Rationally designed pincer-type heck catalysts bearing aminophosphine substituents: PdIV intermediates and palladium nanoparticles

Article abstract of DOI:10.1002/chem.200800441

The aminophosphine-based pincer complexes [C₆H ₃-2,6-{XP(piperidinyl)₂}₂Pd(Cl)] (X = NH 1; X = O 2) are readily prepared from cheap starting materials by sequential addition of 1,1′,1″-phosphinetriyltripiperidine and 1,3-diaminobenzene or resorcinol to solutions of [Pd(cod)(Cl)₂] (cod = cyclooctadiene) in toluene under N₂ in "one pot". Compounds 1 and 2 proved to be excellent Heck catalysts and allow the quantitative coupling of several electronically deactivated and sterically hindered aryl bromides with various olefins as coupling partners at 140°C within very short reaction times and low catalyst loadings. Increased reaction temperatures also enable the efficient coupling of olefins with electronically deactivated and sterically hindered aryl chlorides in the presence of only 0.01 mol% of catalyst. The mechanistic studies performed rule out that homogeneous Pd⁰ complexes are the catalytically active forms of 1 and 2. On the other hand, the involvement of palladium nanoparticles in the catalytic cycle received strong experimental support. Even though pincer-type Pd^IV intermediates derived from 1 (and 2) are not involved in the catalytic cycle of the Heck reaction, their general existence as reactive intermediates (for example, in other reactions) cannot be excluded. On the contrary, they were shown to be thermally accessible. Compounds 1 and 2 show a smooth halide exchange with bromobenzene to yield their bromo derivatives in DMF at 100°C. Experimental observations revealed that the halide exchange most probably proceeded via pincer-type Pd^IV intermediates. DFT calculations support this hypothesis and indicated that aminophosphine-based pincer-type Pd^IV intermediates are generally to be considered as reactive intermediates in reactions with aryl halides performed at elevated temperatures.

Full text of DOI:10.1002/chem.200800441

FULL PAPER
DOI: 10.1002/chem.200800441
Rationally Designed Pincer-Type Heck Catalysts Bearing Aminophosphine
Substituents: PdIV Intermediates and Palladium Nanoparticles
Jeanne L. Bolliger, Olivier Blacque, and Christian M. Frech*[a]
Abstract: The aminophosphine-based
pincer complexes [C6H3-2,6-{XP(piperi-
dinyl)2}2Pd(Cl)] (X=NH 1; X=O 2)
are readily prepared from cheap start-
ing materials by sequential addition of
1,1’,1’’-phosphinetriyltripiperidine and
1,3-diaminobenzene or resorcinol to
ficient coupling of olefins with elec-
tronically deactivated and sterically
hindered aryl chlorides in the presence
of only 0.01 mol% of catalyst. The
mechanistic studies performed rule out
that homogeneous Pd0 complexes are
the catalytically active forms of 1 and
2. On the other hand, the involvement
of palladium nanoparticles in the cata-
lytic cycle received strong experimental
support. Even though pincer-type PdIV
intermediates derived from 1 (and 2)
are not involved in the catalytic cycle
of the Heck reaction, their general ex-
istence as reactive intermediates (for
example, in other reactions) cannot be
excluded. On the contrary, they were
shown to be thermally accessible. Com-
pounds 1 and 2 show a smooth halide
exchange with bromobenzene to yield
their bromo derivatives in DMF at
1008C. Experimental observations re-
vealed that the halide exchange most
probably proceeded via pincer-type
PdIV intermediates. DFT calculations
support this hypothesis and indicated
that aminophosphine-based pincer-type
PdIV intermediates are generally to be
considered as reactive intermediates in
reactions with aryl halides performed
at elevated temperatures.
solutions of [PdACHTREUNG(cod)(Cl)2] (cod=cy-
clooctadiene) in toluene under N2 in
“one pot”. Compounds 1 and 2 proved
to be excellent Heck catalysts and
allow the quantitative coupling of sev-
eral electronically deactivated and ster-
ically hindered aryl bromides with vari-
ous olefins as coupling partners at
1408C within very short reaction times
and low catalyst loadings. Increased re-
action temperatures also enable the ef-
Keywords: aminophosphines · C–C
coupling · nanoparticles · palladi-
um · pincer reagents
Introduction
Although recent developments have achieved a considera-
ble increase in the activity of Heck catalysts, a typical proto-
col for this reaction still requires prolonged reaction times
and relatively high catalyst loadings, and there is still a clear
need for more efficient systems. Pincer complexes of palladi-
um are among the most efficient Heck catalysts and contin-
uously attract attention because of their unique balance be-
tween stability and reactivity. Seemingly slight electronic
and steric modifications of the pincer core and/or the phos-
phine substituents have been demonstrated to dramatically
influence their catalytic activities.[2a,d,6,7] Although nowadays
pincer complexes are in most cases considered as depot
forms of palladium nanoparticles, the involvement of PdIV
intermediates in the catalytic cycle still cannot be excluded
completely.[8–10]
The palladium-catalyzed arylation of alkenes has proven to
be one of the most important methods for carbon–carbon
bond formation in organic chemistry.[1] Various types of pal-
ladium complexes have been employed to promote the
Heck reaction, but even though some efficiently couple ster-
ically hindered substrates or occasionally even aryl chlor-
ides, their syntheses are often time-consuming, difficult, and/
or require the use of expensive starting materials. In addi-
tion, many catalytically active systems suffer from their poor
thermal stability, as well as poor stability towards air and
moisture.[2–5]
We report herein the catalytic activity of aminophos-
phine-based pincer complexes of palladium with the general
formula [(C6H3-2,6-{XP(piperidinyl)2}2Pd(Cl)] (X=NH 1;
X=O 2) in the arylation of olefins. Aminophosphine-based
systems were chosen because of their high s-donor strength
and the possibility to facilitate the accessibility of PdIV inter-
[a] J. L. Bolliger, Dr. O. Blacque, Dr. C. M. Frech
Department of Inorganic Chemistry
University of Zürich, 8057 Zürich (Switzerland)
Fax : (+41)44-635-6802
E-mail: chfrech@aci.uzh.ch
Chem. Eur. J. 2008, 14, 7969 – 7977
ꢀ 2008 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
7969

mediates by additional electron donation through the nitro-
gen lone pairs. On the other hand, if pincer complexes are
stable and clean sources of palladium nanoparticles, the ami-
nophosphines should promote their formation and lead to
enhanced catalytic activities when compared with their
phosphine and phosphite analogues. Our results indicate
that homogeneous Pd0 species derived from 1 (and 2) are
not the active forms of the catalysts, whereas the involve-
ment of palladium nanoparticles in the catalytic cycle re-
ceived strong experimental support. Although pincer-type
PdIV intermediates are not involved in Heck reactions cata-
lyzed by 1 (and 2) and also never have been detectable with
other pincer-type complexes, their general accessibility
cannot be excluded in reactions with aryl halides performed
at elevated temperatures. In contrast, our experimental find-
ings clearly indicate that pincer-type PdIV intermediates (de-
rived from 1 and 2) are accessible and thus, should be con-
sidered as intermediates in reactions involving aryl halides
at elevated temperatures.
bromotoluene (Table 1, entries 5–8). A decrease in activity
was observed by using 2-bromo-m-xylene as substrate, for
which 95% conversion was obtained after 8 h (Table 1,
entry 9). Heck reactions performed with N,N-dimethyl acryl-
amide exhibit very similar conversion rates and yields to
those with styrene (Table 1, entries 10–16). Complete prod-
uct formation and excellent selectivities but slightly retarded
conversion rates were observed with n-butyl acrylate as cou-
pling partner (Table 1, entries 17–21). For instance, using de-
activated 4-bromoanisole or sterically hindered 2-bromoto-
luene as substrate led to quantitative (ꢀ97%) product for-
mation within 4.5 h in the presence of only 0.005 mol% of
1. A conversion of 65% was observed after 12 h when 2-
bromo-m-xylene was coupled with n-butyl acrylate (Table 1,
entry 22). Quantitative product formation but further retar-
dation of the conversion rates accompanied by low selectivi-
ty was observed after 5–8 h with 0.005 mol% of 1 with the
electronically deactivated n-butyl vinyl ether (Table 1, en-
tries 23–28). Even the sterically hindered 2-bromo-m-xylene
was converted to 72% of product after only 8 h (Table 1,
entry 28). When the amount of catalyst was increased to
0.02 mol%, 4-vinylpyridine undergoes quantitative coupling
with bromobenzene, electronically deactivated 4-bromoani-
sole, and sterically hindered 2-bromotoluene within 8–12 h
(Table 1, entries 29–31). An 85% yield was obtained within
44 h when 2-bromo-m-xylene was used as coupling partner
(Table 1, entry 32). Heck reactions performed with 4-vinyl-
pyridine are slower due to the ligating property of the sub-
strate, as shown by reactions between [(C6H3-
Results and Discussion
The aminophosphine-based pincer complexes 1 and 2 were
readily prepared by the sequential addition of 1,1’,1’’-phos-
phinetriyltripiperidine and 1,3-diaminobenzene or resorcinol
to solutions of [PdACHTREUNG(cod)(Cl)2] (cod=cyclooctadiene) in tolu-
ene under N2 in “one pot”.[11] Removal of the volatiles
under reduced pressure and subsequent extractions with di-
ethyl ether gave pure 1 and 2 in high yields.[7]
ACHRTE{UGN NHP(piperidinyl)2}2Pd]ACHRTE[UGN BF4] (3) and pyridine, which result-
ed in its stable adduct 4.[13] Further retardation was noticed
with 2-vinylpyridine (Table 1, entry 33), most probably due
to chelation.
The exceptional high catalytic activity of 1 was demon-
strated further in an exemplary “large-scale” reaction, in
which bromobenzene (210 mL; 2.0 mol) and styrene
(250 mL; 2.4 mol) were coupled in the presence of only
0.00002 mol% of catalyst. Quantitative conversion was ach-
ieved after 36 h (Table 1, entry 3). When the reaction tem-
perature was raised to 1608C, 1,1’,1’’-ethene-1,1,2-triyltriben-
zene was quantitatively formed within 20 h on addition of
1.1 equivalents of bromobenzene to solutions of (E)-stilbene
in DMF (Table 1, entry 34). Notably, the same product was
formed quantitatively within 24 h under identical reaction
conditions by either using 1,1-diphenylethene as substrate,
or by adding 2.2 equivalents of bromobenzene to solutions
of styrene in DMF.
Remarkably, in contrast to Heck reactions performed
with aryl bromides, catalysts 1 and 2 show the same level of
activity with aryl chlorides as substrates.[14] For instance,
when in the presence of 0.01 mol% of catalyst and about
15% of tetrabutylammonium bromide in 1-methyl-2-pyrroli-
done (NMP) at 1608C, the electronically activated 4-chloro-
Both complexes show exceptional high activity in the cat-
alytic arylation of olefins with aryl bromides, leading to very
high conversion rates and quantitative yields in short reac-
tion times using low catalyst loadings, even for the use of
electronically deactivated and sterically hindered substrates
(Table 1). In Heck reactions performed with aryl bromides,
catalyst 2 is generally less active than 1.[12] For example, bro-
À
mobenzene and styrene underwent complete C C coupling
in the presence of only 0.002 mol% of 1 and K2CO3 within
2.5 h in DMF at 1408C, whereas a reaction time of 10 h was
necessary with catalyst 2 (Table 1, entries 1 and 2). Com-
plete conversion of 1,3-dibromobenzene and styrene into 1-
[2-phenylvinyl]-3-[2-phenylvinyl]benzene was achieved after
3.5 h (Table 1, entry 4). The same level of activity was ob-
served with the electronically deactivated 4-bromoanisole
and 4-methoxystyrene, as well as the sterically hindered 2-
AHCTREaUNG cetophenone and N,N-dimethyl acrylamide or styrene were
coupled almost quantitatively to (2E)-3-(4-acetylphenyl)-
N,N-dimethylprop-2-enamide and 1-{4-[(E)-2-phenylethe-
nyl]phenyl}ethanone, respectively, within 2.5 h (Table 2, en-
7970
www.chemeurj.org
ꢀ 2008 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 7969 – 7977

Pincer-Type Heck Catalysts
FULL PAPER
Table 1. Heck coupling of aryl bromides with various olefins catalyzed by [C6H3-2,6-{XP(piperidinyl)2}2Pd(Cl)] (X=NH 1; X=O 2).[a]
Entry
Aryl halide
Olefin
Cat. (ppm)
Conv. [%][b]
t [h]
TOF[c]
TON[d]
(cis/trans/gem)
1
2
bromobenzene
bromobenzene
bromobenzene
1,3-dibromobenzene
4-bromoanisole
4-bromoanisole
bromobenzene
2-bromotoluene
2-bromo-m-xylene
bromobenzene
bromobenzene
1,3-dibromobenzene
4-bromoanisole
4-bromoanisole
2-bromotoluene
2-bromo-m-xylene
bromobenzene
bromobenzene
1,3-dibromobenzene
4-bromoanisole
2-bromotoluene
2-bromo-m-xylene
bromobenzene
bromobenzene
1,3-dibromobenzene
4-bromoanisole
2-bromotoluene
2-bromo-m-xylene
bromobenzene
4-bromoanisole
2-bromotoluene
2-bromo-m-xylene
bromobenzene
bromobenzene
styrene
styrene
styrene
styrene
styrene
styrene
4-methoxystyrene
styrene
styrene
1 (20)
2 (20)
1 (0.2)
1 (20)
1 (20)
2 (20)
1 (20)
1 (20)
1 (50)
1 (20)
2 (20)
1 (20)
1 (20)
2 (20)
1 (20)
1 (50)
1 (50)
2 (50)
1 (50)
1 (50)
1 (50)
1 (50)
1 (50)
2 (50)
1 (50)
1 (50)
1 (50)
1 (50)
1 (200)
1 (200)
1 (200)
1 (200)
1 (200)
1 (50)
>99 (0/10/1)
96 (0/10/1)
>99 (0/10/1)
>99 (1/7/0)[f]
99 (0/10/1)
97 (0/10/1)
97 (0/10/1)
98 (0/20/1)
95 (2/80/1)
100 (1/20/0)
100 (1/25/0)
>99 (1/10/0)[f]
99 (1/20/0)
95 (2/80/1)
100 (1/40/0)
79 (0/1/0)
100 (1/100/1)
93 (1/100/0)
>99 (1/20/0)[f]
99 (1/100/0)
97 (1/200/0)
65 (1/100/0)
100 (2/4/3)
99 (2/4/3)
2.5
10
36
3.5
2.5
11
2.5
2.5
8
19880
4800
138333
14157
19800
4409
19400
19600
2375
25000
5000
19872
19800
3958
12500
1975
4444
775
4436
4400
4311
1083
4000
1238
2493
3840
3267
1800
589
49700
48000
4980000
49550
49500
48500
48500
49000
19000
50000
50000
49680
49500
47500
50000
15800
20000
9300
19960
19800
19400
13000
20000
19800
19940
19200
19600
14400
5000
3[e]
4
5
6
7
8
9[g]
10
N,N-dimethyl acrylamide
N,N-dimethyl acrylamide
N,N-dimethyl acrylamide
N,N-dimethyl acrylamide
N,N-dimethyl acrylamide
N,N-dimethyl acrylamide
N,N-dimethyl acrylamide
n-butyl acrylate
n-butyl acrylate
n-butyl acrylate
n-butyl acrylate
n-butyl acrylate
n-butyl acrylate
n-butyl vinyl ether
n-butyl vinyl ether
n-butyl vinyl ether
n-butyl vinyl ether
n-butyl vinyl ether
n-butyl vinyl ether
4-vinylpyridine
4-vinylpyridine
4-vinylpyridine
4-vinylpyridine
2-vinylpyridine
(E)-stilbene
2
11
12
13
14
10
2.5
2.5
12
4
15
16[g]
17
8
4.5
12
4.5
4.5
4.5
12
5
16
8
5
18
19
20
21
22[g]
23
24
25
26
27
>99[h]
96 (1/2/3)
98 (2/4/3)
72 (5/3/5)
100 (2/50/0)
>99 (1/20/0)
100 (1/20/0)
85 (1/40/0)
53 (1/10/0)
98
6
8
28[g]
29[g]
30[g]
31[g]
32[g]
33[g]
34[i]
8.5
8.5
12
44
60
20
585
417
97
44
4972
5000
4250
2650
980
19600
[a] Reaction conditions: 4.0 mmol aryl halide, 4.4 mmol olefin, 4.4 mmolK2CO3, 5 mL DMF, catalyst (synthesized in one pot and used without purifica-
tion) added in solution (toluene), reaction performed at 1408C under N2 atmosphere. [b] Determined by GC/MS, based on aryl halide. [c] Defined as
mol product per mol of catalyst per hour. [d] Defined as mol product per mol of catalyst. [e] 2.0 mol aryl halide, 2.4 mol olefin, 2.4 mol K2CO3, 1 L DMF.
[f] Product distribution refers to (cis–trans/trans–trans/gem–trans). [g] Reactions performed in NMP. [h] All possible isomers with similar yields were ob-
tained. [i] Reaction performed at 1608C.
tries 1–4). When the reaction temperature was raised to
2008C, even nonactivated, deactivated, and ortho-substitut-
ed aryl chlorides were successfully coupled with olefins. For
example, reactions performed with chlorobenzene and N,N-
dimethyl acrylamide afforded the coupling product in 77%
yield in the presence of catalyst 1 and in 91% yield with cat-
alyst 2 after 16 h (Table 2, entries 5 and 6). Remarkably,
even the sterically hindered 2-chloro-m-xylene was convert-
ed to about 60% of the product after 28 h in the presence
of 1 with N,N-dimethyl acrylamide as coupling partner
(Table 2, entry 7). A prolonged reaction time was required
with the electronically deactivated 4-chloroanisole as sub-
strate (Table 2, entry 8). A conversion of 80% was achieved
within 12 h when chlorobenzene was coupled with 4-methyl-
styrene (Table 2, entry 9). Even higher conversions were ob-
served after 18 h when chlorobenzene or 4-chlorotoluene
were allowed to react with 4-methylstyrene or 4-methoxy-
styrene as coupling partners (Table 2, entries 10–13).
lyst solutions are readily prepared from very cheap starting
materials in “one pot”, whose solutions in toluene can be
used directly for catalytic reactions without purification. The
catalyst solutions remain stable for several months at room
temperature, affording the coupling products at essentially
the same conversion rates and yields as freshly prepared cat-
alyst solutions from pure 1 and 2, respectively. In addition, 1
(and to a minor extent 2) are more efficient in the majority
of the coupling reactions performed with aryl bromides than
the reference systems [(PCNHCP)Pd(Cl)]Cl (PCNHCP=1,3-
bis[2-(diphenylphosphanyl)ethyl]-1H-imidazol-3-ium-2-
ide),[15] [Pd
(TFA=trifluoroacetate)
A
ACHTREUNG
(NHCMe)Pd(Br)]+ (NHCMe =3,3’-(pyridine-2,6-diyldimetha-
nediyl)bis(1-methyl-1H-imidazol-3-ium-2-ide)[17] and also
Herrmannꢁs systems such as [(NHCMe)2Pd], [(C6H4-2-{CH2P-
A
N
(PPh3)Pd(Cl)2].[18] Comparisons
ACHTREUNG
C
ACHTRE(UNG tBu)3 (dba=dibenzyl-
Overall, 1 and 2 belong to the most active and most con-
venient Heck catalysts reported up to date, since their cata-
ideneactone)[3b,4] are difficult, since the Heck reactions were
generally performed at room temperature. Similarly, Heck
Chem. Eur. J. 2008, 14, 7969 – 7977
ꢀ 2008 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
www.chemeurj.org
7971

C. M. Frech et al.
Table 2. Heck coupling of aryl chlorides with various olefins catalyzed by [C6H3-2,6-{XP(piperidinyl)2}2Pd(Cl)] (X=NH 1; X=O 2).[a]
Entry
Aryl halide
Olefin
Cat.
Conv. [%][b]
t [h]
TOF[c]
TON[d]
(cis/trans/gem)
1[e]
2[e]
3[e]
4[e]
5
6
7
8
9
10
11
12
13
4-chloroacetophenone
4-chloroacetophenone
4-chloroacetophenone
4-chloroacetophenone
chlorobenzene
chlorobenzene
2-chloro-m-xylene
4-chloroanisole
chlorobenzene
4-chlorotoluene
chlorobenzene
chlorobenzene
styrene
styrene
1
2
1
2
1
2
1
1
1
1
1
2
2
100 (5/100/1)
95 (0/1/0)
99 (0/1/0)
96 (0/20/1)
77 (3/100/1)
91 (0/60/1)
57 (0/1/0)
66 (0/1/0)
82 (1/80/10)
80 (1/7/0)
90 (1/100/10)
100 (1/100/10)
98 (0/8/1)
2.5
2.5
2.5
2.5
16
16
28
72
12
18
18
18
18
4000
3800
3960
3840
583
683
204
92
683
444
500
556
544
10000
9500
9900
9600
7000
8200
5700
6600
8200
8000
9000
10000
9800
N,N-dimethyl acrylamide
N,N-dimethyl acrylamide
N,N-dimethyl acrylamide
N,N-dimethyl acrylamide
N,N-dimethyl acrylamide
N,N-dimethyl acrylamide
4-methylstyrene
4-methylstyrene
4-methoxystyrene
4-methoxystyrene
4-methoxystyrene
4-chlorotoluene
[a] Reaction conditions: 4.0 mmol aryl halide, 6.0 mmol olefin, 4.4 mmolK2CO3, 0.6 mmol tetrabutylammonium bromide, 5 mL NMP, 0.01 mol% of cata-
lyst (synthesized in one pot and used without purification) added in solution (toluene), reaction performed at 2008C under N2 atmosphere. [b] Deter-
mined by GC/MS, based on aryl halide. [c] Defined as mol product per mol of catalyst per hour. [d] Defined as mol product per mol of catalyst. [e] Re-
action performed at 1608C.
reactions performed with aryl chlorides are difficult to com-
pare because most of the reference systems, such as [C6H3-
[3b]
2,6-(OPiPr2)2Pd(Cl)],[2d] [Pd
2ACHTERNGU(dba)3]/PCAHTRE(UGN tBu)3 and [Pd(Cl)2]/
[19]
PR3 perform the Heck reaction at different reaction tem-
peratures.
The change of base or solvent strongly influences the con-
version rates and yields: best results were observed with
K2CO3 or K3PO4 as base, whereas sodium acetate, sodium
hydroxide, or potassium tert-butoxide were not appropriate,
since they lead to fast deposition of inactive palladium
black. Also the use of amines, such as NEt3, is not practical
due to their ligating properties and inhibits catalysis effi-
ciently. Similarly, while comparable activities were found in
DMF and NMP, replacement of these solvents by p-xylene
or DMSO led to a dramatic drop in activity.[20] Lowering the
reaction temperature has the same effect: reactions per-
formed with bromobenzene and styrene in DMF in the pres-
ence of 0.002 mol% of catalyst 1 only led to about 7% con-
version after 3 h at 1008C. Almost no catalytic activity was
observed below 808C. Catalyst concentrations of
>0.1 mol% are also not appropriate and led to fast deposi-
tion of inactive palladium black.
Accordingly, experiments were carried out to gain mecha-
nistic insights. The addition of mercury to reaction mixtures
of aryl bromides, olefins, and 1 (or 2) efficiently inhibit cata-
lysis.[8f–h,21,22] Sigmoidal-shaped kinetics with induction peri-
ods between 30 and 45 min were observed in all the cou-
pling reactions performed with aryl bromides (Figure 1).
The same observations were made with [C6H3-2,6-
Figure 1. Kinetics of the coupling reaction of 4-bromoanisole with styrene
in the presence and in the absence of additives, catalyzed by 0.002 mol%
of 1 at 1408C in DMF (Table 1, entry 5).
coupling partners: whereas only low conversions were ob-
served in the absence of tetrabutylammonium bromide (fast
deposition of inactive palladium black was observed), signif-
icantly improved conversion rates and yields were found
when about 15 mol% of tetrabutylammonium bromide was
present in the reaction mixtures. The presence of 0.3 equiva-
lents of thiophene or 0.05 equivalents of PPh3 relative to the
catalyst neither had an influence on the conversion rate nor
on the yield.[26,27] This is in sharp contrast to CS2, which ef-
fectively stopped catalysis when 0.5 equivalents (relative to
catalyst) were present. The effect of CS2 on the performance
in the Heck reaction catalyzed by 1 (and 2) is due to catalyst
{NHP(piperidinyl)2}2Pd]ACHTREU[NG BF4] (3). The addition of 2–3 drops
of water to the reaction mixtures (to promote the formation
of palladium nanoparticles) leads to significantly decreased
induction periods.[23] Whereas the induction period remains
unchanged when tetrabutylammonium bromide (ca. 15
mol% relative to the aryl halide) was added to catalytic re-
actions performed with aryl bromides, retarded conversion
rates were noticed.[24,25] This is in sharp contrast to the ob-
servations made at 2008C in NMP with aryl chlorides as
À
degradation involving P N bond cleavage and (partial) for-
mation of the bis(piperidine-1-carbodithioato)palladium
complex [Pd(S2CNpiperidyl)2] (5), which shows no catalytic
activity in the Heck reaction. Indeed, reactions of 1 (and 2)
with an excess (ca. 20 equiv) of CS2 at room temperature
yielded complex 5 in almost quantitative yield. The identity
7972
www.chemeurj.org
ꢀ 2008 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 7969 – 7977

Pincer-Type Heck Catalysts
FULL PAPER
1
of 5 was confirmed by Hand 13C{1H} NMR spectroscopy as
DMF containing either an excess of benzyl chloride or reac-
tion mixtures of benzyl chloride, bromobenzene, styrene,
and K2CO3) no dibenzyl was detectable after 4 h at 1408C,
offering the possibility to trace homogeneous Pd0 complexes
by dibenzyl detection under catalytic reaction conditions.[34]
Since in the Heck reactions catalyzed by 1 (or 2) no diben-
zyl was detectable,[35] the involvement of complexes with Pd0
centers as the active form of the catalysts were excluded.
The same results were obtained with 6 (and 14), which
thereby rules out that homogeneous Pd0 complexes are the
catalytically active form of phosphine- and phosphite-based
pincer complexes. This is in line with previous investigations
performed on 6 and 14.[2a,b,h]
well as by X-ray diffraction crystallographic analysis.[13,28]
Under catalytic reaction conditions, piperidine-1-carbodi-
thioate probably dissociates from 5, and blocks the active
centers of the palladium nanoparticles, if formed. Overall, 1,
2, (and 3) most probably are pre-catalysts and palladium
nanoparticles are their active form. This would be in line
with the fact that neither 1 nor 2 (nor their bromo deriva-
tives) were detectable in the reaction mixtures after catalysis
(different catalyst concentrations were tested) and the ob-
servation that the turnover frequency (TOF) increases with
increasing substrate/catalyst ratios.[29]
To test if transformations of 1 (or 2) into homogeneous
Pd0 complexes are involved in the catalytic cycle, the follow-
ing new method to trace such species under catalytic reaction
conditions was developed. Benzyl chloride is known to
induce clean and fast electron transfers from Pd0 complexes,
thereby forming dibenzyl. For instance, [C6H3-2,6-
(CH2PiPr2)2Pd(Cl)] (6) undergoes collapse of the pincer
framework to form a binuclear Pd0/PdII complex [Pd{C6H3-
Although PdIV intermediates derived from 1 (and 2) are
not involved in the catalytic cycle in the Heck reaction, it is
important to mention that their thermal accessibility re-
ceived strong experimental and computational (see below)
support. Reaction mixtures of 1 (and 2) and bromobenzene
show a halide exchange in DMF (or NMP) at 1008C, lead-
ing to their bromo derivatives [C6H3-2,6-{XP(piperidi-
nyl)2}2Pd(Br)] (X=NH 12; X=O 13) and phenyl chloride,
as detected by GC/MS.[36,37] In analogy to the catalytic aryla-
tion of olefins, lowering the reaction temperatures reduces
the conversion rates. Similarly, the halide exchange is
strongly solvent dependent and exhibits the highest conver-
sion rates in DMF and NMP, whereas the lowest rates are
observed in p-xylene and DMSO.[39] This observation could
be interpreted by the formation of ion pairs, such as [C6H3-
2,6-ACHTREUNG(CH2PiPr2)2}Pd] (7) incorporating a 14-electron linear
Pd0 center and a “butterfly”-type PdII 16-electron center
under strongly reducing reaction conditions. Subsequent ad-
dition of an excess (5.0 equiv) of benzyl chloride to solutions
of 7 in DMF (or NMP) instantly induced an electron trans-
fer, regenerating the pincer complex 6 accompanied by the
formation of a corresponding amount of dibenzyl.[31] Similar-
ly, reactions between benzyl chloride and [PdACHTRE(UGN PPh3)4] (8),
[Pd(C6H5CH2PiPr2)2][32] (9) or [Pd{P(piperidinyl)3}2][13] (10),
which was prepared in high yields (93%) by direct reduction
of [Pd{P(piperidinyl)3}2(Cl)2] (11) with sodium metal over-
night, cleanly gave their PdII chloro derivatives and a corre-
sponding amount of dibenzyl, as detected by GC/MS. An X-
ray structure analysis of 10 is shown in Figure 2. Compound
11 and a corresponding amount of dibenzyl were detected
when benzyl chloride was added to pre-heated (1408C) reac-
tion mixtures of 10, bromobenzene, styrene, and K2CO3. Im-
portantly, in the reference reactions (solutions of 1 (or 2) in
2,6-{XP(piperidinyl)2}2Pd]Cl
or
[C6H3-2,6-{XP(piperidi-
nyl)2}2Pd(Br)(Ph)]Cl (X=NHor O). This anticipation got
experimental support from the addition of equimolar
amounts (relative to bromobenzene) of pyridine to solutions
of 1 (or 2) and bromobenzene in DMF, which significantly
retarded the halide exchange due to adduct formation. The
oxidative addition of bromobenzene to 1 (and to 2) became
evident from the observation that no halide exchange at all
was observed when 6 or [C6H3-2,6-(OPiPr2)2Pd(Cl)] (14)
were treated with bromobenzene under identical reaction
conditions.[2h] Steric parameters of these complexes are
almost identical and hence, should have no effect. Appa-
rently, the electron density on the metal centers is not rele-
vant either: whereas the electron density of 6 is comparable
to that of 1, the one of 14 is the lowest.[40] Thus, reaction
paths involving a nucleophilic attack of chloride on the Cipso
atom of coordinated bromobenzene can be excluded. The
nitrogen lone pairs of the aminophosphines, which apparent-
ly facilitate the formation of PdIV intermediates, provide the
only explanation for the striking difference in the reactivity
of 1 and 2 when compared with that of the pincer complexes
6 and 14, respectively.[37] To gain more mechanistic insights
into the halide-exchange reaction of 1 (and 2) with bromo-
benzene, DFT calculations were performed using the experi-
mental system with the sterically demanding P(piperidinyl)2
groups. The potential energy surface is simulated in DMF as
solvent and is shown in Figure 3 along with a schematic il-
lustration of the local minima and transition states.[42] Disso-
ciation of the chloride ligand from 1 to form ion pair A (+
Figure 2. ORTEP diagram of a molecule of 10, showing the atom labeling
scheme (50% probability).[33] Hydrogen atoms are omitted for clarity. Se-
À
À
lected bond lengths [Š] and angles [8]: Pd1 P1 2.2590(11), Pd1 P2
À
À
À
À
2.2567(12), P1 N1 1.716(4), P1 N2 1.673(4), P1 N3 1.684(4), P2 N4
À
À
1.659(5), P2 N5 1.691(4), P2 N6 1.722(4); P1-Pd1-P2 174.88(4).
Chem. Eur. J. 2008, 14, 7969 – 7977
ꢀ 2008 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
www.chemeurj.org
7973

C. M. Frech et al.
Figure 3. Calculated reaction pathway of the reaction of 1 with bromobenzene in DMF.
23.8 kcalmolÀ1) is anticipated to initiate the reaction se-
Conclusion
quence.[43] Oxidative addition of bromobenzene to A+ re-
sults in the cationic PdIV intermediate B+ (+28.8 kcal
molÀ1), which is only 5.0 kcalmolÀ1 higher in energy than
A+. The energetic barrier, that is, the energy difference be-
tween A+ and the transition state TS1+, is only +19.9 kcal
molÀ1. Re-coordination of the chloride generates C, which
has a computed energy of +25.6 kcalmolÀ1. The reductive
elimination of chlorobenzene from C to yield complex 12 is
exothermic (the energy difference between 1 and 12 is
À5.3 kcalmolÀ1) and accompanied by a calculated barrier
(TS2) of +27.8 kcalmolÀ1.
In conclusion, the halide exchange demonstrated that
pincer-type PdIV intermediates (derived from 1 and 2) are
principally accessible at elevated reaction temperatures (in
particular when polar solvents, such as DMF or NMP, or
pincer complexes containing labile anionic ligands are used)
in reactions with aryl halides.[44] Although palladium nano-
particles are the active form in the Heck reaction catalyzed
by aminophosphine- (and phosphite-) based pincer complex-
es,[2h] this is not necessarily the case for phosphine-based
pincer complexes, where PdIV species might indeed be the
key intermediates in the catalytic cycle.[2a,45]
The aminophosphine-based pincer complex 1 is nowadays
one of the most efficient and convenient Heck catalysts, pro-
À
moting quantitative C C coupling of deactivated and steri-
cally hindered aryl bromides with various olefins in very
short reaction times and with low catalyst loadings. In-
creased reaction temperatures offer the possibilities to
either quantitatively prepare trisubstituted olefins or to per-
form coupling reactions even with deactivated and sterically
hindered aryl chlorides with various olefins as coupling part-
ner. Mechanistic investigations indicate that homogeneous
Pd0 complexes are not the active form of pincer-type palla-
dium catalysts (a new method to test this possibility was de-
veloped), which is in line with previous investigations. On
the other hand, the involvement of palladium nanoparticles
in the Heck reactions catalyzed by 1 (and 2) received strong
experimental evidence (sigmoidal-shaped kinetics, of which
the induction periods were significantly shortened, for ex-
ample, by the addition of few drops of water). Although our
results indicate that palladium nanoparticles are the active
form of aminophosphine-based pincer complexes, the acces-
sibility of aminophosphine-based PdIV intermediates, howev-
er, was experimentally and computationally shown. This im-
plies that its PdIV intermediates are generally to be consid-
7974
www.chemeurj.org
ꢀ 2008 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 7969 – 7977

Pincer-Type Heck Catalysts
FULL PAPER
ed with pentane (3”10 mL), filtered, and dried again. The product was
ered as reactive intermediates in reactions (such as the
Suzuki–Miyaura cross-coupling) catalyzed by pincer com-
plexes of palladium with aryl halides performed at elevated
temperatures. Detailed computational studies addressing
this issue in the Heck reaction catalyzed by pincer com-
plexes are currently in progress.
obtained in 93% yield (194 mg, 0.288 mmol).
31P{1H} NMR (CD2Cl2, TMS): d=105.8 ppm (s, P(CH
1
A
(CD2Cl2, TMS): d=3.17 (br.s, 24H; NCH2), 1.57 (br.s, 12H;
NCH2CH2CH2), 1.51 ppm (br.s, 24H; NCH2CH2); 13C{1H} NMR
(CD2Cl2): d=48.5 (vt,
JP,C =13.6 Hz; NCH2), 27.2 (vt, JP,C =8.3 Hz;
NCH2CH2), 25.7 (s, NCH2CH2CH2); elemental analysis calcd (%) for
C30H60N6P2Pd: C 53.53, H8.98, N 12.48; found: C 53.26, H8.98, N 12.38.
Preparation of [C6H3-2,6-(OPiPr2)2Pd(CO)]
ACHTRE[UNG BF4]: AgBF4 (13.0 mg,
0.07 mmol) was added to solution of [C6H3-2,6-(OPiPr2)2Pd(Cl)]
a
Experimental Section
(32.2 mg, 0.07 mmol) in dichloromethane (10 mL). The reaction mixture
was stirred for 15 min under rigorous exclusion of light. The reaction
mixture was filtered through celite. The solvent was removed under re-
duced pressure. Treating a solution of [C6H3-2,6-(OPiPr2)2Pd]ACHTREUNG[BF4] in
CD2Cl2 with an excess (ca. 50 equiv) of CO gas yielded its carbonyl
adduct. CO release was observed under reduced pressure, hence an ele-
General procedures: All synthetic operations were carried out in oven-
dried glassware using a combination of glovebox (M. Braun 150B-G-II)
and Schlenk techniques under a dinitrogen atmosphere. Solvents were re-
agent grade or better, freshly distilled under a N2 atmosphere by standard
procedures, and degassed by freeze–thaw cycles before use. Deuterated
solvents were purchased from Armar, stored in a Schlenk tube (Teflon
tap) over CaH2, distilled, and degassed prior to use. All the chemicals
were purchased from Aldrich Chemical Co., Acros Organics, or Fluka,
and used without further purification.
Analysis: 1H, 13C{1H}, and 31P{1H} NMR data were recorded at 500.13,
125.76, and 202.46 MHz, respectively, on a Bruker DRX-500 spectrome-
ter. Chemical shifts (d) are expressed in parts per million (ppm), coupling
constants (J) are in Hz. The 1Hand 13C NMR chemical shifts are report-
ed relative to tetramethylsilane; the resonance of the residual protons of
the solvent was used as an internal standard for 1H (d=5.32 ppm di-
chloromethane) and all-D solvent signals for 13C (d=53.5 ppm dichloro-
methane). 31P{1H} NMR data are reported downfield relative to external
85% H3PO4 in D2O at d=0.0 ppm. All measurements were carried out
at 298 K. Abbreviations used in the description of NMR data are as fol-
lows: s, singlet; d, doublet; t, triplet; dist q, distinct quartet; m, multiplet;
v, virtual. IR spectra were obtained by ATR methods with a Bio-Rad
FTS-45 FTIR spectrometer. Elemental analyses were performed on a
Leco CHNS-932 analysator at the University of Zurich, Switzerland.
mental analysis was not obtained.
1
31P{1H} NMR (CD2Cl2, TMS): d=186.4 ppm (s, P(CH
G
(CD2Cl2, TMS): d=7.26 (t, 3JH,H =8.3 Hz, 1H; ArH), 6.78 (d, JH,H
8.3 Hz, 2H; ArH), 2.61 (m, 4H; CH(CH3)2), 1.31 (dist q, J=7.8 Hz, 12H;
3
AHCTREUNG
(CH3)2); 13C{1H} NMR
CHACTHERUNG(CH3)2), 1.22 ppm (dist q, J=7.8 Hz, 12H; CHACHTREUGN
(CD2Cl2, TMS): d=180.8 (t, 2JP, C =11.1 Hz, CO), 167.4 (vt, JP,C =6.2 Hz,
2
Ar), 134.8 (s, Ar), 133.2 (s, Ar), 107.6 (t, JP, C =7.1 Hz, Ar), 30.6 (vt, JP, C
11.6 Hz, CH
ACHTREU(GN CH3)2), 17.8 (s, CHAHCETRU(NG CH3)2), 17.1 ppm (s, CHACHTRE(UNG CH3)2); IR
(ATR): n˜ =2141 cmÀ1 (s, CO).
Reactions with benzyl chloride: Benzyl chloride (3.0 equiv) was added to
solutions of 8, 9, and 10 (30 mg), respectively, in dichloromethane, and
stirred for about 15 min. Samples taken from the reaction mixtures were
diluted with dichloromethane and analyzed by GC/MS. After evapora-
tion of the solvent the reaction mixtures were washed with small portions
of cold pentane (3”2 mL), which gave the pure dichloropalladium com-
plexes with the general formula [PdACHTREU(NG PR3)2(Cl)2] (PR3 =PPh3,
C6H5CH2PiPr2 or P(piperidinyl)3) as shown by various NMR techniques.
Procedure for the “one-pot” synthesis of catalyst solutions of 1 and 2: In
a Young Schlenk [PdACHTRE(UNG cod)Cl2] (100 mg, 0.35 mmol) was suspended in tol-
Preparation of [C6H3ACHTREUNG{NHP(piperidinyl)2}2PdAHCETRU(NG NC5H5)]ACHTREU[GN BF4] (4): AgBF4
uene (50 mL). After the addition of solutions containing two equivalents
of 1,1’,1’’-phosphinetriyltripiperidine (198.5 mg, 0.70 mmol) in toluene
(20 mL), the reaction mixture was stirred for 10 min. Subsequently, an
equimolar amount of resorcinol or 1,3-diaminobenzene, was added to
these solutions. The reaction mixtures were heated up to 1008C and
stirred until decoloration occurred. After the mixtures had been cooled
to room temperature and the insoluble reaction products had been pre-
cipitated, the concentrations of the catalyst solutions were determined.
Appropriate amounts from these solutions were used for catalysis.
(14.5 mg, 0.07 mmol) was added to a solution of 1 (48.0 mg, 0.07 mmol)
in dichloromethane (10 mL). The reaction mixture was stirred for 15 min
under rigorous exclusion of light and then filtered through celite. Subse-
quent addition of two drops of pyridine yielded complex 4. After removal
of the solvent and residual pyridine under reduced pressure, the solid
was washed with diethyl ether (2”5 mL). The residue was dried in vacuo
to 4 as a white solid (53.7 mg, 0.07 mmol; 93%).
31P{1H} NMR (CD2Cl2, TMS): d=117.5 (s, P{CH(CH3)2}2); 1HNMR
ACHTREUNG
(CD2Cl2, TMS): d=8.56 (dt, 4JH,H =1.8 Hz, 3JH,H =1.5 Hz, 2H; Ar’ortho),
8.04 (tt, 3JH,H =6.3 Hz, 4JH,H =1.8 Hz, 1H, Ar’para), 7.64 (dt, 3JH,H =6.3 Hz,
3JH,H =1.5 Hz, 2H; Ar’meta), 6.87 (tt, 3JH,H =7.8 Hz, 4JP, H =2.1 Hz, 1H;
Arpara), 6.21 (d, 3JH,H =7.8 Hz, 2H; Armeta), 4.93 (s, 2H; NH), 3.01 (broad
s, 16H; NCH2), 1.56 (m, 8H; NCH2CH2CH2), 1.47 ppm (m, 16H;
NCH2CH2): 13C{1H} NMR (CD2Cl2, TMS); 152.3 (s, Ar’ortho), 151.1 (vt,
General procedure for Heck reactions: In a Young Schlenk (10 mL) were
placed appropriate amounts of the olefin, aryl halide, K2CO3, tetrabuty-
lammonium bromide (in reactions performed with aryl chlorides), and
the solvent. Then the correct amount of catalyst was added by syringe as
a solution in toluene. The mixture was vigorously stirred and heated up
to the reaction temperature. Samples, taken from the reaction mixture,
were diluted with dichloromethane and analyzed by GC/MS. At the end
of catalysis the reaction mixtures were allowed to cool to room tempera-
ture, quenched with aqueous HCl (2m, 40 mL), extracted with dichloro-
methane (3”40 mL), and the combined extracts were dried (MgSO4) and
evaporated to dryness. The crude material was purified by flash chroma-
tography on silica gel.
J
P, C =63.5 Hz, ArNHP), 146.1 (s, Ar’para), 129.9 (s, Arpara), 127.0 (s, Ar’meta),
120.0 (unresolved t, Aripso), 103.5 (vt, JP, C =39.0 Hz, Armeta), 47.4 (vt, JP, C
13.2 Hz, PNCH2), 26.9 (vt, P, C =11.9 Hz, PNCH2CH2), 25.2 ppm (s,
=
J
PNCH2CH2CH2); elemental analysis calcd (%) for C31H50BF4N7P2Pd: C
47.99, H6.49, N 12.64; found: C 48.31, H6.65, N 12.42.
Preparation of [Pd(S2CNpiperidyl)2] (5): An excess (ca. 20 equiv) of CS2
was added to a solution of 1 (28.0 mg, 0.04 mmol) in dichloromethane
(2 mL). Degradation of complex 1 was noticed within a few minutes, as
shown by 31P{1H} NMR spectroscopy. Large crystals of
5 (17.9 mg,
0.04 mmol; 97%) were grown within two weeks from the reaction mix-
ture. The crystals were isolated, washed with small portions of cold pen-
tane, dried, and analyzed by various NMR techniques and X-ray diffrac-
tion.[21]
Acknowledgement
This work was supported by the University of Zurich and the Swiss Na-
tional Science Foundation (SNSF).
Preparation
of
[Pd{P(piperidinyl)3}2]
(10):
A
solution
of
[Pd(Cl)2{P(piperidinyl)3}2] (230 mg, 0.310 mmol) in THF (20 mL) and a
large excess of sodium metal (ca. 100 equiv) was stirred at room tempera-
ture overnight. After filtration of the reaction mixture through celite, the
solvent was removed under reduced pressure. Compound 10 was extract-
[1] For reviews of palladium-catalyzed Heck reaction, see: a) R. F.
Heck in Palladium Reagents in Organic Syntheses (Eds.: A. R. Ka-
tritzky, O. Meth-Cohn, C. W. Rees), Academic Press, London, 1985,
Chem. Eur. J. 2008, 14, 7969 – 7977
ꢀ 2008 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
www.chemeurj.org
7975

C. M. Frech et al.
À
p. 2; b) R. F. Heck, “Vinyl Substitution with Organopalladium Inter-
mediates” in Comprehensive Organic Synthesis, Vol. 4 (Eds.: B. M.
Trost, I. Fleming), Pergamon, Oxford, 1991; c) J.-L. Malleron, J.-C.
Fiaud, J.-Y. Legros in Handbook of Palladium-Catalysed Organic
Reactions, Academic Press, London, 1997; d) M. T. Reetz in Transi-
tion Metal Catalysed Reactions, (Eds.: S. G. Davies, S.-I. Murahashi),
Blackwell, Oxford, 1999; e) J. T. Link, L. E. Overman in Metal-Cata-
lyzed Cross-Coupling Reactions (Eds.: F. Diederich, P. J. Stang),
Wiley-VCH, Weinheim, 1998, Chapter 6; f) S. Bräse, A. de Meijere
in Metal-Catalyzed Cross-Coupling Reactions (Eds.: F. Diederich,
P. J. Stang), Wiley-VCH, Weinheim, 1998, Chapter 3.6; g) K. C. Nic-
olaou, E. J. Sorensen in Classics in Total Synthesis, Wiley-VCH,
Weinheim, 1996, Chapter 31; h) R. A. de Vries, P. C. Vosejpka,
M. L. Ash in Catalysis of Organic Reactions (Ed.: F. E. Herkes), M.
Dekker, New York, 1998, Chapter 37; i) L. F. Tietze, G. Kettschau,
U. Heuschert, G. Nordmann, Chem. Eur. J. 2001, 7, 368.
(probably by P O cleavage), this is not necessarily the case for a
phosphine-based PCP system.
[11] The resulting catalyst solutions were used for catalytic reactions
without purification and remain stable in solution for several
months at room temperature, affording the coupling products at es-
sentially the same conversion rates and yields as freshly prepared
catalyst solutions from pure 1 and 2, respectively. The concentra-
tions of the catalyst solutions were determined prior to use.
[12] Since the steric bulk of 1 and 2 is almost identical, the higher cata-
lytic activity of 1 can be attributed to electronic factors (the metal
center of 1 being more electron-rich than 2)[7] if PdIV intermediates
are involved in the catalytic cycle. The higher stability of 2 com-
pared with 1 could provide an explanation for its lower catalytic ac-
tivity if palladium nanoparticles are the active form.
[13] See Experimental Section.
[14] The higher reaction temperature could provide an explanation for
the leveled out catalytic activity of catalysts 1 and 2, if palladium
nanoparticles are the catalytically active form of the catalysts.
[15] H. M. Lee, J. Y. Zeng, C.-H. Hu, M.-T. Lee, Inorg. Chem. 2004, 43,
6822.
[16] A. H. M. de Vries, J. M. C. A. Mulders, J. H. M. Mommers, H. J. W.
Henderickx, J. G. de Vries, Org. Lett. 2003, 5, 3285.
[17] S. Gründemann, M. Albrecht, J. A. Loch, J. W. Faller, R. H. Crab-
tree, Organometallics 2001, 20, 5485.
[18] a) W. A. Herrmann, M. Elison, J. Fischer, C. Kçcher, G. R. J. Artus,
Angew. Chem. 1995, 107, 2602; Angew. Chem. Int. Ed. Engl. 1995,
34, 2371; b) G. D. Frey, C. Reisinger, E. Herdtweck, W. A. Herr-
mann, J. Organomet. Chem. 2005, 690, 3193; c) W. A. Herrmann, K.
Öfele, S. K. Schneider, E. Herdtweck, S. D. Hoffmann, Angew.
Chem. 2006, 118, 3943; Angew. Chem. Int. Ed. 2006, 45, 3859.
[19] A. Schnyder, T. Aemmer, A. F. Indolese, U. Pittelkow, M. Studer,
Adv. Synth. Catal. 2002, 344, 495.
[20] The low conversion rates observed in these solvents can be ex-
plained by both, ion pair and palladium nanoparticle formation.
Whilst their formation is not promoted in p-xylene, the ligating
properties of DMSO could either suppress the coordination of sub-
strate molecules at the metal centers or stabilize the catalysts,
which, in turn, would retard the formation of palladium nanoparti-
cles.
[21] a) J. A. Widegren, R. G. Finke, J. Mol. Catal. A 2003, 198, 317; e) C.
Rocaboy, J. A. Gladysz, Org. Lett. 2002, 4, 1993; b) G. M. White-
sides, M. Hackett, R. L. Brainard, J.-P. P. M. Lavalleye, A. F. Sowin-
ski, A. N. Izumi, S. S. Moore, D. W. Brown, E. M. Staudt, Organome-
tallics 1985, 4, 1819; c) P. Foley, R. Di Cosimo, G. M. Whitesides, J.
Am. Chem. Soc. 1980, 102, 6713; d) D. R. Anton, R. H. Crabtree,
Organometallics 1983, 2, 855.
[22] It is important to note that the mercury drop test is not applicable
here, since halide-exchange reactions of 1 (and 2) with bromoben-
zene at 1008C in DMF to form chlorobenzene and their bromo de-
rivatives 11 (and 12) also were inhibited in the presence of mercury
(for more details regarding the halide exchange, see below).
[23] The reason for the effect of water on the induction period could be
explained by decomposition of the pincer core to form the active
catalyst (i.e. palladium nanoparticles). However, it should be men-
tioned that Shaw et al. envisioned a possible pathway in which the
olefin first coordinates to the Pd center and a nucleophile then at-
tacks the olefin, forming a palladium alkyl complex, which could fa-
cilitate the oxidative addition of the aryl halide.[9b] Water could
serve as a nucleophile, thereby explaining the shortened induction
period in the presence of water. To test this possibility tetrabutylam-
monium bromide (ca. 15 mol%) was added to the reaction mixtures.
If the mechanism would proceed according to Shawꢁs mechanism,
the added bromide should greatly reduce the induction period. The
induction period remained unchanged in the presence of added bro-
mide, which is inconsistent with Shawꢁs mechanism.
[2] a) M. Ohff, A. Ohff, M. E. van der Boom, D. Milstein, J. Am. Chem.
Soc. 1997, 119, 11687; b) K. Kiewel, Y. Liu, D. E. Bergbreiter, G. A.
Sulikowski, Tetrahedron Lett. 1999, 40, 8945; c) F. Miyazaki, K. Ya-
maguchi, M. Shibasaki, Tetrahedron Lett. 1999, 40, 7379; d) D. Mo-
rales-Morales, R. Redon, C. Yung, C. M. Jensen, Chem. Commun.
2000, 1619; e) D. Morales-Morales, C. Grause, K. Kasaoka, R.
Redon, R. Cramer, C. M. Jensen, Inorg. Chim. Acta 2000, 300–302,
958; f) S. Sjovall, O. P. Wendt, C. J. Anderson, Chem. Soc. Dalton
Trans. 2002, 1396; g) D. Morales-Morales, R. E. Cramer, C. M.
Jensen, J. Organomet. Chem. 2002, 654, 44; h) M. R. Eberhard, Org.
Lett., 2004, 2125; i) N. Withcombe, K. K. Hii (Mimi) S. Gibson, Tet-
rahedron 2001, 57, 7449.
[3] a) W. A. Herrmann, C. Brossmer, K. Öfele, C. Reisinger, T. Rierme-
ier, M. Beller, H. Fisher, Angew. Chem. 1995, 107, 1989; Angew.
Chem. Int. Ed. Engl. 1995, 34, 1844; b) A. Littke, G. C. Fu, J. Org.
Chem. 1999, 64, 10; c) X. Gai, R. Grigg, I. Ramzan, V. Sridharan, S.
Collard, J. Muir, Chem. Commun. 2000, 2053; d) S. Gibson, D.
Foster, D. Eastham, R. Tooze, D. Cole-Hamilton, Chem. Commun.
2001, 779; e) M. Albrecht, G. van Koten, Angew. Chem. 2001, 113,
3866; Angew. Chem. Int. Ed. 2001, 40, 3750; f) M. E. van der Boom,
D. Milstein, Chem. Rev. 2003, 103, 1759.
[4] A. F. Littke, G. C. Fu, J. Am. Chem. Soc. 2001, 123, 6989.
[5] J. P. Stambuli, S. R. Stauffer, K. H. Shaughnessy, J. F. Hartwig, J.
Am. Chem. Soc. 2001, 123, 2677.
[6] a) E. Peris, J. A. Loch, J. Mata, R. H. Crabtree, Chem. Commun.
2001, 201; b) S. Gründemann, M. Albrecht, J. A. Loch, J. W. Faller,
R. H. Crabtree, Organometallics 2001, 20, 5485; c) J. A. Loch, M.
Albrecht, E. Peris, J. Mata, J. W. Faller, R. H. Crabtree, Organome-
tallics 2002, 21, 700; d) D. J. Nielsen, K. J. Cavell, B. W. Skelton,
A. H. White, Inorg. Chim. Acta 2002, 327, 116; e) W. A. Herrmann,
V. P. W. Bçhm, C. W. K. Gstçttmayr, M. Grosche, C.-P. Reisinger, T.
Weskamp, J. Organomet. Chem. 2001, 617, 616; f) C. Yang, H. M.
Lee, S. P. Nolan, Org. Lett. 2001, 3, 1511; g) N. Tsoureas, A. A. Da-
nopoulos, A. A. D. Tulloch, M. E. Light, Organometallics 2003, 22,
4750.
[7] J. L. Bolliger, O. Blacque, C. M. Frech, Angew. Chem. 2007, 119,
6634; Angew. Chem. Int. Ed. 2007, 46, 6514.
[8] Reviews that include catalysis of the Heck reaction by palladacycles:
a) see refs [2a,b]; b) P. L. Beletskaya, A. V. Cheprakov, Chem. Rev.
2000, 100, 3009; c) M. Dupont, M. Pfeffer, J. Spencer, Eur. J. Inorg.
Chem. 2001, 1917; d) R. B. Bedford, Chem. Commun. 2003, 1787;
e) G. J. de Vries, Dalton Trans. 2006, 421; f) D. E. Bergbreiter, P. L.
Osburn, J. D. Frels, Adv. Synth. Catal. 2005, 347, 172; g) K. Yu, W.
Sommer, J. M. Richardson, M. Weck, C. W. Jones, Adv. Synth. Catal.
2005, 347, 161; h) K. Yu, W. Sommer, M. Weck, C. W. Jones, J.
Catal. 2005, 226, 101.
[9] a) W. A. Herrmann, V. P. W. Bçhm, C.-P. Reisinger, J. Organomet.
Chem. 1999, 576, 23; b) B. L. Shaw, S. D. Perera, E. A. Staley, Chem.
Commun. 1998, 1362.
[10] Palladium nanoparticles are proposed to be involved when phos-
phinite PCP-type Heck catalysts are utilized (see ref. [2f, h]). While
PCP phosphinite systems might decompose at elevated temperatures
and prolonged reaction times under the basic Heck conditions
[24] Tetrabutylammonium bromide is known to stabilize palladium nano-
particles. See for instance: a) T. Jeffery in Advances in Metal-Organ-
ic Chemistry, Vol. 5 (Ed.: L. S. Liebeskind), JAI Press, Greenwich,
1996, p. 153; b) M. T. Reetz, E. Westermann, Angew. Chem. 2000,
7976
www.chemeurj.org
ꢀ 2008 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 7969 – 7977

Pincer-Type Heck Catalysts
FULL PAPER
112, 170; Angew. Chem. Int. Ed. 2000, 39, 165; c) A. Zapf, M.
Beller, Chem. Eur. J. 2001, 7, 2908.
mobenzene when compared with that of 6 and 14, they cannot be
excluded completely: Decomposition of a minor amount of 1 (or 2)
to form the anionic complexes [ArPdBr2]À or its dimer via Pd0 inter-
mediates may occur to promote the halide exchange with the pincer
complexes. Alternatively, a bromide ligand could dissociate from
[ArPdBr2]À or its dimer (possibly as a dimethylammonium salt from
DMF) and exchange with the pincer complexes directly.
[25] The presence of tetrabutylammonium bromide is anticipated to
retard the conversion rates if PdIV intermediates are involved in the
catalytic cycle; the bromide anions suppress the coordination of the
substrate molecules at the metal center. On the other hand, the low
concentration of palladium nanoparticles (if formed) may not need
to be stabilized. In contrast, tetrabutylammonium bromide may slow
down their formation and/or reduces their activity.
[26] a) J. A. Widegren, R. G. Finke, J. Mol. Catal. A 2003, 198, 317; b) Y.
Lin, R. G. Finke, Inorg. Chem. 1994, 33, 4891; c) K. S. Weddle, J. D.
Aiken III, R. G. Finke, J. Am. Chem. Soc. 1998, 120, 5653.
[27] Quantitative poisoning experiments are usually performed at tem-
peratures below 508C, whereas our experiments were conducted at
1408C. High temperatures may lead to dissociation of thiophene or
PPh3 from heterogeneous systems, thereby regenerating the catalyst.
However, examples of Heck reactions performed at 1808C with re-
sorcinol-derived pincer complexes of palladium were effectively
stopped after addition of the corresponding amounts of CS2, thio-
phene, or PPh3.[2g]
[38] a) M. L. Clarke, D. J. Cole-Hamilton, M. Z. Slawin, J. D. Woollins,
Chem. Commun. 2000, 2065; b) K. G. Moloy, J. L. Peterson, J. Am.
Chem. Soc. 1995, 117, 7696.
[39] For example, only 10% conversion was observed in p-xylene at
1008C over night.
[40] The CO stretching frequency of the cationic carbonyl derivatives
[C6H3-2,6-(CH2PCy2)2Pd(CO)]
(OPiPr2)2Pd(CO)][BF4]
[C6H3{XP(piperidinyl)2}2Pd(CO)]AHCTREUNG ;
[BF4] (X=NH, nCO =2106 cmÀ1
[OTf] (nCO =2105 cmÀ1),[41] [C6H3-2,6-
ACHTREUNG
A
(nCO =2141 cmÀ1
)
and
X=O, nCO =2133 cmÀ1) are indicative of the electron density at the
metal centers.[7,13]
[41] Y. Jong Gook, S. Jung Min, L. Kap Duk, K. Sangha, P. Soonheum,
Bull. Korean Chem. Soc. 1996, 17, 311.
[28] a) F. Shaheen, A. Badshah, M. Gielen, M. Dusek, K. Fejfarova, D.
de Vos, B. Mirza, J. Organomet. Chem. 2007, 692, 3019; b) F. Sha-
heen, A. Badshah, S. Anjum, A. Saqib, Acta Crystallogr. Sect. E
2006, 62, m329; c) L. Marcheselli, C. Preti, M. Tagliazucchi, V. Cher-
chi, L. Sindellari, A. Furlani, A. Papaioannou, V. Scarcia, Eur. J.
Med. Chem. 1993, 28, 347; d) C. G. Sceney, R. Magee, J. Inorg. Nucl.
Chem. 1974, 10, 323.
[29] Increasing TOF with increasing substrate/catalyst ratios is believed
to be caused by the palladium which is present in the form of nano-
particles.[16a,30] At higher substrate/catalyst ratios more palladium
will be in the catalytic cycle. In addition, the size of the nanoparti-
cles will be smaller at higher substrate/catalyst ratios, making the
ratio between palladium on the outer rim and palladium on the
inside of the particles more favorable.
[30] M. T. Reetz, J. G. de Vries, Chem. Commun. 2004, 1559.
[31] C. M. Frech, L. J. W. Shimon, D. Milstein, Angew. Chem. 2005, 117,
1737; Angew. Chem. Int. Ed. 2005, 44, 1709.
[32] C. M. Frech, G. Leitus, D. Milstein, Organometallics 2008, 27, 894.
[33] CCDC-671437 contains the supplementary crystallographic data for
this paper. These data can be obtained free of charge from The
Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/
data_request/cif
[42] Full geometry optimizations were performed with the Gaussian03
program package using the mPW1PW91 functional. Computational
details and a complete scheme with all calculated intermediates and
transition states, simulated in different solvents is given in the Sup-
porting Information.
[43] Oxidative addition of bromobenzene on 1 is not possible, since the
calculated energy barrier is 71.8 kcalmolÀ1 [42]
.
[44] According to our findings it seems indeed to be the case that PdIV
intermediates are the key intermediates in the Suzuki–Miyaura
cross-coupling reaction mediated by 1 (and 2), as it was recently
suggested.[7]
[45] This anticipation got experimental support from Heck reactions per-
formed with bromobenzene and n-butyl acrylate in NMP at 1408C
with 0.01 mol% of catalyst and K2CO3 as base. Whereas dramatical-
ly retarded conversion rates were observed with the phosphine-
based pincer complex 6 in the presence of 1-methyl-1,4-cyclohexa-
diene,[2b] only a marginal effect was noticed under identical reaction
conditions with catalyst 1 (as well as with 14). These results rule out
that the catalytically active species derived from 1 (or 14) and from
6 are of the same type, and in turn imply that different reaction
mechanisms are operative in Heck reactions catalyzed by amino-
phosphine- (or phosphite-) and phosphine-based pincer complexes.
The addition of 1-methyl-1,4-cyclohexadiene also had a negligible
[34] Dibenzyl formation was not observed when [PdCAHTRE(UGN OAc)2]—a system
known to form palladium nanoparticles—was used as catalyst.
[35] Various catalyst concentrations were tested: 0.05, 0.02, 0.002, and
0.0005 mol% of 1 and 2, respectively.
[36] A 60% conversion of 1 into 11 and a 32% conversion of 2 into 12
was observed after 5 h.
[37] The aminophosphines provide an explanation for the striking differ-
ence in the reactivity of 1 and 2 in the halide exchange when com-
pared with that of 6 and 14, if PdIV intermediates are traversed.
Apart from the high s-donor strength of the P atom of aminophos-
phines, which is similar to those of its phosphine analogues, the ni-
trogen lone pairs may donate additional electron density towards
the phosphorus atom and hence, towards the metal center.[38] Al-
though the following mechanisms would not provide an explanation
for the striking difference in the reactivity of 1 (and 2) towards bro-
influence on the conversion rate of [PdACHTRE(UNG OAc)2]-catalyzed Heck reac-
tions, adding further support to the fact that palladium nanoparticles
probably are not the active form of phosphine-based pincer com-
plexes. This, however, would indicate that pincer-type PdIV inter-
mediates are involved in the catalytic cycle of phosphine-based
pincer complexes, since the formation of homogeneous Pd0 com-
plexes was already excluded. Nevertheless, it is important to note
that when styrene was used as coupling partner, a dramatic retarda-
tion in the conversion rate was observed in all cases when 1-methyl-
1,4-cyclohexadiene was present, implying that this test is not defini-
tive.
Received: March 11, 2008
Published online: July 11, 2008
Chem. Eur. J. 2008, 14, 7969 – 7977
ꢀ 2008 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
www.chemeurj.org
7977