using polymer-
immobilized catalysts, see: Styring, P.; Grindon, C.; Fisher, C. M.
Catal. Lett. 2001, 77, 219.
(4) (a) Wolfe, J . P.; Wagaw, S.; Marcoux, J .-F.; Buchwald, S. L. Acc.
Chem. Res. 1998, 31, 805. (b) Hartwig, J . F. Acc. Chem. Res. 1998, 31,
852. (c) Yang, B. H.; Buchwald, S. L. J . Organomet. Chem. 1999, 576,
125.
(14) Lipshutz, B. H.; Tasler, S.; Chrisman, W.; Spliethoff, B.; Tesche,
B. J . Org. Chem. 2002, 67, 1177, and references therein.
(15) (a) Stu¨rmer, R. Angew. Chem. 1999, 111, 3509; Angew. Chem.,
Int. Ed. 1999, 38, 3307, and references therein. (b) Alcazar-Roman, L.
M.; Hartwig, J . F. J . Am. Chem. Soc. 2001, 123, 12905, and references
therein. (c) Farina, V.; Krishnan, B. J . Am. Chem. Soc. 1991, 113, 9585.
(d) Portnoy, M.; Milstein, D. Organometallics 1993, 12, 1665. (e)
Herrmann, W. A.; Brossmer, C.; O¨ fele, K.; Reisinger, C.-P.; Priermeier,
T.; Beller, M.; Fischer, H. Angew. Chem. 1995, 107, 1989; Angew.
Chem., Int. Ed. Engl. 1995, 34, 1844. (f) Beller, M.; Fischer, H.;
Herrmann, W. A.; O¨ fele, K.; Brossmer, C. Angew. Chem. 1995, 107,
1992; Angew. Chem., Int. Ed. Engl. 1995, 34, 1848. (g) Wolfe, J . P.;
Singer, R. A.; Yang, B. H.; Buchwald, S. L. J . Am. Chem. Soc. 1999,
121, 9550. (h) Mathews, C. J .; Smith, P. J .; Welton, T. J . Chem. Soc.,
Chem. Commun. 2000, 1249. (i) Andreu, M. G.; Zapf, A.; Beller, M. J .
Chem. Soc., Chem. Commun. 2000, 2475. (j) Dai, C.; Fu, G. C. J . Am.
Chem. Soc. 2001, 123, 2719. (k) Carpentier, J . F.; Petit, F.; Mortreux,
A.; Dufaud, V.; Basset, J .-M.; Thivolle-Cazat, J . J . Mol. Catal. 1993,
81, 1.
(5) Hartwig, J . F. Angew. Chem. 1998, 110, 2154-2177; Angew.
Chem., Int. Ed. 1998, 37, 2046.
(6) Prim, D.; Campagne, J .-M.; J oseph, D.; Andrioletti, B. Tetrahe-
dron 2002, 58, 2041.
(7) The oxidation potential for the system Pd(II)/Pd(0) is much
higher than that of Ni(II)/Ni(0), therefore Ni(0) is more prone to
oxidative addition; Dean, J . A. Lange’s Handbook of Chemistry, 14th
ed.; McGraw-Hill: New York, 1992.
(8) Grushin, V. V.; Alper, H. Chem. Rev. 1994, 94, 1047.
(9) The bond strength for nickel halides increases from iodine to
chlorine, whereas those for palladium halides decreases in that order.
Therefore, the energy gain from the formed metal-halide bond during
oxidative addition is highest for Ni-Cl but lowest for Pd-Cl. See ref
7 and Henry, P. M. Palladium Catalyzed Oxidation of Hydrocarbons;
D. Reidel Publishing Company: Boston, 1980; pp 11-12.
(16) Huang, J .; Nolan, S. P. J . Am. Chem. Soc. 1999, 121, 9889.
(17) Hughes, E. C.; Veatch, F.; Elersich, V. Ind. Eng. Chem. 1950,
42, 787.
10.1021/jo020297e CCC: $25.00 © 2003 American Chemical Society
Published on Web 08/16/2002
1190
J . Org. Chem. 2003, 68, 1190-1199

Aromatic Aminations and Kumada Couplings
TABLE 1. Ar om a tic Am in a tion s of 1 w ith 2 Ca ta lyzed
by Ni/C
advantages of heterogeneous catalysis, has led to devel-
opment of a catalytic system based on nickel-on-charcoal
(‘Ni/C’). Applications of this reagent to several different
C-C bond-forming reactions (including Kumada-type
couplings), hydrodehalogenations, and aromatic amina-
tions have already been reported.13 Although an extensive
study has strongly implicated solution-based catalysis,
Ni/C can be considered as a heterogeneous catalyst in
that essentially complete recovery of the nickel in solution
is ultimately achieved on the support after filtration.14
In the course of our investigations on the nature of the
catalytically active species in Ni/C-mediated reactions,
several interesting mechanistic and synthetic aspects of
aromatic aminations19 and Kumada couplings20 have
been uncovered, which are the subject of this contribu-
tion.
entry phosphine ligand BuLi time [h] conv [%]a 3 [%]a
1
2
3
4
5
6
7
8
dppfb
dppfb
dppfb
dppfb
yes
no
no
1
1
1.5
2
1
1.5
2
2
3
6
99
78
93
100
72
96
100
67
83
100
12
11
83c
65
77
86
47
75
78
50
63
76
no
d
PPh3
PPh3
PPh3
PPh3
PPh3
PPh3
yes
yes
yes
no
no
no
yes
no
no
d
d
d
d
d
9
10
11
12
13
dpppe
1
2
2
0.4
-
2
dpppe
f
-
3
a
b
By quantitative GC. 0.5 equiv relative to nickel. c Isolated
yield: 85%. 4.0 equiv relative to nickel. e 1.0 equiv relative to
d
nickel. f No nickel source added.
Resu lts a n d Discu ssion
SCHEME 1. Am in a tion of Ch lor oben zon itr ile 1
w ith Mor p h olin e (2)
Ar om a tic Am in a tion Rea ction s. Aminations of aryl
chlorides using Ni/C, as previously described,19 required
a 90 min premixing period of Ni(II)/C and dppf in toluene
prior to the addition of n-BuLi for reduction of this
catalyst precursor to active Ni(0). From more recent
studies on this catalyst, which revealed an equilibrium
involving nickel in solution (Nisol) located both inside and
outside the charcoal matrix which occurs during the first
minutes of a reaction,14 such a long premixing is no longer
justified. Furthermore, potential replacement of dppf by
PPh3 was targeted for economic reasons, in addition to
exploring the possible use of reductants other than BuLi
for Ni(II)/C.
Reduction of Ni(II)/C by heat (425 °C) in a hydrogen
atmosphere has already been shown to cause migration
of nickel atoms and sintering, leading to bigger nickel
particles with decreased activity.14 Since phosphines and
phosphites reduce Pd(II),21 while phosphites22 and amines18
mediate reduction of Ni(II), amination of p-chloroben-
zonitrile (1) with morpholine (2) was carried out in the
presence of dppf without prior reduction of Ni(II)/C by
BuLi. The mixture was simply heated at reflux after the
substrate and all reagents had been added without any
period of incubation. As seen from Table 1, the reaction
proceeds somewhat more slowly but goes to completion
with similar yields of product 3 (entry 1 vs entries 2-4).
When 4 equiv of PPh3 (instead of 0.5 equiv of dppf) were
used relative to nickel, the reaction again was slower.
However, the amination eventually went to completion
either with (entry 1 vs entries 6, 7) or without (entries
3, 4 vs entries 8-10) prior reduction of Ni(II) by BuLi
(albeit with differing amounts of the two ligands; vide
infra), although yields varied slightly (compare entries
7, 10 vs entries 1, 4). Also examined was dppp, as it is
electronically relatively similar to PPh3 but offers biden-
tate complexation as with dppf. It has also been described
as the ligand of choice in nickel-catalyzed Kumada
couplings.1,23,24 The outcome of these experiments (entries
11, 12) was that no product was formed above the amount
found in an uncatalyzed reaction (entry 13).25 Clearly,
bidentate complexation is not sufficient for high conver-
sion, and complexes of nickel ligated by dppf or dppp can
be assumed to be quite different. Comparing results from
aminations with Ni(0)/C (from BuLi-reduced Ni(II)/C)
with those obtained with Ni(II)/C directly raised the
question as to whether the decreased reaction rate26 with
the latter catalyst was due to involvement of a different
nickel species [e.g. Ni(0) vs Ni(I)]. Alternatively, rate
differences could be a consequence of an induction period
during which Ni(II)/C was being reduced and/or the
equilibrium involving Nisol was being established. Figure
1 shows the plot of conversion versus time for aminations
with and without prior BuLi reduction of Ni(II)/C, using
either 0.5 equiv of dppf or 4 equiv of PPh3 as ligand. With
both phosphines, the conversion for reactions after BuLi
treatment progressed linearly, indicating that the reac-
tion is basically (pseudo) zero order in aryl chloride.27 The
BuLi-free variation, however, showed a delay of ca. 15
to 50 min with dppf and PPh3, respectively, during which
essentially no aryl chloride was consumed. After this
(23) Park, M.; Buck, J . R.; Rizzo, C. J . Tetrahedron 1998, 54, 12707.
(24) Sofia, A.; Karlstro¨m, E.; Itami, K.; Ba¨ckvall, J . E. J . Org. Chem.
1999, 64, 1745.
(18) Cramer, R.; Coulson, D. R. J . Org. Chem. 1975, 40, 2267.
(19) Lipshutz, B. H.; Ueda, H. Angew. Chem. 2000, 112, 4666;
Angew. Chem., Int. Ed. 2000, 39, 4492.
(20) Lipshutz, B. H.; Tomioka, T.; Blomgren, P. A.; Sclafani, J . A.
Inorg. Chim. Acta 1999, 296, 164.
(21) (a) Malatesta, L.; Angoletta, M. J . Chem. Soc. 1957, 1186. (b)
Mason, M. R.; Verkade, J . G. Organometallics 1992, 11, 2212. (c)
Amatore, C.; J utand, A.; M’Barki, M. A. Organometallics 1992, 11,
3009. (d) Ozawa, F.; Kubo, A.; Hayashi, T. Chem. Lett. 1992, 2177. (e)
Amatore, C.; J utand, A. Acc. Chem. Res. 2000, 33, 314.
(22) Vinal, R. S.; Reynolds, L. T. Inorg. Chem. 1964, 3, 1062.
(25) Similar findings were mentioned in refs 11 and 19, although
PPh3 was also unsuitable for these aromatic aminations.
(26) Comparison of rates is based on the assumption that similar
amounts of nickel are involved in the catalysis, not regarding possible
differences in lifetime of the catalytic species due to differing reaction
conditions. Previous work has shown that most likely around 80% of
Ni/C is actually available in solution for homogeneous catalysis within
the charcoal matrix, not depending on ligand concentration.14
(27) Colon, I.; Kelsey, D. R. J . Org. Chem. 1986, 51, 2627.
J . Org. Chem, Vol. 68, No. 4, 2003 1191

Tasler and Lipshutz
SCHEME 2. P ossible Mech a n istic P a th w a ys for
th e F or m a tion of Nick el(II) Am id e
ligand, and either morpholine or LiO-t-Bu did not result
in such a color change, even upon heating in refluxing
toluene (i.e., our standard reaction conditions for ami-
nations), again excluding the phosphine as the reducing
agent.29 In both cases, the yellow color prevailed after 1
h. However, as soon as both morpholine and LiO-t-Bu
were present, these yellow mixtures developed the dark
purple color, usually observed for aromatic aminations
catalyzed by NiCl2 without prior reduction with BuLi,
within 10 minutes. Therefore, the combination of base
and amine affords a reducing agent for Ni(II). Aromatic
amines can likewise be used for this purpose, as the same
tests performed on aniline resulted in similar observa-
tions.29
F IGURE 1. Extent of conversion of aryl chloride 1 in
aminations with morpholine (2) in the presence of dppf (top)
or PPh3 (bottom) using preformed Ni(0)/C (b) or Ni(II)/C ([).
The question then arises as to which of three pathways
(A-C, Scheme 2) is involved en route to the amide
complex iii. These include (1) complexation of Ni(II) by
the amine to form i, which then undergoes deprotonation
(path A); (2) prior complexation with tert-butoxide to
generate ii and subsequent deprotonation/complexation
of the amine (path B); (3) an already deprotonated amide
is directly complexed, with the latter species generated
via an acid/base equilibrium between amine and tert-
butoxide as the rate-limiting factor (path C). Insight was
gained from a reaction in which lithium morpholide (from
BuLi plus morpholine), was directly added to a NiCl2/
PPh3 mixture in toluene. Here, the darkening process was
slightly slower, with a brown-black color reached after
30 min. Consequently, pregeneration of uncomplexed
amide (path C) appears not to play a major rate-limiting
role (for differentiation between pathways A and B: vide
infra). Whether Ni(0) is eventually generated from Ni-
morpholide (iii) by single electron transfer (SET)28 or by
a direct two electron process was not further investigated.
Interestingly, the ligandless reaction with NiCl2 (not
reduced with BuLi) led to 56% consumption of aryl
chloride 1 and 35% product formation (Table 2, entry 4),
whereas a ligandless reaction using Ni/C with or without
prior BuLi reduction of Ni(II) (entries 1, 2), or with NiCl2
reduced with BuLi (entry 3), resulted only in yields
comparable to the uncatalyzed reaction (Table 1, entry
13), even though there was partial consumption of aryl
chloride (but not into product 3). Apparently, BuLi
time, the amination of 1 proceeded at a nearly identical
rate26 as for that with Ni(0)/C arrived at via BuLi
reduction of Ni(II)/C. Hence, similar catalytically active
species seem to be involved as soon as a reduced nickel
species is formed. Attempts to monitor and kinetically
analyze the amination reaction using React-IR failed due
to significant overlap of IR signals of educt and product.14
Given that Nisol was the only catalytically active species,
linear product formation right from the start in reactions
of Ni(0)/C indicates that the postulated14 equilibrium for
Nisol had to be established during the first few minutes,
otherwise a delay in conversion should have occurred.
As potential reducing agents in the absence of BuLi,
phosphine ligands21,22 and the tertiary amine 328 were
excluded experimentally.29 Another possibility was that
morpholine serves as a reducing agent for Ni(II)/C. Its
reduction potential was checked simply by monitoring the
color of a reaction mixture with anhydrous NiCl2 as a
Ni(II) source. Reduction of a yellow suspension of NiCl2
in toluene with BuLi at room temperature resulted in a
quick color change to black, indicating significant genera-
tion of Ni(0). On the other hand, mixtures of NiCl2,
(28) For a similar reduction of Pd(II) by tertiary amines, see: (a)
McCrindle, R.; Ferguson, G.; Arsenault, G. J .; McAlees, A. J .; Stephen-
son, D. K. J . Chem. Res. (S) 1984, 360. (b) Beckwith, A. L. J .; Eichinger,
P. H.; Mooney, B. A.; Prager, R. H. Aust. J . Chem. 1983, 36, 719.
(29) For experimental details, see Supporting Information.
1192 J . Org. Chem., Vol. 68, No. 4, 2003

Aromatic Aminations and Kumada Couplings
TABLE 2. Liga n d less Am in a tion Rea ction s of 1 w ith 2
TABLE 3. Ar om a tic Am in a tion s of 1 w ith 2 in th e
P r esen ce of Differ en t P h osp h in es
entry
source of Ni
BuLi
conv [%]a
3 [%]a
entry
liganda
BuLi
time [h]
conv. [%]b
10 [%]b
1
2
3
4
Ni/C
Ni/C
NiCl2
NiCl2
yes
no
yes
no
27
22
45
56
2
3
3
1
2
3
4
5
6
7
8
dppf
yes
yes
yes
yes
yes
no
no
no
no
no
1
1
1
1
1
2
2
2
2
2
99
84
70
72
60
100
85
69
67
67
83
66
55
47
42
86
67
58
48
50
P(p-anis)3
P(p-tol)3
PPh3
P(p-F-Ph)3
dppf
P(p-anis)3
P(p-tol)3
PPh3
35
a
By quantitative GC, reaction time 2 h.
reduces Ni(II) so rapidly that the Ni(0) formed cannot
be stabilized by weak ‘ligands’; e.g., a substrate, perhaps
leading to nickel colloids (black in color)12,30 that seem to
be catalytically inactive under these amination condi-
tions. Not so easily explained, however, is the difference
between NiCl2- and Ni/C-catalyzed couplings without
prior BuLi reduction. Reduction of NiCl2 by amine/LiO-
t-Bu appears to be sufficiently mild and somewhat slower
than that by BuLi such that at least a minor stabilizing
effect can be exerted by a substrate, probably by oxidative
addition,18 resulting in partial product formation. That
Ni/C fails to catalyze formation of product 3 suggests that
NiCl2 may not mimic the Ni/C system. This may be due
to differing nickel species involved in the catalysis
(dependent on the nickel source, not the reduction method
used). Alternatively, the lack of product formation could
be attributed to additional steric and/or electronic effects
exerted by the charcoal matrix on Ni(0), perhaps in the
form of π-interactions of graphite-like substructures,31
thereby rendering this phosphine-free system unreactive
in these aminations.
While in early investigations aromatic aminations were
catalyzed by ligandless NiCl2 with reduction of
Ni(II) assumed to be effected by the amine,17,18 Buchwald
et al. noted that such a process did not occur in their
system.11 Reduction of Ni(II), therefore, seems to be a
function of reaction temperature. The original literature
work was performed in sealed tubes under pressure and
with temperatures above 150 °C being described as
necessary,17,18 the amine serving simultaneously as nu-
cleophile, base, and reductant. Apparently, addition of
lithium or sodium tert-butoxide increases the reducing
ability of the mixture, although temperatures above 100
°C are still essential (i.e. no reduction occurs at 100 °C,11
but in overheated toluene reduction does take place; vide
supra). Under conditions described herein, even aniline
is capable of mediating the reduction, an observation
questioned long ago by Cramer et al. for N-methyla-
niline.18
9
10
P(p-F-Ph)3
a
0.5 equiv of dppf; otherwise 4 equiv of PR3 were used (relative
b
to nickel). By quantitative GC.
ligand type, with tri(p-anisyl)-, tri(p-tolyl)-, and tri(p-
fluorophenyl)phosphines chosen for this purpose. For
both variants of these aminations [i.e., with and without
BuLi reduction of Ni(II)/C], the same reactivity patterns
were noted, with reaction rates decreasing in the order
dppf > P(p-anis)3 > P(p-tol)3 > PPh3 g P(p-fluorophenyl)3
(Table 3). For the monodentate phosphines, the major
trend is such that greater electron density at phosphorus
leads to faster reactions. Oxidative addition is generally
rapid with Ni(0) (cf. Kumada couplings; vide infra).32 On
the other hand, reductive elimination is usually consid-
ered rate-determining in palladium-catalyzed amina-
tions5 and is accelerated by ligands with less electron
density, a trend not observed in these Ni/C-mediated
reactions. Reductive elimination should be further en-
hanced in the presence of highly nucleophilic morpho-
line,5,6 and the rate-determining step should thus be
associated with the deprotonation/complexation of the
amine.33 In principle, two routes can be envisioned
leading to nickelamido complex iii (pathways A and B,
Scheme 2); either by complexation of tert-butoxide and
subsequent deprotonation/complexation of the amine (B),
or by complexation of the amine and deprotonation of this
activated species by free tert-butoxide (A).5 Whereas the
rate for pathway B should be enhanced by electron-rich
ligands which increases the basicity of tert-butoxide in
species ii, the rate for pathway A should decrease in this
case since the N-H bond in ammonium ion i should be
less activated toward deprotonation with weaker Lewis
acidic nickel. Using electron-rich ligands, the Lewis
acidity of the original nickel(II) fragment might be so low
that complexation of tert-butoxide (pathway B) is highly
favored over complexation of free amine (pathway A),
which is indeed reflected in the order of reactivity for PR3
(Table 3).
The experiments discussed above lead to the conclusion
that aromatic aminations can be performed by combining
all ingredients in the reaction flask at the same time;
that premixing of certain reagents and prior reduction
of Ni(II)/C with BuLi are not required, resulting in a
faster and simplified procedure. That aromatic amina-
tions of 1 with morpholine (2) can proceed with PPh3
allowed for evaluation of electronic effects within this
Whether the newly established reaction conditions for
coupling p-chlorobenzonitrile (1) and morpholine (2) are
applicable to other substrate combinations was next
examined. Aromatic aminations of p-chlorobenzonitrile
(32) Kumada couplings displayed higher levels of conversion even
with electron-rich aryl chlorides (i.e., deactivated for oxidative addition)
in the same reaction times as aromatic aminations (albeit in different
solvents, THF vs toluene). Hence, oxidative addition is not likely to
be rate-limiting.
(33) That different mechanistic pathways are operating for reductive
elimination with different ligands cannot be ruled out; e.g., triggering
of reductive elimination by SET: (a) Tsou, T. T.; Kochi, J . K. J . Am.
Chem. Soc. 1978, 100, 1634. (b) Morrell, D. G.; Kochi, J . K. J . Am.
Chem. Soc. 1975, 97, 7262. (c) Koo, K.; Hillhouse, G. L. Organometallics
1995, 14, 4421. See also ref 40.
(30) Formation of palladium colloids was postulated to generate
reactive species in ligandless Heck reactions and homocoupling of
chlorobenzene; see: (a) Reetz, M. T.; Westermann, E. Angew. Chem.
2000, 112, 170; Angew. Chem., Int. Ed. 2000, 39, 165. (b) Mukho-
padhyay, S.; Rothenberg, G.; Gitis, D.; Sasson, Y. J . Org. Chem. 2000,
65, 3107.
(31) Vol’pin, M. E.; Novikov, Y. N.; Lapkina, N. D.; Kasatochkin, V.
I.; Struchkov, Y. T.; Kazakov, M. E.; Stukan, R. A.; Povitskij, V. A.;
Karimov, Y. S.; Zvarikina, A. V. J . Am. Chem. Soc. 1975, 97, 3366.
J . Org. Chem, Vol. 68, No. 4, 2003 1193

Tasler and Lipshutz
TABLE 7. Sca le-u p of Ar om a tic Am in a tion s of 1 w ith 2
SCHEME 3. Ar om a tic Am in a tion s w ith a Va r iety
of Su bstr a tes
equiv of
entry
dppfa
BuLi
time [h]
conv [%]b
3 [%]b
1
2
3
0.5
0.5
1.5
yes
no
no
1
6
3
100
80
99
85
65
88
a
Relative to nickel; 0.05 equiv of nickel used (relative to aryl
b
chloride). By quantitative GC.
identical yields in comparable reaction times to those
using BuLi as reducing agent (Tables 4, 6, entries 3, 4;
Table 5, entries 3, 5). For the reaction of m-chlorotoluene
(4) with morpholine (2), longer reaction times and ad-
ditional dppf were necessary for complete conversion
(Table 5, entries 4, 5). Generally, all three transforma-
tions were expectedly slower than that between 1 and 2
and required increased amounts of dppf (from 0.5 to at
least 1.0 equiv) relative to nickel. This increase is most
likely needed to further stabilize the active nickel species
which might otherwise decompose during the extended
reaction times, or due to decomposition of dppf. Replace-
ment of dppf by PPh3 (4 equiv) proved to be detrimental
in all three cases (Tables 4, 5, entry 6; Table 6, entry 5);
thus, this ligand cannot be considered of general utility
in these reactions. The moderate yield (53%) for amina-
tion of chlorotoluene 4 with hexylamine 5 (Table 6,
entries 3, 4) is in good agreement with results published
by Buchwald, who stated that hexylamine leads to decent
yields only with activated aryl chlorides due to competing
â-elimination and hydrodehalogenation in slower reac-
tions with nonactivated substrates.11 This was verified
by the amination reaction of activated aryl chloride 1
with this amine, resulting in a higher (74%) yield (Table
4, entries 3, 4). Using the electron-rich substrate chlo-
rotoluene 4, small amounts of homocoupling to give 3,3′-
bistoluene (7) were detectable (Tables 5, 6).
TABLE 4. Ar om a tic Am in a tion s of 1 w ith 5 Ca ta lyzed
by Ni/C
equiv of
Nia
equiv of
entry
ligand liganda BuLi time [h] conv. [%]b 6 [%]b
1
2
3
4
5
6
0.05
0.05
0.05
0.05
0.075
0.075
dppf
dppf
dppf
dppf
dppf
PPh3
0.025 yes
0.025 yes
1
8
7
7
5
5
22
30
100
100
100
46
6
8
73
74c
73
13
0.1
0.1
yes
no
0.15
0.3
yes
yes
a
b
Relative to aryl chloride. By quantitative GC. c Isolated
yield: 76%.
TABLE 5. Ar om a tic Am in a tion s of 4 w ith 2 Ca ta lyzed
by Ni/C
equiv of
entry ligand liganda BuLi time [h] conv [%]b 6 [%]b 7 [%]b
1
2
3
4
5
6
dppf
dppf
dppf
dppf
dppf
PPh3
0.5
0.5
1
1
1.5
4
yes
yes
yes
no
no
yes
5
10
10
10
17
10
66
82
100
77
100
36
54
66
88c
65
85
5
1.8
3.1
1.1
1.0
2.5
-
a
Relative to nickel; 0.05 equiv of nickel used (relative to aryl
chloride). By quantitative GC. c Isolated yield: 84%.
b
The feasibility of scaling up an amination was also
studied, with and without prior reduction of Ni(II)/C with
BuLi. Conducting the coupling of 1 on a 15.0 mmol scale,
as opposed to using 0.75 mmol as the standard amount
of aryl chloride, amination with morpholine (2) using Ni-
(II)/C reduced with BuLi was reproduced in exactly the
same time and with identical results as with the smaller
scale reaction (Table 7, entry 1). For the BuLi-free
variant, however, the amount of dppf had to be increased
from 0.5 to 1.5 equiv (relative to nickel; entries 2, 3) to
obtain identical results in slightly increased reaction
times compared to our standard reactions (cf. Table 7,
entry 3 vs. Table 1, entry 4).
TABLE 6. Ar om a tic Am in a tion s of 4 w ith 5 Ca ta lyzed
by Ni/C
equiv of
entry ligand liganda BuLi time [h] conv [%]b 6 [%]b 7 [%]b
1
2
3
4
5
dppf
dppf
dppf
dppf
PPh3
0.5
0.5
1
1
4
yes
yes
yes
no
5
16
19
19
19
33
84
100
100
81
20
35
53
53c
-
0.7
1.1
1.0
1.3
yes
-
a
Relative to nickel; 0.05 equiv of nickel used (relative to aryl
chloride). By quantitative GC. c Isolated yield: 53%.
b
Ku m a d a Cou p lin gs. On the basis of our study of
aromatic aminations, premixing of ligands and Ni(II)/C
for a given period of time is not essential for these
reactions to take place. A similar simplification was
anticipated for Kumada couplings20 as well. Furthermore,
reduction of Ni(II)/C was likely to be easily achieved by
the Grignard reagent being used, in place of BuLi. For
inexpensive Grignard reagents this alternative would be
another improvement in the original procedure. A time
savings was also expected by adding the Grignard at
room temperature rather than at -78 °C, thus avoiding
initial cooling followed by warming.
(1) and m-chlorotoluene (4) were therefore performed
using either n-hexylamine (5) or morpholine (2) (Scheme
3) under both our original and newly developed condi-
tions, and with either dppf or PPh3 as ligand. Aminations
with anilines were not tested in this study, but a
successful outcome without prior BuLi reduction of Ni-
(II)/C is likely based on our results from NiCl2 reductions
performed with this aromatic amine (vide supra). Ami-
nations with anilines in the presence of PPh3 are not
likely to proceed, however, according to our preliminary
results.19
As can be seen from Tables 4-6, aminations without
prior Ni(II)/C reduction by BuLi resulted in essentially
Initial studies were performed on m-chlorotoluene (4),
which was treated in separate experiments with PhMgCl,
1194 J . Org. Chem., Vol. 68, No. 4, 2003

Aromatic Aminations and Kumada Couplings
SCHEME 4. Ku m a d a Cou p lin gs of Ch lor otolu en e 4
w ith Gr ign a r d Rea gen ts
attributed to either product 8 or homocoupled material
7 (compare with Tables 8, 9). If the transmetalation step
were rate-determining, reactions with PhMgCl should
have been slower than those with the electron-richer (p-
anisyl)MgCl, assuming limited interactions of the latter
reagent with magnesium salts in the pot, which might
diminish the electron-donating abilities of the p-methoxy
groups. Moreoever, BuMgCl is more nucleophilic than
PhMgCl and thus should have led to a faster transmeta-
lation. However, it cannot be ruled out that the observed
‘rates’ are just a result of differing lifetimes of the active
catalytic species in the presence of different Grignard
reagents.
TABLE 8. Ku m a d a Cou p lin gs of 4 w ith P h MgCl
Further mechanistic insight was expected from experi-
ments using phosphine ligands of varying electronic
properties. Additionally, the optimal amount of phos-
phine relative to nickel had yet to be determined. An
initial test reaction between chlorotoluene 4 and PhMgCl
was chosen. With PPh3 as ligand, 0-6 equiv were used
in the new coupling protocol (no LiBr, Grignard addition
at room temperature), with comparisons in yield and
conversion being made after a 1 h reaction time (Figure
2). The reaction rate26 increased continuously with
decreasing amounts of ligand resulting in a maximum
yield of 88% for product 8 when using 1 equiv PPh3, while
the amount of homocoupled product 7 increased (from
0.6% to 4.3%). Interestingly, the ligandless reaction
resulted in complete conversion but led to substantial
homocoupling (25%, along with 64% of product 8). Thus,
a compromise between extent of homocoupling and rate
was made, choosing to use 2 equiv of ligand. Given the
levels of homocoupled product found, the mechanistic
pathway leading to 7 seems to be based on nickel
complexes without participation of the specific ligand
present. If true, different ligands should not significantly
influence the amount of homocoupling observed. Indeed,
using P(p-anis)3 or P(p-tol)3 in place of PPh3, the yields
for 7 were essentially identical (Figure 2, bottom). The
only exceptions turned out to be reactions using P(p-
fluorophenyl)3 as ligand, which gave rise to lesser amounts
of homocoupling product (below 1%), although a drastic
decrease was also observed for the yields of product 8.
The drop in efficiency using this ligand was far greater
than expected and was independent of the amount of
ligand used. GCMS analysis of reaction mixtures re-
vealed involvement of the ligand itself in a Kumada
coupling with fluorine as the leaving group,1 as well as
in hydrodehalogenation and ligand scrambling.27,34 There-
fore, the nickel catalyst was diverted from its intended
role in Kumada couplings with aryl chloride 4, preventing
complete conversion under the given conditions.
addition of
entry PhMgCl at LiBr time [h] conv. [%]a 8 [%]a 7 [%]a
1
2
3
4
5
-78 °Cb
-78 °Cb
-78 °C
rt
yes
yes
no
yes
no
1.5
1
1
1
1
100
86
75
100
79
83
74
68
80
70
7.2
5.6
0.9
7.8
1.2
rt
a
By quantitative GC; reactions run with 4 equiv of PPh3
(relative to nickel). Procedure according to ref 20.
b
TABLE 9. Ku m a d a Cou p lin gs of 4 w ith (p-An isyl)MgCl
addition of
ArMgCl at
entry
LiBr
conv. [%]a
8 [%]a
7 [%]a
1
2
3
4
-78 °Cb
yes
no
yes
no
75
68
75
69
66
62
63
61c
7.1
1.6
11.2
1.8
-78 °C
rt
rt
a
By quantitative GC; reactions run with 4 equiv of PPh3
b
(relative to nickel), reaction time 1 h. Procedure according to ref
20. c Isolated yield: 58%.
TABLE 10. Ku m a d a Cou p lin gs of 4 w ith Bu MgCl
addition of
BuMgCl at
entry
LiBr
conv [%]a
8 [%]a
7 [%]a
1
2
3
4
-78 °Cb
yes
no
yes
no
74
66
85
65
49
43
60
45
6.4
2.1
7.7
2.6
-78 °C
rt
rt
a
By quantitative GC; reactions run with 4 equiv of PPh3
(relative to nickel), reaction time 1 h. Procedure according to ref
b
20.
(p-anisyl)MgCl, and BuMgCl in refluxing THF (Scheme
4). The results are listed in Tables 8-10. Some general
trends found from these experiments include: (1) The
additive LiBr (1.0 equiv relative to aryl chloride), origi-
nally used to avoid homocoupling of 4,20 appears to
actually encourage this side reaction. (2) Without added
LiBr, reactions are generally slower but cleaner. (3) The
new procedure involving Grignard addition to Ni(II)/C
at room temperature produces slightly more homocou-
pling product. (4) Of the two procedures, the one more
recently developed is faster in the presence of LiBr.
Without LiBr, both protocols result in essentially identi-
cal conversions and product formations.
The order of reaction rates26 was quite unexpected as
well: P(p-tol)3 g PPh3 g P(p-anis)3 [> P(p-fluorophenyl)3].
If oxidative addition were the rate-determining step,
higher rates would have been expected with more electron-
rich ligands. Either the rate-determining step is changing
for P(p-anis)3 or its strong electron-donating ability due
to the p-methoxy groups is altered by complexation with
Using a ratio of Ni:PPh3:4 ) 1:4:20, the reaction rate
(based on consumption of 4)26 decreases in the order
PhMgCl > (p-anisyl)MgCl g BuMgCl. Reactions run with
the latter Grignard (Table 10) seemed to be more
‘destructive’, resulting in ca. 10% (absolute percentage)
greater losses of starting material which cannot be
(34) (a) Segelstein, B. E.; Butler, T. W.; Chenard, B. L. J . Org. Chem.
1995, 60, 12. (b) Goodson, F. E.; Wallow, T. I.; Novak, B. M. J . Am.
Chem. Soc. 1997, 119, 12441. (c) Grushin, V. V. Organometallics 2000,
19, 1888. (d) Kwong, F. Y.; Lai, C. W.; Tian, Y.; Chan, K. S. Tetrahedron
Lett. 2000, 41, 10285. (e) Kwong, F. Y.; Chan, K. S. Organometallics
2001, 20, 2570.
J . Org. Chem, Vol. 68, No. 4, 2003 1195

Tasler and Lipshutz
SCHEME 5. Ku m a d a Cou p lin g of Ch lor oa n isole 9
w ith P h MgCl
TABLE 11. Ku m a d a Cou p lin gs of Ch lor oa n isole 9 w ith
P h MgCl
PR3
R )
equiv of
PR3
time conv
[h]
10
11
a
entry
LiBr
[%]b [%]b [%]b
1
2
3
4
5
6
7
8
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
p-tol
p-anisyl
p-F-Ph
2
2
2
2
2c
-
1
2
3
5
2
2
2
yes
yes
no
no
no
no
no
no
no
no
no
no
no
1.5
2.5
1
3
2
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
94
100
69
99
97
100
87
84
76
64
79
84
60
87
85
73
78
74
64
56
84
73
23
6.6
8.3
1.6
3.5
3.3
17
3.0
2.0
1.1
0.6
1.7
2.4
0.4
9
10
11
12
13
92
82
30
a
b
Relative to nickel. By quantitative GC; Grignard addition at
room temperature. c Run with 0.075 equiv of nickel instead of 0.05
equiv relative to aryl chloride.
anisole (9) was selected for a Kumada coupling with
PhMgCl (Scheme 5). In the event, the same order of
reactivity for the four different ligands (Table 11, entries
8, 11-13) was found, also confirming our earlier observa-
tions that without LiBr the reaction is slower but cleaner
(entries 1-4) and that decreasing amounts of phosphine
result in faster reactions along with more homocoupling
(entries 6-10). These findings do not allow for an
interpretation regarding the rate-determining step of
these reactions, nor can additional steps involved in the
catalytic cycle be ruled out; e.g. electron transfer leading
to reductive elimination.33 In general, reductive elimina-
tion from Ni(II) species is supposed to be relatively slow
compared to oxidative additon with Ni(0) to aryl halides,
whereas this is variable for Ni(I)/Ni(III) systems.27,35,36
For couplings with p-chloroanisole (9), the turnover
numbers (TON) for Ni/C seemed lower than those for aryl
chloride 4. That is, with LiBr added (1.0 equiv relative
to aryl chloride) an additional hour was required to
achieve full conversion of the remaining 6% educt (Table
11, cf. entries 1 and 2). Without LiBr, either a 3 h reaction
time (entry 4), or more catalyst (entry 5), was necessary
for the reaction to reach completion.
In changing the Grignard reagent from aryl to an alkyl
species in couplings with chlorotoluene 4, the reactivity
order using different ligands was essentially identical to
that observed in previous couplings (Table 12). However,
the behavior of the reaction varied dramatically as the
amount of ligand was changed (Table 13). The reaction
F IGURE 2. Dependence of Kumada couplings of 4 with
PhMgCl on amount of phosphine ligand PR3 (equiv relative
to nickel; Grignard addition at room temperature, no LiBr,
reaction time 1 h).29
magnesium salts accruing in the reaction mixture. One
approach to address this question was to use a strongly
deactivated aryl chloride so as to slow the oxidative
addition, ensuring that this step in the mechanistic
picture becomes rate-determining. Therefore, p-chloro-
(35) Amatore, C.; J utand, A. Organometallics 1988, 7, 2203.
(36) (a) Amatore, C.; J utand, A.; Mottier, L. J . Electroanal. Chem.
1991, 306, 125. (b) J utand, A.; Mosleh, A. J . Org. Chem. 1997, 62, 261.
1196 J . Org. Chem., Vol. 68, No. 4, 2003

Aromatic Aminations and Kumada Couplings
TABLE 12. In flu en ce of Liga n d s on Ku m a d a Cou p lin gs
betw een 4 a n d Bu MgCl
which the interaction of d-orbitals on nickel with the π*
LUMO of the aryl moiety facilitates electron transfer
from nickel to the arene.40 In the case of Ni(II) reduction,
interaction between d-orbitals (hybrids) and the π HOMO
in i would be required to mediate reduction of Ni(II) to
Ni(I), or a rearrangement from a π- to a σ-complex has
to precede the actual reduction step leading to formation
of Ni(I) complex iii. The newly created Ni(I) species iii
can then undergo a second oxidative addition to iv
followed by reductive elimination of the biaryl, and
regeneration of a species within the catalytic cycle by one
of several ways postulated in the literature, e.g. by
reduction to Ni(0).27,35-38 Product formation from complex
i also requires a π-/σ-rearrangement followed by reduc-
tive elimination of cross-coupled product from ii and
regeneration of Ni(0).41,42
The involvement of charge transfer π-complexes would
account for the complete loss of activity (product forma-
tion and homocoupling) in Kumada couplings with
BuMgCl in the absence of ligands, since the butyl group
cannot form such a complex.42 The small amounts of
homocoupling found with more equivalents of ligand
present in Kumada couplings with (aryl)MgCl or with
BuMgCl in general can best be explained by standard
SET from a σ-complex associated with a species stabilized
by phosphines.
PR3
R )
entry
conv [%]a
8 [%]a
7 [%]a
1
2
3
4
p-tol
Ph
p-anisyl
p-F-Ph
69
69
61
52
51
49
42
35
2.1
3.0
1.7
0.7
a
By quantitative GC; Grignard addition at room temperature,
3 equiv of PR3 (relative to nickel), no LiBr, reaction time 1 h.
TABLE 13. Dep en d en ce of Ku m a d a Cou p lin gs of 4 w ith
Bu MgCl on Am ou n t of P P h 3
equiv of
PPh3
a
entry
conv [%]b
8 [%]b
7 [%]b
1
2
3
4
5
6
7
-
1
2
20
39
66
69
65
-
20
40
49
45
66
78
-
1.1
2.7
3.0
2.6
5.5
7.0
3
4
3c
3d
87
100
a
b
Relative to nickel. By quantitative GC; Grignard addition at
room temperature, no LiBr, reaction time 1 h. c Reaction time 2
d
h. Reaction time 4 h; isolated yield: 75%.
rate increased with increasing amounts of PPh3, reaching
a maximum conversion at ca. 3 equiv (Table 13, entry
4), while the corresponding ligandless reaction afforded
no cross-coupling or homocoupled product (entry 1). With
this substrate combination, fewer side reactions involving
the fluorophenyl group in the ligand were detectable by
GCMS, thus leading to reasonable conversions of aryl
chloride (Table 12, entry 4), to be expected given the
electronic properties of this ligand. Interestingly, the
TONs for Ni/C in these reactions with BuMgCl again
appeared to be lower than those with PhMgCl, since
extended reaction times were required to go from 69 to
100% conversion (Table 13, entries 4, 6, 7), indicating
that the active nickel species had essentially lost its
activity.
Homocoupling of aryl halides, triflates, or mesylates
with catalytic amounts of nickel requires the presence
of a reducing agent to regenerate the catalytically active
species.35-38 In the case of Kumada couplings, this should
be provided by the Grignard reagent itself. The discrep-
ancy between couplings with PhMgCl and BuMgCl,
however, cannot be accounted for by differences in
reducing capabilities of these two reagents, since based
on their electrochemical potentials,39 BuMgCl should be
a more powerful reducing agent than PhMgCl. Homo-
coupling of an aryl chloride seems to be catalyzed by
nickel species nonligated by phosphine ligands (vide
supra). Hence, PhMgCl might act as a π-basic ligand for
nickel(II) following initial oxidative addition, forming a
charge-transfer complex which leads to SET to the nickel
center (Scheme 6). Such species are analogous to Ni(0)
charge-transfer complexes involving aryl halides, in
The effects observed upon addition of LiBr (1.0 equiv
relative to aryl chloride) to Kumada couplings might be
explained as follows. Stille et al. introduced additives
such as LiCl, LiBr, or LiI for coupling reactions of vinyl
triflates with organostannanes and explained their effects
by anion exchange at the palladium center after oxidative
addition, resulting in a complex more prone to participate
further in the catalytic cycle.43 With bromide or iodide
counterions on palladium, the reaction rates increased
but led to more rapid catalyst decomposition.43 In our
case, the oxidative addition directly leads to a Ni-Cl
bond, whereupon addition of bromide leads to slightly
faster reactions, as seen previously.27,43 In general, the
addition of halide additives could accelerate either oxida-
tive addition or reductive elimination by formation of
nickelate complexes,15b,21e,27 thus leading to a higher
overall rate of the reaction. Furthermore, halide exchange
would result in Ni-Br species, which facilitate reduction
at nickel,27 resulting in higher amounts of homocoupling.
In this vein, it should be recalled that the addition of
halide only influences the amount of homocoupling in the
presence of phosphines; for ligandless reactions, no
significant difference was observed (100% conversion,
60% yield of product 8, 25% homocoupling with added
LiBr compared to 100/64/25% without LiBr).29
Recent studies have shown that Kumada couplings can
be carried out at ambient temperatures.3,24,44,45 Our
attempts to couple chloride 4 with both PhMgCl and
(40) Tsou, T. T.; Kochi, J . K. J . Am. Chem. Soc. 1979, 101, 6319.
(41) For further examples of cross-coupling reactions of aryl halides
with only aromatic Grignard reagents catalyzed by nonligated nickel,
see: (a) Ikoma, Y.; Taya, F.; Ozaki, E.-i.; Higuchi, S.; Naoi, Y.; Fuji-i,
K. Synthesis 1990, 147. (b) Ikoma, Y.; Ando, K.; Naoi, Y.; Akiyama,
T.; Sugimori, A. Synth. Commun. 1991, 21, 481.
(42) NiCl2 was also used recently for cross-couplings of alkyl halides
with alkyl Grignard reagents. Similar to our results, ligandless
reactions did not provide the desired products, addition of 1,3-butadiene
derivatives as ligands was required. Terao, J .; Watanabe, H.; Ikumi,
A.; Kuniyasu, H.; Kambe, N. J . Am. Chem. Soc. 2002, 124, 4222.
(43) Scott, W. J .; Stille, J . K. J . Am. Chem. Soc. 1986, 108, 3033.
(37) (a) Percec, V.; Bae, J .-Y.; Zhao, M.; Hill, D. H. J . Org. Chem.
1995, 60, 176. (b) Amatore, C.; J utand, A.; Perichon, J .; Rollin, Y.
Monatsh. Chem. 2000, 131, 1293.
(38) Tsou, T. T.; Kochi, J . K. J . Am. Chem. Soc. 1979, 101, 7547.
(39) (a) Carloni, P.; Greci, L.; Stipa, P.; Eberson, L. J . Org. Chem.
1991, 56, 4733. (b) Bruni, P.; Carloni, P.; Conti, C.; Giorgini, E.; Greci,
L.; Iacussi, M.; Stipa, P.; Tosi, G. Tetrahedron 1996, 52, 6795.
J . Org. Chem, Vol. 68, No. 4, 2003 1197

Tasler and Lipshutz
SCHEME 6. P ossible Mech a n istic P a th w a y for Liga n d less Cr oss- a n d Hom ocou p lin gs of Ar yl Ch lor id es
BuMgCl at room temperature were unsuccessful (28 and
12% yields, respectively). Substrates bearing electrophilic
ester or nitrile substituents are also unfortunately not
as yet amenable to Ni/C-catalyzed cross-couplings with
Grignard reagents, unlike a novel palladium-catalyzed
process using imidazolium ligands16 or couplings based
on iron.45
Finally, the simplified protocol described herein for
Kumada couplings using Ni/C without prior reduction by
BuLi was tested on a larger scale reaction (15.0 mmol of
aryl chloride 4) using PhMgCl. After 1 h in refluxing
THF, the coupling was complete, with the yield of 8 (77%)
comparable to that realized in the corresponding small
scale reaction (81%).29
amination conditions, reduction of Ni(II)/C requires an
induction period of several minutes to allow for Ni(0) to
be formed. Thereafter, reaction rates are similar to
reactions mediated by preformed Ni(0)/C, suggesting that
similar nickel species are involved. Experiments with
phosphine ligands with different electronic properties led
to the conclusion that deprotonation/complexation of the
amine by preformed nickel(II) tert-butoxide is likely to
be the rate-limiting step for aromatic aminations cata-
lyzed by nickel complexes. The new protocol proved to
be applicable to transformations involving either acti-
vated and nonactivated aryl chlorides with primary or
secondary aliphatic amines. Aromatic amines were not
tested in the course of this work, but it was shown that
aniline is capable of reducing Ni(II) under these reaction
conditions. A consistent trend for Kumada couplings
based on electronic effects in the ligand was not found,
potentially due to alternative catalytic cycles involving
other species in addition to Ni(II)/Ni(0). With aromatic
Grignard reagents, ligandless Kumada couplings led to
complete consumption of the aryl chloride, yielding a 2:1
ratio of cross- and homocoupled products. This interesting
reactivity might be explained by the formation of π-com-
plexes with the aromatic residue in PhMgCl which can
stabilize nickel as a ligand and simultaneously facilitate
electron transfer to nickel, a necessary precondition for
homocoupling of an aryl halide.
Con clu sion s
Simpler procedures for utilizing Ni/C as a ‘heteroge-
neous’ catalyst in aromatic aminations and Kumada
couplings have been developed. Since Ni/C is serving as
a reservoir for active nickel in solution which is the
species involved in the catalytic cycle,14 premixing periods
of catalyst precursor Ni(II)/C with ligands prior to reduc-
tion of Ni(II) was shown to be unnecessary. Moreover, a
separate reduction of Ni(II)/C with n-BuLi is no longer
required. The conversion of Ni(II)/C to Ni(0) can be
achieved either by the amine in combination with LiO-
t-Bu (for aromatic aminations), or by the Grignard
reagent involved in Kumada couplings. Under the new
Scaled-up versions of these reactions (15.0 mmol of aryl
chloride) succeeded for both reaction types using these
new procedures. Combined with the advantages of a
‘heterogeneous’ catalyst (i.e., only traces of nickel can be
found in the product after filtration of the reaction
mixture), these protocols may be applicable to selected
industrial processes.
(44) (a) Bo¨hm, V. P. W.; Weskamp, T.; Gsto¨ttmayr, C. W. K.;
Herrmann, W. A. Angew. Chem. 2000, 112, 1672; Angew. Chem., Int.
Ed. 2000, 39, 1602. (b) Li, G. Y.; Marshall, W. J . Organometallics 2002,
21, 590.
(45) Fu¨rstner, A.; Leitner, A. Angew. Chem. 2002, 114, 632; Angew.
Chem., Int. Ed. 2002, 41, 609.
1198 J . Org. Chem., Vol. 68, No. 4, 2003

Aromatic Aminations and Kumada Couplings
Sca le-u p of Am in a tion Rea ction s. Using p-chlorobenzo-
nitrile (1, 2.06 g, 15.0 mmol) and toluene (25 mL), all other
reagents were added in identical ratios as stated in the general
procedures. Variations in reaction conditions (amount of
ligand, reaction times) and results can be found in Table 7.
Exp er im en ta l Section
Gen er a l. All reactions, and catalyst transfers, were con-
ducted under an argon atmosphere. Morpholine (2), m-chloro-
toluene (4), n-hexylamine (5), and p-chloroanisole (9) were
distilled prior to use; PPh3 was recrystallized from hexanes.
PhMgCl and BuMgCl (both 1.8 M in THF) and n-BuLi (2.1 M
in hexanes) were obtained from commercial sources and
titrated before use. All reactions were analyzed by quantitative
GC using 1,3,5-trimethoxybenzene and p-fluorotoluene as
standards for aromatic aminations and Kumada couplings,
respectively. GC response factors were determined using
commercially available compounds [3,3′-bistoluene (7), 3-phe-
nyltoluene (8, R ) Ph), 4-phenylanisole (10), 4,4′-bisanisole
(11)] or isolated materials [N-(p-cyanophenyl)morpholine (3),
N-(p-cyanophenyl)-n-hexylamine, N-(m-tolyl)morpholine, N-
(m-tolyl)-n-hexylamine (6; the three previous amines), 3-butyl-
toluene (8, R ) Bu), 4-(3-tolyl)anisole (8, R ) p-MeO-Ph)]. For
chromatographic purposes, 200-425 mesh Fisher Brand silica
gel was used. Unequivocal identification of products 7, 8 (R )
Ph), 10, and 11 was accomplished by GC coinjection, GCMS
analyses, and TLC comparison with authentic materials.
P r ep a r a tion of Ni(II)/C. Activated charcoal (5.00 g) was
impregnated with Ni(NO3)2‚6H2O (727 mg, 2.50 mmol) in H2O
(75 mL).46
Gen er a l P r oced u r e for Ni/C-Ca ta lyzed Ar om a tic Am i-
n a tion s of p-Ch lor oben zon itr ile (1) w ith Mor p h olin e (2)
[p r ior r ed u ction of Ni(II) w ith n -Bu Li]. These reactions
were performed according to a literature protocol,19 using 0.750
mmol of 1 (103 mg), 2.0 equiv of 2 (132 µL, 131 mg, 1.50 mmol),
and toluene (1.4 mL, total). Generally, 0.05 equiv of Ni/C (63.1
mg, 0.038 mmol, loading: 0.594 mmol/g), 0.025 equiv of dppf
(10.4 mg, 0.019 mmol) or 0.2 equiv of PPh3 (39.3 mg, 0.150
mmol), and LiO-t-Bu (72.1 mg, 0.900 mmol) were added. The
premixing period of catalyst and ligand in toluene was
shortened from 90 to 25 min and the time for reduction after
addition of BuLi (35.7 µL, 0.075 mmol, 2.1 M) was lowered
from 30 to 15 min. After each reaction (see Tables), the mixture
was cooled to room temperature, and with stirring was added
CH2Cl2 (3 mL). The suspension was filtered through a Bu¨chner
funnel and the cake washed with CH2Cl2, petroleum ether (10
mL each), and diethyl ether (25 mL). The filtrate was refil-
tered, and the solvent volume was reduced to ca. 1 mL, after
which it was filtered through a pipet filled with glass wool,
Celite, and silica gel. The GC standard was added to the
resulting clear solution, and the extent of conversion and
percent product formed were determined by GC. Variations
in reaction conditions (ligand, amount of ligand, nickel source,
reaction times) and results for different reactions can be found
in Tables 1, 3 and Tables S1, S2 in the Supporting Information.
Gen er a l P r oced u r e for Ar om a tic Am in a tion s of p-
Ch lor oben zon itr ile (1) w ith Mor p h olin e (2) [with ou t
p r ior r ed u ction of Ni(II) w ith Bu Li]. To a mixture of Ni/C
(63.1 mg, 0.038 mmol, loading: 0.594 mmol/g), dppf (10.4 mg,
0.019 mmol) or PPh3 (39.3 mg, 0.150 mmol), LiO-t-Bu (72.1
mg, 0.900 mmol), and p-chlorobenzonitrile (1, 103 mg, 0.750
mmol) in a Schlenck tube were added toluene (1.4 mL) and
morpholine (2, 132 µL, 131 mg, 1.50 mmol). The tube was
placed in a preheated oil bath (130 °C) and kept at reflux for
the time indicated in the Tables. Workup and analysis of the
product mixtures obtained were identical to the procedure
above. Variations of reaction conditions (ligand, amount of
ligand, nickel source, reaction times) and results for different
reactions can be found in Tables 1, 3 and Tables S1, S2 in the
Supporting Information..
Ku m a d a Cou p lin gs of m -Ch lor otolu en e (4) a n d Gr ig-
n a r d Rea gen ts [p r ior r ed u ction of Ni(II)/C w ith Bu Li].
Following a literature procedure,20 aryl chloride 4 (118 µL, 127
mg, 1.00 mmol) was treated with a Grignard (PhMgCl or
BuMgCl: 835 µL, 1.50 mmol, 1.8 M) at -78 °C after reduction
of 0.05 equiv of Ni(II)/C catalyst precursor (84.2 mg, 0.050
mmol, loading: 0.594 mmol/g) by BuLi (47.6 µL, 0.100 mmol,
2.1 M) in the presence of 0.2 equiv of PPh3 (52.5 mg, 0.200
mmol) and 1.0 equiv of LiBr (86.9 mg, 1.00 mmol) in THF (1.5
mL), and then heated at reflux. After a selected reaction time
(see Tables), the mixture was cooled to room temperature and
upon stirring, 3 mL of aq NH4Cl solution were carefully added.
The suspension was filtered through a Bu¨chner funnel and
the cake was washed with H2O, diethyl ether, petroleum ether,
CH2Cl2 (each 10 mL), and again with diethyl ether (25 mL).
The biphasic filtrate was shaken, the aqueous phase was
removed with a pipet, and the organic layer was concentrated
under reduced pressure. A GC standard was added to the
resulting clear solution and the extent of conversion and yield
of product analyzed by GC. Variations of reaction conditions
and results for different reactions can be found in Tables 8-10.
Ku m a d a Cou p lin gs of m -Ch lor otolu en e (4) a n d Gr ig-
n a r d Rea gen ts [with ou t p r ior r ed u ction of Ni(II) w ith
Bu Li]. To a mixture of Ni(II)/C (84.2 mg, 0.050 mmol,
loading: 0.594 mmol/g) and PPh3 (52.5 mg, 0.200 mmol) in
THF (1.5 mL) were added m-chlorotoluene (4, 118 µL, 127 mg,
1.00 mmol) and RMgCl (1.50 mmol) at room temperature, and
the suspension was placed in a preheated oil bath (75 °C) and
kept at reflux. No LiBr was used in this case. After a selected
reaction time (see Tables), workup and analyses were per-
formed as described above. Variations in reaction conditions
(especially amounts of ligand) and results for different reac-
tions can be found in Tables 8-10, 12, 13, and Tables S3-S6
in the Supporting Information. The best compromise between
reaction rate and amount of homocoupled side product was
obtained with 2.0 equiv of PPh3 (26.3 mg, 0.100 mmol) relative
to nickel using ArMgCl, and 3.0 equiv of PPh3 (39.3 mg, 0.150
mmol) when using BuMgCl.
Sca le-u p of Ku m a d a Cou p lin gs. Using m-chlorotoluene
(4, 1.77 mL, 1.90 g, 15.0 mmol) and THF (20 mL), all other
reagents were added in identical ratios as stated in the general
procedure without prior reduction of Ni(II)/C with n-BuLi,
except for PPh3 which was used only in a 2:1 ratio relative to
nickel. Results can be found in Table S3, Supporting Informa-
tion (100% conversion, 77% 8, 5.4% 7).
Ack n ow led gm en t. Financial support provided by
the NIH (GM 40287) and the DAAD (fellowship to S.
Tasler, Hochschulsonderprogramm III) is gratefully
acknowledged. Thanks also go to Randy Vivian and Will
Chrisman for inspiring discussions.
Su p p or tin g In for m a tion Ava ila ble: Further experimen-
tal details of amination and Kumada reactions, isolation, and
characterization of all products. This material is available free
of charge via the Internet at http://pubs.acs.org.
(46) Lipshutz, B. H.; Tasler, S. Adv. Synth. Catal. 2001, 343, 327.
J O020297E
J . Org. Chem, Vol. 68, No. 4, 2003 1199