Cyanation of aryl bromides with K4[Fe(CN)6] catalyzed by dichloro[bis{1-(dicyclohexylphosphanyl)piperidine}]palladium, a molecular source of nanoparticles, and the reactions involved in the catalyst-deactivation processes

Article abstract of DOI:10.1002/chem.201102936

Dichloro[bis{1-(dicyclohexylphosphanyl)piperidine}]palladium [(P{(NC ₅H₁₀)(C₆H₁₁)₂}) ₂PdCl₂] (1) is a highly active and generally applicable C-C cross-coupling catalyst. Apart from its high catalytic activity in Suzuki, Heck, and Negishi reactions, compound 1 also efficiently converted various electronically activated, nonactivated, and deactivated aryl bromides, which may contain fluoride atoms, trifluoromethane groups, nitriles, acetals, ketones, aldehydes, ethers, esters, amides, as well as heterocyclic aryl bromides, such as pyridines and their derivatives, or thiophenes into their respective aromatic nitriles with K₄[Fe(CN)₆] as a cyanating agent within 24 h in NMP at 140 °C in the presence of only 0.05 mol % catalyst. Catalyst-deactivation processes showed that excess cyanide efficiently affected the molecular mechanisms as well as inhibited the catalysis when nanoparticles were involved, owing to the formation of inactive cyanide complexes, such as [Pd(CN)₄]^2-, [(CN)₃Pd(H)]^2-, and [(CN)₃Pd(Ar)]^2-. Thus, the choice of cyanating agent is crucial for the success of the reaction because there is a sharp balance between the rate of cyanide production, efficient product formation, and catalyst poisoning. For example, whereas no product formation was obtained when cyanation reactions were examined with Zn(CN)₂ as the cyanating agent, aromatic nitriles were smoothly formed when hexacyanoferrate(II) was used instead. The reason for this striking difference in reactivity was due to the higher stability of hexacyanoferrate(II), which led to a lower rate of cyanide production, and hence, prevented catalyst-deactivation processes. This pathway was confirmed by the colorimetric detection of cyanides: whereas the conversion of β-solvato-α-cyanocobyrinic acid heptamethyl ester into dicyanocobyrinic acid heptamethyl ester indicated that the cyanide production of Zn(CN)₂ proceeded at 25 °C in NMP, reaction temperatures of >100 °C were required for cyanide production with K₄[Fe(CN) ₆]. Mechanistic investigations demonstrate that palladium nanoparticles were the catalytically active form of compound 1. A balancing act: Compound 1 (see scheme) is a highly active cyanation catalyst. Furthermore, a sharp balance between the rates of cyanide generation, efficient product formation, and catalyst deactivation owing to excess cyanide was observed in deactivation processes. Copyright

Full text of DOI:10.1002/chem.201102936

DOI: 10.1002/chem.201102936
Cyanation of Aryl Bromides with K4[Fe(CN)6] Catalyzed by
DichloroACHTUNGTRENNUNG[bis{1-(dicyclohexylphosphanyl)piperidine}]palladium, a Molecular
Source of Nanoparticles, and the Reactions Involved in the
Catalyst-Deactivation Processes
Roman Gerber, Miriam Oberholzer, and Christian M. Frech*[a]
Abstract: Dichloro
phosphanyl)piperidine}]palladium
[(P{(NC5H10)(C6H11)2})2PdCl2] (1) is a
[bis{1-(dicyclohexyl-
that excess cyanide efficiently affected
the molecular mechanisms as well as
inhibited the catalysis when nanoparti-
cles were involved, owing to the forma-
tion of inactive cyanide complexes,
such as [Pd(CN)4]2À, [(CN)3Pd(H)]2À,
and [(CN)3Pd(Ar)]2À. Thus, the choice
of cyanating agent is crucial for the
success of the reaction because there is
a sharp balance between the rate of cy-
anide production, efficient product for-
mation, and catalyst poisoning. For ex-
ample, whereas no product formation
was obtained when cyanation reactions
were examined with Zn(CN)2 as the
cyanating agent, aromatic nitriles were
smoothly formed when hexacyanofer-
rate(II) was used instead. The reason
for this striking difference in reactivity
was due to the higher stability of hexa-
cyanoferrate(II), which led to a lower
rate of cyanide production, and hence,
prevented catalyst-deactivation pro-
cesses. This pathway was confirmed by
the colorimetric detection of cyanides:
whereas the conversion of b-solvato-a-
cyanocobyrinic acid heptamethyl ester
into dicyanocobyrinic acid heptamethyl
ester indicated that the cyanide pro-
duction of Zn(CN)2 proceeded at 258C
in NMP, reaction temperatures of
>1008C were required for cyanide
production with K4[Fe(CN)6]. Mecha-
nistic investigations demonstrate that
palladium nanoparticles were the cata-
lytically active form of compound 1.
G
ACHTUNGTRENNUNG
highly active and generally applicable
À
C C cross-coupling catalyst. Apart
from its high catalytic activity in
Suzuki, Heck, and Negishi reactions,
compound 1 also efficiently converted
various electronically activated, non-
ACHTUNGTRENNUNGactivated, and deactivated aryl bro-
mides, which may contain fluoride
atoms, trifluoromethane groups, ni-
triles, acetals, ketones, aldehydes,
ethers, esters, amides, as well as hetero-
cyclic aryl bromides, such as pyridines
and their derivatives, or thiophenes
into their respective aromatic nitriles
with K4[Fe(CN)6] as a cyanating agent
within 24 h in NMP at 1408C in the
presence of only 0.05 mol% catalyst.
Catalyst-deactivation processes showed
Keywords: aminophosphines · cya-
nides · deactivation · nanoparticles ·
palladium
Introduction
meyer reaction, and the ammoxidation of toluene (and its
derivatives).[13,14] However, these methods suffer from major
drawbacks, such as the generation of equimolar amounts of
heavy-metal waste and the need for very high reaction tem-
peratures. Therefore, the transition-metal-catalyzed cyana-
tion of aryl halides (Scheme 1) offers an efficient and power-
ful alternative for the high-yielding synthesis of aromatic ni-
triles under relatively mild reaction conditions, and thus has
applications in process and medicinal chemistry as well as in
ligand synthesis.[15]
Aromatic nitriles are of great importance in organic and me-
dicinal chemistry as they are integral moieties in various
natural products, pharmacologically and biologically active
compounds, herbicides, agrochemicals, and dyes.[1,2] More-
over, they also have applications as synthetic intermediates
that can be converted into various other functional groups,
such as aldehydes, amines, nitrogen-containing heterocycles,
benzoic-acid derivatives, etc.[3–11] Thus, benzonitriles are
highly valuable synthetic targets. Classical synthetic methods
for their preparation include the Rosenmund von Braun re-
action,[12] the diazotization of anilines followed by the Sand-
Scheme 1. Palladium-catalyzed cyanation of aryl bromides.
[a] R. Gerber, M. Oberholzer, Dr. C. M. Frech
Department of Inorganic Chemistry
University of Zꢀrich
Besides nickel- and copper-catalyzed cyanation reac-
tions,[16,17] palladium, for which molecular mechanisms (Pd0/
PdII) as well as catalytic cycles that involve nanoparticles
have been proposed, is still the metal of choice for this
transformation.[18–21] Even though recent developments have
8057 Zꢀrich (Switzerland)
Fax : (+41)44-635-46-92
E-mail: chfrech@aci.uzh.ch
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201102936.
2978
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 2978 – 2986

FULL PAPER
considerably increased the activity of cyanation catalysts to
the point at which some tolerate the reaction of sterically
hindered substrates or nonactivated aryl chlorides,[19e–i] even
at temperatures below 1008C,[19j,k] a typical reaction proce-
dure still requires prolonged reaction times (up to 18 h),
high reaction temperatures (140 or 1608C), and catalyst
loadings of up to 10 mol% for efficient product formation.
Moreover, modified reaction conditions (catalyst loadings
and reaction temperatures) are often required for different
substrates. In addition, most catalysts suffer from low func-
tional-group tolerance and/or from an unpredictable re-
Whereas palladium nanoparticles were demonstrated to
be the catalytically active form of compound 1 in Suzuki
and Heck reactions,[24a,b] a homogeneous Pd0/PdII mechanism
was found to be operative in the Negishi reaction.[24c] How-
ever, mechanistic investigations demonstrated that palladi-
um nanoparticles were the catalytically active form of com-
pound 1 in the cyanation of aryl bromides.
Reaction conditions: To establish the most-efficient reaction
conditions for the cyanation of aryl bromides with com-
pound 1, the reaction temperature, solvent, and base were
varied, using 1-bromo-4-methoxybenzene as the aryl bro-
mide (Table 1). The cyanating agent, potassium hexacyano-
ferrate(II) was not varied in these optimization experiments
because all other cyanide sources tested, such as NaCN,
KCN, and Zn(CN)2 were found to be inferior.
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
widen the scope of this reaction is of great importance for
both academic and industrial applications.
Herein, we report the catalytic performance of dichloro-
ACHTUNGTRENNUNG
Table 1. Initial optimization experiments for the cyanation of 1-bromo-4-
G
ACHTUNGTRENNUNG
methoxybenzene with K4[Fe(CN)6], catalyzed by
AHCTUNGTRENNUNG
1 under N2-atmo-
sphere.[a]
À
C C cross-coupling catalyst, in the cyanation of (heterocy-
clic) aryl bromides with K4[Fe(CN)6], which, apart from its
environmental advantages, is cheap and—in contrast to all
other sources of cyanide—nontoxic.[22] Furthermore, investi-
gations into catalyst-deactivation processes have shown that
excess cyanide affects both the molecular mechanisms at
every step of the cycle as well as catalytic cycles where
nanoparticles are involved. Moreover, colorimetric cyanide
detection has shown that cyanide production with hexacya-
noferrate(II) requires reaction temperatures of >1008C (in
contrast to, for example, Zn(CN)2, where cyanides are al-
ready produced at 258C) and thus, provides a simple explan-
ation for the high catalytic activity of compound 1 in the cy-
anation of aryl bromides with K4[Fe(CN)6] as the cyanation
agent, whilst almost no product formation was observed
when Zn(CN)2 was applied.
Entry Base
Solvent
Cat.
Equiv.
Conv.
AHCTUNGTRENNUNG
[mol%] K2[Fe(CN)6] [%][b]
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Na2CO3
NMP
NMP
NMP
NMP
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.167
0.333
0.5
97
47
30
1
0
3
0
0
0
0
89
76
5
3
0
K2CO3
NaOAc
K3PO4
K3PO4·xH2O NMP
NaOH
tBuOK
tBuONa
NEt3
lutidine
Na2CO3
Na2CO3
Na2CO3
Na2CO3
Na2CO3
Na2CO3
Na2CO3
Na2CO3
Na2CO3
Na2CO3
Na2CO3
Na2CO3
Na2CO3
Na2CO3
Na2CO3
NMP
NMP
NMP
NMP
NMP
DMAc
DMF
1-hexanol
ethylene glycol 0.05
DMSO
p-xylene
NMP (2 mL)
NMP (8 mL)
NMP
NMP
NMP
NMP
NMP
NMP
NMP
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.5
0
100
24[c]
78
100
99
56
27
93
3
Results and Discussion
Catalyst synthesis: [(P
quantitatively formed within a few minutes by the treatment
of a suspension of commercially available [PdCl2A(cod)]
(cod=cycloocta-1,5-diene) in toluene with two equivalents
of 1-(dicyclohexylphosphanyl)piperidine at 258C under a N2
atmosphere (Scheme 2).[23] Compound 1 was assumed to be
the ideal cyanation catalyst because it has been demonstrat-
ed to efficiently promote both molecular (Pd0/PdII) mecha-
nisms and catalytic cycles involving nanoparticles.[24]
ACHUTNGTNERGUN{(NC5H10)AHCTUNTGNERN(UNG C6H11)2})2PdCl2] (1) was
1.0
0.25
0.25
0.25
0.2
0.01
CTHUNGTRENNUNG
[a] Reaction
conditions:
1-bromo-4-methoxybenzene
(1.0 mmol),
K4[Fe(CN)6], base (1.0 mmol), solvent (4.0 mL total, unless otherwise
stated), 1408C, 6 h. [b] Determined by GCMS, based on the aryl halide.
[c] Conversion obtained after 18 h.
Interestingly, whilst comparable conversion rates and
yields were obtained for compound 1 at 140 and 1208C,[25]
a
dramatic drop in activity was noticed when the reaction
temperature was 1008C; no catalytic activity was obtained
when the reaction temperature was lowered further. Dra-
matic effects were also noticed when changing the solvent
or base: For example, whereas 1-bromo-4-methoxybenzene
was almost quantitatively converted into 4-methoxybenzoni-
Scheme 2. Synthesis of dichloroACHTNUTRGNEUNG[bis{1-(dicyclohexylphosphanyl)piperidi-
ne}]palladium (1).
Chem. Eur. J. 2012, 18, 2978 – 2986
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2979

C. M. Frech et al.
trile within 6 h in N-methylpyrrolidone (NMP, 4 mL total) at
1408C with Na2CO3 as the base (0.05 mol% catalyst and
0.25 equiv K4[Fe(CN)6] relative to the aryl bromide), only
47% and 30% of the benzonitrile was obtained when
K2CO3 and NaOAc were used as the base, respectively
(Table 1, entries 1–3). Only trace amounts (<3%) of
4-methoxybenzonitrile were detected when NaOH,
tBuONa, tBuOK, K3PO4 or its hydrate were employed in-
stead (Table 1, entries 4–8). By analogy, the use of amines,
such as NEt3 or lutidine, were also not appropriate because
they efficiently inhibited the catalysis (Table 1, entries 9 and
10).
Similar effects were noticed by changing the solvent:
whilst comparable performances were obtained for NMP
and N,N-dimethylacetamide (DMAc), replacing these sol-
vents with DMF led to a drop in activity (Table 1, entries 11
and 12). Notably, product formation of up to 5% was found
when protic solvents, such as 1-hexanol and ethylene glycol,
were used (Table 1, entries 13 and 14).[26–28] However, no
product formation was observed at all when the cyanation
reactions were performed in DMSO or p-xylene (Table 1,
entries 15 and 16). Moreover, the amount of solvent also
strongly influenced the catalytic performance of compound
1, according to the expected trend: the smaller the solvent/
substrate ratio, the higher the conversion rate (Table 1, en-
tries 1, 17, and 18; Figure 1). The amount of cyanating agent
also had an influence on the rate of conversion and on prod-
uct yields: whereas a successive increase in activity was ob-
served when the amount of K4[Fe(CN)6] was raised from
0.167 to 0.25 equivalents (relative to the aryl bromide), es-
sentially the same level of activity was found within the
range 0.25–0.5 equivalents. However, when the amount of
cyanating agent was increased further, a successive drop in
activity was noticed (Table 1, entries 19–22). For example,
whereas almost-quantitative product formation was
achieved within 6 h when 1-bromo-4-methoxybenzene was
coupled with 0.25 equiv of K4[Fe(CN)6] (NMP, 1408C,
Na2CO3, 0.05 mol% catalyst), only
Figure 1. Kinetics of the cyanation of 1-bromo-4-methoxybenzene
(1.0 mmol) with K4[Fe(CN)6] (0.25 mmol) in the presence and absence of
additives, catalyzed by 0.05 mol% of dichloroACTHNUTRGNE[NUG bis{1-(dicyclohexylphos-
phanyl)piperidine}]palladium (1) at 1408C in NMP (2.0 mL).
ladium(II)–cyano complexes, such as [Pd(CN)4]2À,
[(CN)3Pd(H)]2À, and [(CN)3Pd(Ar)]2À.[29] Indeed, when an
excess (about 5 equiv) of [Bu4N]13CN (d=161.7 ppm) was
added to a solution of compound 1 in NMP, [Pd
(d=126.4 ppm) was instantly and quantitatively formed at
(13CN)4]2À
ACHTUNGTRENNUNG
258C [Eq. (1)]. Similarly, [PdACHTUNGTRNEUNG
(13CN)4]2À was detected when
an excess (about 20 equiv relative to Pd) of [Bu4N]13CN was
added to a suspension of palladium on charcoal (5%) in
NMP and stirred for 18 h at 258C [Eq. (2)]. Furthermore,
when an excess (about 20 equiv) of [Bu4N]13CN was added
(C6H11)2})2Pd][24c] in NMP at
to a solution of [(PACHTUNGTRNNEUG{(NC5H10)ACHTUGNTRENNUGN
258C, [(13CN)3Pd(H)]2À (1H NMR: d=À10.8 ppm (dt, JCH
=
64.9 and 2.6 Hz), 13C{1H} NMR: d=148.2 and 137.8 ppm)
was smoothly and quantitatively formed at 258C when trace
amounts of water were present.[29] Subsequent heating
(1408C) of these reaction mixtures induced their complete
conversion into [PdACTHNUTRGNEUNG
(13CN)4]2À within 18 h [Eq. (1)].
56% yield was achieved when the
cyanation reaction was performed
with an equimolar amount of
K4[Fe(CN)6]. The same trend was
noted by changing the amount of
catalyst. Whereas the best results
were achieved with 0.05 mol% compound 1, decreased con-
version rates and yields were noted when the amount of cat-
alyst was lowered. The same observations were made when
the reaction was performed with increased amounts of cata-
lyst, presumably owing to the formation of palladium black
(Table 1, entries 23–25).
Even though [(CN)3Pd(Ar)]2À may be considered as an in-
termediate in both molecular mechanisms and in catalytic
cycles where palladium nanoparticles are involved, reductive
elimination of the respective benzonitrile was recently
shown to occur from tricoordinate [(CN)2Pd(Ar)]À, and
therefore required the predissociation of one of the cyanide
groups from [(CN)3Pd(Ar)]2À.[30] However, cyanide dissocia-
tion from [(CN)3Pd(Ar)]2À is known to be efficiently inhibit-
ed in the presence of even small quantities of cyanide in so-
Catalyst-deactivation processes and colorimetric cyanide de-
tection: The decreased catalytic activity of compound 1 in
the presence of a large excess of cyanating agent was as-
sumed to be a result of the higher concentration of “free”
cyanide ions in solution, and hence, (partial) catalyst deacti-
vation owing to the formation of (catalytically inactive) pal-
2980
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 2978 – 2986

Palladium-Catalyzed Cyanation of Aryl Bromides with K4[Fe(CN)6]
FULL PAPER
lution: whereas reductive elimination of benzonitrile was
observed at 258C when up to 3 equivalents of cyanide
excess in the reaction mixture, and hence affect molecular
catalytic mechanisms as well as inhibit catalysis when nano-
particles were involved.[32] Indeed, whereas no reaction was
([Bu4N]CN) were added to a solution of [(P
ACHTUGNRTNE{NUNG (NC5H10)-
G
observed when about 20 equivalents of K4[FeACTHUNGETRNNUG
(13CN)6] were
matic nitrile was not detectable when 4 equivalents (or
more) of [Bu4N]CN were present.[30] Consequently, the pres-
ence of cyanide inhibits catalysis, independent of the charac-
ter of the catalytically active species of compound 1, and
hence, independent of an operative molecular mechanism or
a catalytic cycle where palladium nanoparticles are involved.
The addition of an excess (about 10 equiv relative to the cat-
alyst) of [Bu4N]13CN to the cyanation reaction catalyzed by
compound 1 efficiently inhibited catalysis owing to the for-
added to a solution of compound 1 in NMP and stirred for
48 h at 258C, slow cyanide transfer, and thus the stepwise
transformation of compound
[Pd
(13CN)4]2À, was observed at 1408C. Complete conversion
of compound 1 into [Pd
(13CN)4]2À was achieved after 48 h.
Similarly, [Pd
(13CN)4]2À was quantitatively formed at 1408C
within 18 h when an excess (about 20 equiv) of
K4[Fe {(NC5H10)-
(13CN)6] was added to a solution of [(P
(C6H11)2})2Pd][24c] in NMP [Eq. (3)]. On the other hand, pal-
(13CN)2] and
1 into [PdACTHUNTGNRUEGN
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
A
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
mation of [Pd
N
U
ladium–cyano complexes were not detected when suspen-
sions of palladium on charcoal (5%) in NMP were thermal-
confirmed by NMR spectroscopy of the reaction mixtures.
The same was true for KCN or Zn(CN)2, which demonstrat-
ed the crucial role of the cyanating agent in the successful
cyanation of aryl halides and illustrated the sharp balance
between the rate of cyanide generation, efficient product
formation, and catalyst deactivation owing to excess cya-
nide. Indeed, whereas, for example, the generation of cya-
nide in a solution of Zn(CN)2 in NMP was indicated by its
colorimetric detection with the
ly treated in the presence of a large excess of K4[FeACHTUNGTRENNUNG
(13CN)6]
for 24 h [Eq. (2)]. Thus, this result indicates that catalyst de-
activation with hexacyanoferrate(II) as the cyanating agent
was too slow (if observable at all) to be relevant for the cat-
alytic performance of compound 1, in particular when palla-
dium nanoparticles were its catalytically active form.
corrin-based
chemosensor
b-solvato-a-cyanocobyrinic acid
heptamethyl ester (Scheme 3)
at 258C (the reaction of b-sol-
Catalysis: Complex 1 is a highly active, reliable, and versa-
tile cyanation catalyst that efficiently and cleanly converts
electronically activated, nonactivated, and deactivated, as
well as heterocyclic aryl bromides into their respective ben-
zonitriles (1408C, NMP solvent, Na2CO3 base, 0.05 mol%
catalyst, 0.25 equiv K4[Fe(CN)6] relative to the aryl bromide,
under a N2 atmosphere) within 6–8 h in typically good to ex-
cellent yields (Table 2). For example, the cyanation of elec-
tronically activated aryl bromides, such as 4-bromobenzoni-
trile, 4-bromobenzaldehyde, 1-(4-bromophenyl)ethanone,
methyl-4-bromobenzoate, 4-bromobenzophenone or 3-bro-
mobenzaldehyde, 3-bromobenzaldehyde diacetal, and
methyl-3-bromobenzoate afforded their respective benzoni-
triles within 6 h in almost quantitative yields. The same
yields but slightly lower conversion rates were noticed when
fluorinated aryl bromides, such as 1-bromo-4-(trifluorome-
thyl)benzene, 1-bromo-3-(trifluoromethyl)benzene, and
1-bromo-4-fluorobenzene were applied. High conversions
were also achieved after 6 h when nonactivated aryl bro-
mides, such as phenyl bromide, 1-bromo-3-methylbenzene,
4-bromobiphenyl, and 1-(4’-bromobiphenyl-4-yl)ethanone
were applied. The same yield but a retarded conversion was
noticed for 4-bromoacetanilide. Further retardation was ob-
tained for 4’-bromobiphenyl-4-yl acetate, for which a moder-
ate yield of 67% was achieved after 14 h. This result was in
contrast to electronically deactivated aryl bromides, such as
Scheme 3. Colorimetric detection of cyanide in NMP solutions of
K4[Fe(CN)4] (at 1408C) and Zn(CN)2 (at 258C) with the corrin-based
chemosensor b-solvato-a-cyanocobyrinic acid heptamethyl ester.
vato-a-cyanocobyrinic acid heptamethyl ester with cyanide
yielded the dicyanocobyrinic acid heptamethyl ester and
was accompanied by a characteristic color change from
orange to violet, with the absorption maxima for the a-band
at lmax =587 nm),[31] thermal treatment (>1008C) was re-
quired to induce cyanide production with hexacyanoferra-
te(II), and hence its colorimetric detection.
Consequently, the (main) reason for the successful cyana-
tion of aryl bromides with K4[Fe(CN)6] as the cyanating
agent (with catalyst 1) was the high stability of hexacyano-
ferrate(II), and thus the slow (optimal) rate of cyanide pro-
duction under the reaction conditions used. Even though
slow cyanide production may also have a retarding effect on
catalysis, too fast cyanide production would lead to its
1-bromo-4-methoxybenzene,
1-bromo-4-phenoxybenzene,
1-(benzyloxy)-4-bromobenzene, and 1-bromo-3,5-dimethox-
Chem. Eur. J. 2012, 18, 2978 – 2986
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2981

C. M. Frech et al.
ybenzene, for which full conversions into their re-
spective benzonitriles were typically achieved
within only 6–8 h. On the other hand, when nitro-
gen-containing substrates, such as 5-bromo-1-
methyl-1H-indole and 4-(4-bromophenyl)morpho-
line were applied, the same yields but retarded
conversions were noticed. For example, whereas
the quantitative formation of 4-morpholin-4-yl-
benzonitrile into 4-(4-bromophenyl)morpholine
was achieved within 8 h, 12 h were required to
fully convert 5-bromo-1-methyl-2,3-dihydro-1H-
indole into 1-methyl-2,3-dihydro-1H-indole-5-car-
bonitrile. Further retardation was found in the cy-
anation of 4-bromo-N,N-dimethylbenzene: 75%
conversion into 4-(dimethylamino)benzonitrile
was obtained after 14 h. Whereas similar rates of
conversion and yields were obtained when (nonac-
tivated, sterically hindered) 1-bromonaphthalene
and 2-bromotoluene were applied, a drop in activ-
ity was noted when using 2-bromobenzonitrile as
a substrate (46% conversion after 24 h). Further
retardation and even lower product yields (about
30%) were obtained when ortho-substituted aryl
bromides with increased electron density, such as
1-bromo-4-methoxy-2-methylbenzene or 1-bromo-
2-methoxybenzene, were applied. Almost no cata-
lytic activity was noticed when 2,6-di-substituted
aryl bromides, such as 2-bromo-1,3-dimethylben-
zene, were applied. Notably, protic substrates,
such as phenols, alcohols, carboxylic acids, acetic
acids, and primary and secondary amines were not
tolerated, most probably owing to HCN forma-
tion, and thus afforded their respective benzoni-
triles in typically<25% yield.[26]
On the other hand, an impressive level of activi-
ty was typically obtained when cyanation re-
AHCTUNGTREGaNNUN ctions were performed with heterocyclic aryl bro-
mides. For example, whereas quantitative product
formation was observed in 14–18 h when 3-bromo-
pyridine, 3-bromoquinoline, and 2-bromothio-
phene were used as substrates, 24 h was required
to achieve conversions of 60–90% when 3-bromo-
thiophene, 3-bromo-1-benzothiophene, 2-bromo-
3-methylthiophene, and 2-bromo-5-methylthio-
phene were applied. No product formation was
observed when 2-bromopyridine was used. How-
ever, this result was in striking contrast to
6-bromo-N-ethyl-N-phenylpyridin-2-amine and 1-
(6-bromopyridin-2-yl)-4-methylpiperazine,
two
randomly chosen 6-bromopyridine-2-amines,[24,33]
for which quantitative conversion into 6-[ethyl-
AHCTUNGTREG(NNUN phenyl)amino]pyridine-2-carbonitrile and 6-(4-
methylpiperazin-1-yl)pyridine-2-carbonitrile was
achieved after 24 h, respectively.[34]
Catalytic performance: Dichloro
ACHTUNGTREN[NUGN bis{1-(dicyclo-
hexylphosphanyl)piperidine}]palladium (1) is
a
2982
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 2978 – 2986

Palladium-Catalyzed Cyanation of Aryl Bromides with K4[Fe(CN)6]
FULL PAPER
highly active, versatile, easy accessible, and convenient pal-
ladium-based cyanation catalyst that was able to convert a
large variety of substituted aryl bromides, containing fluo-
ride atoms, trifluoromethane groups, nitriles, acetals, ke-
tones, aldehydes, ethers, esters, amides, as well as heterocy-
clic aryl bromides, such as pyridines and their derivatives, or
thiophenes into their respective aromatic nitriles with
K4[Fe(CN)6] as the cyanating agent (0.25 equiv) at 1408C in
NMP solvent, with Na2CO3 as the base and 0.05 mol% cata-
lyst (or 0.1 mol% when heterocyclic substrates were used).
Comparisons with the literature data of other catalysts are
difficult to draw because almost every cyanation catalyst op-
erates under different reaction conditions (solvent, reaction
temperature, base, catalyst loading, cyanating agent, addi-
tive, etc.), which are crucial for their successful application.
Therefore, a comparative study was performed between
indeed one of the most-active cyanation catalysts. For exam-
ple, whereas conversions of>95% for 4-bromobiphenyl or
3-bromoquinoline were observed with all of the catalysts
tested (Table 3, entries 3 and 6, respectively), significantly
lower conversions of 33, 68, and 56% were achieved with
palladium acetate in the cyanation of 1-(4-bromophenyl)-
AHCTUNGTREGeNNNU thanone, 1-bromo-4-methoxybenzene or 3-bromopyridine,
respectively (Table 3, entries 1, 4, and 5). On the other
hand, whereas excellent performance was noted for com-
pound 1 and [PdACHTNUTRGNEUNG(OAc)2] when 1-bromo-4-(trifluoromethyl)-
benzene and 3-bromothiophene were used as the substrates
(Table 3, entries 2 and 12), lower conversions were observed
with [Pd
alytic activity was found for catalyst 1 when compared to
both [Pd(OAc)2] and [Pd(OAc)2]/dppf in the cyanation of
ACHTUNGRTEN(NUNG OAc)2]/dppf. Moreover, significantly improved cat-
A
ACHTUNGTRENNUNG
1-(6-bromopyridin-2-yl)-4-methylpiperazine and 2-bromo-5-
methylthiophene (Table 3, entries 7 and 8, respectively):
whilst 6-(4-methylpiperazin-1-yl)pyridine-2-carbonitrile was
almost quantitatively formed with compound 1, conversions
of 40 and 53% were achieved with palladium acetate and
compound 1, [Pd
ACHTUNGTRENNUNG ACHTUNGTRENNUNG
(OAc)2],[19b] and [Pd
1,1’-bis(diphenylphosphino)ferrocene)[19l]
which similar reaction conditions were applied. This study
demonstrated that the aminophosphine-based system was
[PdACTHNUGTRNENG(U OAc)2]/dppf, respectively. Similarly, whilst a conversion
of 60% into 5-methylthiophene-2-carbonitrile was achieved
with compound 1, only 10 and 23% of the reaction product
Table 3. Comparison of compound 1, [PdACHTNUTRGENN(UG OAc)2], and [PdACHTUNGTERN(NUGN OAc)2]/dppf
in the cyanation of aryl bromides with K4[Fe(CN)6] under a N2 atmos-
phere.[a]
was achieved with [Pd
tively. Slightly lower catalytic activity of compound 1 (com-
pared to [Pd(OAc)2]/dppf) was noticed with 2-bromo-3-
methylthiophene as the substrate. Overall, the catalytic ac-
tivity of compound 1 was superior to [Pd(OAc)2] and com-
parable to [Pd(OAc)2]/dppf. Therefore, the aminophos-
ACHTUNGTNERNUG(OAc)2] and [PdCAHTUNGTREN(NUGN OAc)2]/dppf, respec-
Entry
Cyanation
product
Catalyst
Conv.
[%][b]
t
[h]
AHCTUNGTRENNUNG
1
[Pd
[Pd
100
33
96
G
6
AHCTUNGTRENNUNG
1
2
3
R
AHCTUNGTRENNUNG
phine-based system is one of the most-active cyanation cata-
lysts reported to date. The improved catalytic activity of the
aminophosphine-based system in cyanation reactions when
compared to other pre-catalysts (see mechanistic investiga-
tions below), such as [PdACTHNUTRGNE(UGN OAc)2], was attributed to the slow
and controllable heat-induced degradation of catalyst 1 by
1
100
96
83
100
98
100
[Pd
[Pd
1
[Pd
[Pd
G
14
6
R
ACHTUNGTRENNUNG
G
À
hydroxide anions (involving the cleavage of the P N bond),
1
100
68
100
100
56
100
100
100
100
100
40
which resulted in the formation of palladium nanoparticles.
Consequently, Na2CO3 is an excellent promoter for the mild
and controlled formation of nanoparticles from dichloro-
4
[Pd
[Pd
1
[Pd
[Pd
1
[Pd
[Pd
1
[Pd
[Pd
N
6
18
14
24
5[c]
AHCTUNGTREGbNNNU is(aminophosphine) complexes of palladium, because hy-
droxide anions are slowly generated from the equilibrium
6[c]
À
between HCO3 , OHÀ, and CO2.[35] Thus, the formation of
small, highly active nanoparticles (the ratio between palladi-
um on the outer rim and palladium on the inside of the par-
ticles increased favorable with decreasing particle size) was
a direct consequence of the slow release of palladium from
catalyst 1, promoted by hydroxide anions generated by
Na2CO3.[36]
7[c]
53
1
60
10
23
76
57
100
67
72
8[c]
[Pd
[Pd
1
[Pd
[Pd
1
18
18
24
9[c]
Mechanistic investigations: The following experimental ob-
servations indicated that compound 1 was indeed a pre-cata-
lyst, which transformed into palladium nanoparticles:[37]
1) The addition of a large excess of metallic mercury to the
reaction mixture of an aryl bromide, K4[Fe(CN)6], a base,
and a catalyst efficiently stopped the catalysis process. Con-
sequently, no product formation was observed when Hg0
was added at the beginning of the reaction (Figure 1). The
10[c]
[Pd
[Pd
20
[a] Reaction conditions: aryl bromide (1.0 mmol), K4[Fe(CN)6]
(0.25 mmol), base (1.0 mmol), NMP (4.0 mL total), catalyst (0.05 mol%),
1408C. [b] Determined by GCMS based on the aryl halide. [c] Reactions
were performed in the presence of 0.1 mol% catalyst.
Chem. Eur. J. 2012, 18, 2978 – 2986
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2983

C. M. Frech et al.
same observations were also made when poly(4-vinylpyri-
dine) (PVPy; 2% cross-linked with divinylbenzene) was
used.[38,39] 2) Sigmoidal-shaped kinetics with induction peri-
ods, which is characteristic for metal-particle formation, and
autocatalytic surface growth that can lead to soluble mono-
disperse nanoclusters (or insoluble bulk-metal formation),
were observed in all of the reactions examined.[40] 3) Whilst
the presence of 0.1 equivalent of CS2 or triphenylphosphine
(relative to the catalyst) efficiently inhibited the catalysis,
only slightly lowered conversion rates and yields were no-
ticed in the presence of about 20 mol% NBu4Br (relative to
the aryl bromide).[41] 4) The UV/Vis spectrum of the reac-
tion mixtures exhibited a continuous absorption, with a
steep rise in absorbance at shorter wavelengths. This obser-
vation was a clear indication of the presence of palladium
nanoparticles (Figure 2).[42] 5) Lastly, catalyst loadings
>0.2 mol% led to significantly lower conversions, most prob-
ably owing to the formation of inactive palladium black.[43]
ular catalytic mechanisms as well as inhibit the catalysis
when nanoparticles were involved (owing to the formation
of [Pd(CN)4]2À, [(CN)3Pd(H)]2À, and [(CN)3Pd(Ar)]2À).
Even though slow cyanide production may retard catalysis,
too fast cyanide production led to excess cyanide in the re-
action mixtures, thereby affecting the molecular catalytic
mechanisms at every step of the cycle as well as inhibiting
the catalysis when nanoparticles were involved. Consequent-
ly, the choice of cyanating agent is crucial for the successful
cyanation of aryl halides. Therefore, the (main) reason for
the successful cyanation of aryl bromides with K4[Fe(CN)6]
as the cyanating agent (at least with catalyst 1) was its high
stability and hence the optimal (slow) rate of cyanide pro-
duction (at least under the reaction conditions applied). Col-
orimetric cyanide detection involving the conversion of
b-solvato-a-cyanocobyrinic acid heptamethyl ester into di-
cyanocobyrinic acid heptamethyl ester in NMP indicated
that cyanide production with hexacyanoferrate(II) started at
reaction temperatures>1008C. Mechanistic investigations
demonstrated that palladium nanoparticles were the catalyt-
ically active form of compound 1.
Experimental Section
General procedure for the cyanation reactions: In a Young Schlenk tube
(10 mL) were added dried K4[Fe(CN)6] (0.25 mmol), K2CO3 (1.0 mmol),
anhydrous 1-methyl-2-pyrrolidinone (NMP, 2.0 mL), aryl bromide
(1.0 mmol), and an appropriate amount of catalyst (dissolved in NMP)
under a N2 atmosphere. The mixture was vigorously stirred at 1408C.
Samples were periodically taken from the reaction mixture, quenched
with H2O (or with sat. NH4Cl when pyridines were used as substrates),
extracted with EtOAc, and analyzed by GCMS. At the end of reaction,
the mixtures were allowed to cool to RT, quenched with H2O (or sat.
NH4Cl), and extracted with EtOAc (3ꢂ15 mL). The combined extracts
were dried (MgSO4) and evaporated to remove the solvent. The crude
material was purified by column chromatography on silica gel as necessa-
ry.
Figure 2. UV/Vis spectra for diluted reaction mixtures of the cyanation of
1-bromo-4-methoxybenzene, recorded after 2 h: a) with K4[Fe(CN)6]
(0.25 equiv relative to the aryl bromide), catalyzed by 0.1 mol% of com-
pound 1; b) with K4[Fe(CN)6] (0.25 equiv relative to the aryl bromide) in
the absence of compound 1; and c) in the absence of both K4[Fe(CN)6]
and compound 1.
Acknowledgements
Financial support by the University of Zurich and the Swiss National Sci-
ence Foundation (SNSF) is acknowledge. We also acknowledge
F. Zelder, C. Mꢃnnel-Croise, and Balz Aebli from the Inorganic Chemis-
try Institute of the University of Zꢀrich, Switzerland, for performing the
cyanide-detection experiments.
Conclusion
Dichloro
ACHTUNGTRENNUNG
um [(P{(NC5H10)
A
ACHTUNGTRENNUNG
À
generally applicable C C cross-coupling catalyst that, apart
from successfully catalyzing the Suzuki, Heck, and Negishi
reactions, also efficiently converts various electronically acti-
vated, nonactivated, and deactivated as well as heterocyclic
aryl bromides into their respective benzonitriles in high
[1] a) C. Grundmann in Houben-Weyl: Methoden der organischen
Chemie, Vol. E5, 4th ed., Thieme, Stuttgart, 1985; b) R. C. Larock,
Comprehensive Organic Transformations, VCH, Weinheim, 1989.
[2] A. Kleemann, J. Engel, B. Kutscher, D. Reichert in Pharmaceutical
Substances: Syntheses, Patents, Applications, 4th ed., Thieme, Stutt-
gart, New York, 2001, pp. 241–242, pp. 488–489, pp. 553, pp. 825–
826, p. 1154, pp. 1598–1599.
[3] Z. Rappoport in Chemistry of the Cyano Group, Wiley, London,
1970.
[4] 6-Aminopyridine-2-carbonitriles, for example, are important syn-
thetic intermediates in various pharmaceutically active compounds,
such as 6-aminopyridine-2-carboxamides, which have applications as
serotonin (5-hydroxytryptamine) (5-ht)-receptor ligands (in particu-
yields.
[Pd(OAc)2]/dppf showed that the catalytic activity of the
aminophosphine-based system was superior to [Pd(OAc)2]
and comparable to [Pd(OAc)2]/dppf under the reaction con-
A
comparison study with [PdACTHNGUTRENU(NG OAc)2] and
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
ditions applied. Catalyst-deactivation processes were studied
and excess cyanide was found to efficiently affect the molec-
2984
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 2978 – 2986

Palladium-Catalyzed Cyanation of Aryl Bromides with K4[Fe(CN)6]
FULL PAPER
lar for the 5-ht6 subtype).[5] Hence, these nitriles are useful for the
treatment of diseases for which modulation of the activity of the
5-ht receptor is desired. Thus, 6-aminopyridine-2-carboxamides
could find their application in the enhancement of cognition and
memory in patients with Alzheimerꢄs disease,[6] in the treatment of
attention deficit hyperactivity disorder (ADHD),[7] and in the treat-
ment of diseases of the central nervous system, such as schizophre-
nia and depression.[8] In addition, 5-ht6-receptor antagonists are
useful in the treatment of obesity.[9] Another class of compound with
lein, N. Mꢀller, M. Beller, Synlett 2007, 555; g) T. Schareina, A.
Zapf, A. Cotte, N. Mꢀller, M. Beller, Synthesis 2008, 20, 3351.
[18] For reviews on palladium-catalyzed cyanations, see: a) M. Sunder-
meier, A. Zapf, M. Beller, Eur. J. Inorg. Chem. 2003, 3513; b) J.
Tsuji in Transition Metal Reagents and Catalysts—Innovations in Or-
ganic Synthesis Wiley, Chichester, 2000; c) L. Brandsma, S. F. Vasi-
levsky, H. D. Verkruijsse in Application of Transition Metal Catalysts
in Organic Synthesis Springer, Berlin, 1999, pp. 149–177; d) G. P.
Ellis, T. M. Romney-Alexander, Chem. Rev. 1987, 87, 779; e) R. F.
Heck in Palladium Reagents in Organic Syntheses, Academic Press,
London, 1985.
pharma
ACHTUNGTRENNUNGceutical importance that involve 6-aminopyridine-2-carboni-
triles as synthetic intermediates are 3-substituted imidazo-
A
[19] a) K. Takagi, T. Okamoto, Sakakibara. Y, S. Oka, Chem. Lett. 1973,
471; b) T. Schareina, A. Zapf, M. Beller, J. Organomet. Chem. 2004,
689, 4576; c) B. A. Anderson, E. C. Bell, F. O. Ginah, N. K. Harn,
L. M. Pagh, J. P. Wepsiec, J. Org. Chem. 1998, 63, 8224; d) F. G.
Buono, R. Chidambaram, R. H. Muller, R. E. Waltermire, Org. Lett.
2008, 10, 5325; e) M. Sundermeier, A. Zapf, S. Mutyala, W. Bau-
mann, J. Sans, S. Weiss, M. Beller, Chem. Eur. J. 2003, 9, 1828; f) T.
Schareina, R. Jackstell, T. Schulz, A. Zapf, A. Cottꢅ, M. Gotta, M.
Beller, Adv. Synth. Catal. 2009, 351, 643; g) M. Sundermeier, A.
Zapf, M. Beller, J. Sans, Tetrahedron Lett. 2001, 42, 6707; h) F. Jin,
P. N. Confalone, Tetrahedron Lett. 2000, 41, 3271; i) H. R. Chobani-
an, B. P. Fors, L. S. Lin, Tetrahedron Lett. 2006, 47, 3303; j) A.
Littke, M. Soumeillant, R. F. Kaltenbach, R. G. Cherney, C. M.
Tarby, S. Kiau, Org. Lett. 2007, 9, 1711; k) P. Y. Yeung, C. M. So,
C. P. Lau, F. Y. Kwong, Org. Lett. 2011, 13, 648; l) T. Schareina, A.
Zapf, M. Beller, Chem. Commun. 2004, 1388; m) L.-H. Li, Z.-L.
Pan, X.-H. Duan, Y.-M. Liang, Synlett 2006, 13, 2094; n) O. Gross-
mann, D. Gelman, Org. Lett. 2006, 8, 1189; o) Y. Ren, Z. Liu, S. He,
S. Zhao, J. Wang, R. Niu, W. Yin, Org. Process Res. Dev. 2009, 13,
764; p) P. Y. Yeung, C. M. So, C. P. Lau, F. Y. Kwong, Angew. Chem.
2010, 122, 9102; Angew. Chem. Int. Ed. 2010, 49, 8918.
tors that are useful antianginal and cardioprotective agents.[10,11]
[5] E. M. Collantes, J. B. Schwarz, U. S. Pat. Appl. Publ. (2009), US
20090197859.
[6] a) C. Gꢅrard, M.-P. Martres, K. Lefevre, M. C. Miquel, D. Verge, R.
Lanfumey, E. Doucet, M. Hamon, S. El Mestikawy, Brain Res. 2009,
746, 207; b) J. C. Bentley, A. Boursson, F. G. Boess, F. C. Kone, C. A.
Marsden, N. Petit, A. Sleight, Br. J. Pharmacol. 1999, 126, 1537;
c) L. A. Dawson, H. Q. Nguyen, P. Li, Br. J. Pharmacol. 2000, 130,
23; d) M. L. Woolley, C. A. Marsden, A. J. Sleight, K. C. F. Fone,
Psychopharmacology 2003, 170, 358.
[7] M. Ernst, A. J. Zametkin, J. H. Matochik, P. A. Jons, R. M. Cohen, J.
Neurosci. 1998, 18, 5901.
[8] T. A. Branchek, T. P. Blackburn, Annu. Rev. Pharmacol. Toxicol.
2000, 40, 319.
[9] J. Holenz, P. J. Pauwels, J. Luis Dꢆaz, R. Mercꢇ, X. Codony, H.
Buschmann, Drug. Circ. Drug. Discov. Today 2006, 11, 283.
[10] S. Ahmed, S. C. Wu, S. V. OꢄNeil, K. Ngu, K. S. Atwal, D. S. Wein-
stein, U. S. Pat. Appl. Pub. (2000) US 7326705.
[11] Notably, imidazo
ic light-emitting diodes (OLEDs) and organic thin-layer field-effect
transistors (FETs), are also prepared from pyridine-2-carbonitriles.
For example, see: a) D. Kitazawa, G. Tominaga, A. Takano, Jpn.
Kokai Tokkyo Koho JP 2001057292; b) H. Nakamura and H. Yama-
moto, PCT Int. Appl. WO 2005043630.
[20] In the classical (Pd0/PdII) mechanism, coordinatively unsaturated
(14eÀ) Pd0 complexes of type [Pd(L)2] (L=phosphine, etc.) are as-
sumed to be the catalytically active species. The oxidative addition
of an aryl halide onto the metal center of [Pd(L)n] (n=1 or 2) and
II
À
[12] a) J. Lindley, Tetrahedron 1984, 40, 1433; b) R. G. R. Bacon,
H. A. O. Hill, J. Chem. Soc. 1964, 1097; c) J. von Braun, G. Manz,
Justus Liebigs Ann. Chem. 1931, 488, 111; d) K. W. Rosenmund, E.
Struck, Ber. Dtsch. Chem. Ges. 1919, 52, 1749; e) Meister, Lucius,
Brꢀning DE 271,790, 1913.
[13] a) T. Sandmeyer, Ber. Dtsch. Chem. Ges. 1885, 18, 1946; b) T. Sand-
meyer, Ber. Dtsch. Chem. Ges. 1885, 18, 1492; c) T. Sandmeyer, Ber.
Dtsch. Chem. Ges. 1884, 17, 2650.
[14] a) A. C. Stevenson, Ind. Eng. Chem. 1949, 41, 1846; b) W. I. Denton,
R. B. Bishop, H. P. Caldwell, H. D. Chapman, Ind. Eng. Chem. 1950,
42, 796; c) A. Martin, N. V. Kalevaru, B. Lꢀcke, J. Sans, Green
Chem. 2002, 4, 481.
[15] a) X. Wang, B. Xhi, J. Baum, Y. Chen, R. Crockett, L. Huang, S. Ei-
senberg, J. Ng, R. Larsen, M. Martinelli, P. Reider, J. Org. Chem.
2006, 71, 4021; b) J. X. Qiao, X. Cheng, D. P. Modi, K. A. Rossi,
J. M. Luettgen, R. M. Knabb, P. K. Jadhav, R. R. Wexler, Bioorg.
Med. Chem. Lett. 2005, 15, 29; c) J. M. Veauthier, C. N. Carlson,
G. E. Collis, J. L. Kiplinger, K. D. John, Synthesis 2005, 2683.
[16] For nickel-catalyzed cyanation reactions, see: a) Y. Sakakibara, Y.
Ido, K. Sasaki, M. Sakai, N. Uchino, Bull. Chem. Soc. Jpn. 1993, 66,
2776; b) L. Cassar, S. Ferrara, M. Foꢈ, Adv. Chem. Ser. 1974, 132,
252; c) L. Cassar, M. Foꢈ, F. Montanari, G. P. Marinelli, J. Organo-
met. Chem. 1979, 173, 335; d) Y. Sakakibara, F. Okuda, A. Shimoya-
bashi, K. Kirino, M. Sakai, N. Uchino, K. Takagi, Bull. Chem. Soc.
Jpn. 1988, 61, 1985; e) K. Takagi, Y. Sakakibara, Chem. Lett. 1989,
1957.
[17] For copper-catalyzed cyanation reactions, see: a) H. J. Cristau, A.
Ouali, J. F. Spindler, M. Taillefer, Chem. Eur. J. 2005, 11, 2483; b) J.
Zanon, A. Klapars, S. L. Buchwald, J. Am. Chem. Soc. 2003, 125,
2890; c) J. X. Wu, B. Beck, R. X. Ren, Tetrahedron Lett. 2002, 43,
387; d) T. Schareina, A. Zapf, W. Magerlein, N. Muller, M. Beller,
Chem. Eur. J. 2007, 13, 6249; e) T. Schareina, A. Zapf, M. Beller,
Tetrahedron Lett. 2005, 46, 2585; f) T. Schareina, A. Zapf, W. Mꢃger-
subsequent transmetalation (with M CN) yield Pd intermediates
À
of type [Pd(Ar)(CN)(L)n] (and M X). Reductive elimination of
[19c–e]
À
Ar CN regenerates the catalyst.
[21] For examples where palladium nanoparticles have been proposed as
the catalytically active form: a) K. Chattopadhyay, R. Dey, B. C.
Ranu, Tetrahedron Lett. 2009, 50, 3164; b) I. P. Beletskaya, A. V. Se-
livanova, V. S. Tyurin, V. V. Matveev, A. R. Khokholov, Russ. J. Org.
Chem. 2010, 46, 157; c) G. Cheng, J. Wenig, Z. Zheng, X. Zhu, Y.
Cai, J. Cai, Y. Wan, Eur. J. Org. Chem. 2008, 3524; d) S. Velmathi,
N. E. Leadbeater, Tetrahedron Lett. 2008, 49, 4693.
[22] K4[Fe(CN)6] is nontoxic (the Ld50 of K4[Fe(CN)6] is lower than that
of NaCl) and is even used in the food industry for metal precipita-
tion in wine, see: Rçmpp Lexikon Chemie—Version 2.0, Thieme,
Stuttgart/New York, 1999. K4[Fe(CN)6] was recently introduced as a
cyanating agent by M. Beller and co-workers.[19l]
[23] J. L. Bolliger, C. M. Frech, Chem. Eur. J. 2010, 16, 4075.
[24] a) J. L. Bolliger, M. Oberholzer, C. M. Frech, Adv. Synth. Catal.
2011, 353, 945; b) M. Oberholzer, R. Gerber, C. M. Frech, unpub-
lished results; c) J. L. Bolliger, C. M. Frech, Chem. Eur. J. 2010, 16,
11072.
[25] Even though reaction temperatures of 1208C could be used for most
of the substrates without a noticeable loss of activity, the advantage
of applying 1408C lay in the reliability and scope of the procedure,
which would allow the direct adoption of the procedure to other
aryl bromides without the necessity of its modification.
[26] Cyanation reactions performed in protic solvents or with protic aryl
halides (see below) are not appropriate because hydrolysis of CNÀ
into HCN could occur, which competes with the aryl halide for oxi-
dative addition onto the metal center. We assumed that the oxida-
tive addition of HCN onto a bis-phosphine palladium(0) metal
center
generates
palladium–hydride
complexes
of
type
[(PR3)2Pd(H)(CN)], which can react with another equivalent of
HCN to give complexes of the type [(PR3)2Pd(CN)2] and H2.[27,28]
Chem. Eur. J. 2012, 18, 2978 – 2986
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2985

C. M. Frech et al.
Subsequent phosphine substitution by cyanide generates
[Pd(CN)4]2À, which was indeed formed when the reactions were per-
formed in protic solvents or protic aryl bromides.
hexylphosphanyl)piperidine)palladium (1) in the cyanation reaction,
aminophosphines have an additional advantage when compared to
water-insoluble phosphine-based ligand systems: the treatment of
compound 1 under work-up conditions in air was recently shown to
lead to rapid and complete catalyst degradation into decomposition
products that were easily separated from the coupling products—an
often ignored but very important issue for the preparation of phar-
maceutically active compounds etc.
[27] a) V. V. Grushin, Chem. Rev. 1996, 96, 2011; b) R. Gerber, T. Fox,
C. M. Frech, Chem. Eur. J. 2010, 16, 6771.
[28] For an excellent paper about the mechanisms of catalyst poisoning
in palladium-catalyzed cyanation reactions, see: S. Erhardt, V. V.
Grushin, A. H. Kilpatrick, S. A. Macgregor, W. J. Marshall, D. C.
Roe, J. Am. Chem. Soc. 2008, 130, 4828.
[29] a) J. E. Marcone, K. G. Moloy, J. Am. Chem. Soc. 1998, 120, 8527;
b) J. Huang, C. M. Haar, S. P. Nolan, J. E. Marcone, K. G. Moloy,
Organometallics 1999, 18, 297.
[37] For a comprehensive review on the problem of distinguishing truly
homogeneous catalysis from soluble- or heterogeneous catalysis
with other metal particles under reducing conditions, see: J. A. Wi-
degren, R. G. Finke, J. Mol. Catal. A 2003, 198, 317.
[30] K. D. Dobbs, W. J. Marshall, V. V. Grushin, J. Am. Chem. Soc. 2007,
129, 30.
[38] PVPy is a palladium trap that can effectively inhibit catalysis when
nanoparticles are involved in the catalytic cycle. Nevertheless, Pd
nanoparticles that were immobilized on PVPy spheres have been
applied as catalyst in Heck reactions.[39] Therefore, this test is not
definitive and does neither conclusively confirm nor rule out either
catalysis by the surface of Pd nanoparticles or by molecular Pd0 spe-
cies but can strongly support the conclusions drawn from other tests
performed. For details, see: K. Yu, W. Sommer, M. Weck, C. W.
Jones, J. Catal. 2004, 226, 101 and the references therein.
[39] S. Pathak, M. T. Greci, R. C. Kwong, K. Mercado, G. K. S. Prakash,
G. A. Olah, M. E. Thompson, Chem. Mater. 2000, 12, 1985.
[40] a) M. A. Watzky, R. G. Finke, J. Am. Chem. Soc. 1997, 119, 10382;
b) J. A. Widegren, M. A. Bennett, R. G. Finke, J. Am. Chem. Soc.
2003, 125, 10301.
[41] Tetrabutylammonium bromide is known to stabilize palladium nano-
particles. For example, see: a) T. Jeffery In Advances in Metal-Or-
ganic Chemistry, Vol. 5 (Ed.: L. S. Liebeskind), JAI, Greenwich,
1996, p 153; b) M. T. Reetz, E. Westermann, Angew. Chem. 2000,
112, 170; Angew. Chem. Int. Ed. 2000, 39, 165; c) A. Zapf, M.
Beller, Chem. Eur. J. 2001, 7, 2908.
[31] a) C. Mꢃnnel-Croisꢅ, F. Zelder, Inorg. Chem. 2009, 48, 1272; b) C.
Mꢃnnel-Croisꢅ, B. Probst, F. Zelder, Anal. Chem. 2009, 81, 9493.
[32] Another reason for the high catalytic performance of compound 1
with K4[Fe(CN)6] as the cyanating agent might be the reduction
power of hexacyanoferrate(II), which could reduce (inactive)palla-
dium(II)–cyano complexes into catalytically active palladium(0) spe-
cies. By analogy, hexacyanoferrate(II) could also reduce traces of
oxygen and other impurities from the reaction mixture.
[33] J. L. Bolliger, C. M. Frech, Tetrahedron 2009, 65, 1180.
[34] Notably, even though full conversion of 1-(6-bromopyridin-2-yl)-4-
methylpiperazine into 6-(4-methylpiperazin-1-yl)pyridine-2-carboni-
trile was achieved within 24 h, no product formation was noted
when 2-bromo-6-piperidin-1-ylpyridine or 4-(6-bromopyridin-2-yl)-
morpholine were used. The reason for the striking difference in re-
activity of these structurally closely related compounds is unknown.
Nevertheless, because no product formation at all was observed
when [PdACHTUNGTRENNUNG(OAc)2] or [PdACHTUNGTNER(NUGN OAc)2]/dppf (dppf=1,1’-bis(diphenylphos-
phino)ferrocene) was used as the catalyst, it cannot be excluded that
trace amounts of an impurity is the reason for their unsuccessful cy-
anation.
[42] J. A. Creighton, D. G. Eadon, J. Chem. Soc. Faraday Trans. 1991, 87,
3881.
[35] X. Wang, W. Conway, R. Burns, N. McCann, M. Maeder, J. Phys.
Chem. A 2010, 114, 1734.
[43] a) 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;
b) A. H. M. de Vries, F. J. Parlevliet, L. Schmieder-van de Vonder-
voort, J. H. M. Mommers, H. J. W. Henderickx, M. A. M. Walet, J. G.
de Vries, Adv. Synth. Catal. 2002, 344, 996; c) M. Beller, H. Fischer,
K. Kꢀhlein, C.-P. Reisinger, W. A. Herrmann, J. Organomet. Chem.
1996, 520, 257.
[36] The morphologies of the nanoparticles also determine their catalytic
activity. However, nanoparticle morphology decisively depends on
the conditions in which they are grown, such as the concentration of
palladium and the rate of catalyst degradation, and hence, on the
rate of nanoparticle growth, the presence of ligands and other addi-
tives, as well as on the palladium source itself. This dependence ex-
plains the different catalytic activities of the pre-catalysts. Moreover,
apart from the excellent catalytic activity of dichloro-bis(1-(dicyclo-
Received: September 19, 2011
Published online: February 1, 2012
2986
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 2978 – 2986