Ligand-free palladium-catalyzed cyanation of aryl halides

Article abstract of DOI:10.1021/jo0481250

(Chemical Equation Presented) A practical, ligand-free cyanation of aryl bromides that utilizes as little as 0.1 mol % Pd(OAc)₂ in combination with a nontoxic cyanide source, M₄[Fe(CN)₆] (M = K, Na), is described. The reactions are performed in DMAC at 120°C and provide the corresponding aryl nitrile in 83-96% yield, typically in less than 5 h. TON values of up to 7100 were attained.

Full text of DOI:10.1021/jo0481250

functional group tolerance, air stability, and high cata-
lytic activity, but none of these methods offered the
ligand-free advantage. One constraint of these proce-
dures, which typically utilize M-CN (M ) Na, K, TMS,
Cu) as the nucleophile, is the high level of dissolved
cyanide in the reaction that inhibits the catalytic cycle,
namely, the oxidative insertion, due to formation of
unreactive palladium(II) cyano species. This has led to
the use of additives such as zinc acetate,7 diamines,8 zinc
dust,9 and Me3SnCl10 to enhance the catalytic turnover.
Controlling the CN concentration via defined dosing of
the cyanide has also been used toward this end,11 as has
the use of less soluble cyanide reagents such as zinc
cyanide12 and potassium ferrocyanide(II) (K4[Fe(CN)6]).13
The latter reagent, recently rediscovered14 as a cyanide
source by Beller, is particularly intriguing because all
six CN are available for reaction and it is inexpensive,
easily handled, and nontoxic.15 The use of ligands was
thought to improve the catalytic activity and allowed for
milder reaction conditions and the inclusion of typically
unreactive aryl chlorides. The phosphine ligands though
are often air/moisture-sensitive, more costly than the
palladium species, and difficult to remove from the
product and to recover. This has led to a reexamination
of the role of ligands in Pd-catalyzed aromatic substitu-
tion reactions, as evidenced by recent work describing
ligand-free Heck,16 Suzuki,17 and Sonogashira18 reactions.
Those results suggest that at low Pd(0) concentrations,
the rate of oxidative addition exceeds that of the catalyst
aggregation.14a As noted by Beletskaya,19 “the inherent
reactivity of unligated palladium is sufficient for oxida-
tive addition to most kinds of C-X bonds.” This prompted
Ligand-Free Palladium-Catalyzed
Cyanation of Aryl Halides
Steven A. Weissman,* Daniel Zewge, and Cheng Chen
Department of Process Research, Merck Research
Laboratories, Merck & Co., Inc., Rahway, New Jersey 07065
steven_weissman@merck.com
Received October 22, 2004
A practical, ligand-free cyanation of aryl bromides that
utilizes as little as 0.1 mol % Pd(OAc)2 in combination with
a nontoxic cyanide source, M4[Fe(CN)6] (M ) K, Na), is
described. The reactions are performed in DMAC at 120 °C
and provide the corresponding aryl nitrile in 83-96% yield,
typically in less than 5 h. TON values of up to 7100 were
attained.
Aromatic nitriles constitute a key component of nu-
merous commercial compounds, including pharmaceuti-
cals, agrochemicals (herbicides, pesticides), and pigments
and dyes.1 Their utility also stems from the myriad of
possible nitrile transformations, including the synthesis
of benzoic acids/esters, amidines, amides, imidoesters,
benzamidines, amines, heterocycles, and aldehydes.2 The
traditional method for preparing aromatic nitriles from
the corresponding aryl bromides/iodides, Rosemund-Von
Braun reaction,3 requires stoichiometric copper(I) cyanide
at elevated temperatures, often with complicated work-
ups. In 1973, Takagi described the first metal-catalyzed
cyanation of aryl halides that also happened to be a
ligand-free system.4 This methodology employed KCN
and 2 mol % Pd(CN)2 (for the aryl bromides substrates)
with conversions ranging from 64 to 91% at g140 °C. A
subsequent modification using catalytic KOH and KI in
conjunction with 1.5 mol % Pd(OAc)2 at 90 °C was also
reported by that group.5 Since then, palladium-based
methods6 have garnered most of the attention due to their
(7) Chidambaram, R. Tetrahedron Lett. 2004, 1441-1444.
(8) Sundermeier, M.; Zapf, A.; Mutyala, S.; Baumann, W.; Sans, J.;
Weiss, S.; Beller, M. Chem. Eur. J. 2003, 9, 1828-1836.
(9) Okano, T.; Iwahara, M.; Kiji, J. Synlett 1998, 243-244.
(10) Yang, C.; Williams, J. M. Org. Lett. 2004, 6, 2837-2840.
(11) TMS-CN: Sundermeier, M.; Mutyala, S.; Zapf, A.; Spannen-
berg, A.; Beller, M. J. Organomet. Chem. 2003, 684, 50-55. Acetone
cyanohydrin: Sundermeier, M.; Zapf, A.; Beller, M. Angew. Chem., Int.
Ed. 2003, 42, 1661-1664.
(12) Use of zinc cyanide in this capacity was introduced by: Tschaen,
D. M.; Desmond, R.; King, A. O.; Fortin, M. C.; Pipik, B.; King, S.;
Verhoeven, T. R. Synth. Commun. 1994, 24, 887-890.
(13) Schareina, T.; Zapf, A.; Beller, M. Chem. Commun. 2004, 12,
1388-1389. We observed no need to dehydrate (3 equiv of water
present) this reagent as was described by the authors in this reference.
The reagent particle size is critical, as material obtained from Alfa/
Aesar performed well (mean particle size 312 µm; 90% <509 µm), while
material from Acros (mean particle size 464 µm; 90% <822 µm) showed
only 5% conversion to benzonitrile after 10 h at 120 °C. The latter
material performed typically (99% conversion/1 h) after pulverization
to values of 184 and 374 µm, respectively.
(1) For example, The Merck Index (13th ed.; O’Neil, M. J., Ed.;
Merck: 2001) lists 22 compounds that incorporate this functionality:
Monograph nos. 926, 1098, 1204, 1426, 1446, 1477, 2185, 2283, 2299,
2342, 2712, 2720, 3068, 3958, 5469, 5481, 5576, 6690, 7242, 9615.
(2) Chemistry of the Cyano Group; Rappoport, Z., Ed.; John Wiley
& Sons: London 1970.
(3) (a) For a review, see: Mowry, D. T. Chem. Rev. 1948, 42, 189-
283. (b) For improved conditions, see: Friedman, L.; Shechter, H. J.
Org. Chem. 1961, 26, 2522-2524.
(4) Takagi, K.; Okamoto, T.; Sakakibara, Y.; Oka, S. Chem. Lett.
1973, 5, 471-4.
(5) Takagi, K.; Okamoto, T.; Yasumasa, S.; Ohno, A.; Oka, S.;
Hayama, N. Bull. Chem. Soc. Jpn. 1975, 48, 3298-3301.
(6) For reviews, see: (a) Sundermeier, M.; Zapf, A.; Beller, M. Eur.
J. Inorg. Chem. 2003, 3513-3526. (b) Takagi, K. In Handbook of
Organopalladium Chemistry for Organic Synthesis; Negishi, E., Ed.;
J. Wiley & Sons: Hoboken, NJ, 2002; Vol. 1, pp 657-672.
(14) First use of potassium ferrocyanide in this capacity (uncatalyzed
reaction): Merz, V.; Weith, W. Ber. 1877, 10, 746-765.
(15) K4[Fe(CN)6] (aka yellow prussiate of potash) is used as a food
additive in table salt to prevent “caking” and in wine production to
precipitate metals. The lower toxicity is attributed to the tight bond
between the iron center and the CN groups.
(16) (a) de Vries, A. H. M.; Mulders, J. M. C. A.; Mommers, J. H.
M.; Henderickx, H. J. W.; de Vries, J. G. Org. Lett. 2003, 5, 3285-
3288. (b) Reetz, M. T.; de Vries, J. G. Chem. Commun. 2004, 1559-
1563.
(17) de Vries, J. G.; de Vries, A. H. M. Eur. J. Org. Chem. 2003,
799-811.
(18) Urgaonkar, S.; Verkade, J. G. J. Org. Chem. 2004, 69, 5752-
5755.
(19) Beletskaya, I. P.; Cheprakov, A. V. Chem. Rev. 2000, 100, 3009-
3066.
10.1021/jo0481250 CCC: $30.25 © 2005 American Chemical Society
Published on Web 01/26/2005
1508
J. Org. Chem. 2005, 70, 1508-1510

TABLE 1. Ligand-free Cyanation of Aryl Bromidesa
FIGURE 1. Ligand-free cyanation of bromoketone 1.
a study into the potential for similar reactivity with the
aromatic cyanation reaction. Herein, we describe a practi-
cal, ligand-free cyanation of aryl bromides using homeo-
pathic amounts of palladium.20
We required a practical procedure to convert bromo-
ketone 1 into cyanoketone 2 (Figure 1). Application of
the Beller protocol (K4[Fe(CN)6]/Pd(OAc)2/dppf) in DMF
with sodium carbonate gave an 86% assay yield of 2
within 4 h at 120 °C (>99% conversion). A recent result
describing the use of 1,4-diazabicyclo-[2.2.2]octane (DAB-
CO) as a palladium ligand in the Suzuki coupling21 led
us to try a similar tactic in the aryl cyanation reaction.
By replacing the dppf with 2 mol % DABCO, in conjunc-
tion with 1 mol % Pd(OAc)2 and 22 mol % K4[Fe(CN)6]‚3
H2O in DMF, a 90% assay yield of nitrile 2 was obtained
within 4 h at 120 °C; no bis cyanide was observed. A
control reaction in DMF in the absence of DABCO
showed only 23% conversion after 5 h at 120 °C
This result suggests that in DMF, the DABCO acts as
a ligand to accelerate the reaction. The role of sodium
carbonate is essential, as only 3% conversion was ob-
served after 3 h at 120 °C in its absence. Subsequently,
the use of dimethylacetamide (DMAC) solvent was
preferred, as it led to more robust and faster reactions
(2 h). The fact that DMF is known to thermally degrade
to dimethylamine (DMA) could be a contributing factor,
as recent work from these laboratories has demon-
strated22 that DMA enhances cyanide solubility in the
case of ZnCN2. When a control reaction using 0.5 mol %
Pd(OAc)2 in DMAC was performed in the absence of the
DABCO (i.e., ligand-free), surprisingly an 87% assay yield
was obtained within 2 h.
This result led to a study into the scope and limitations
of this new aryl cyanation methodology.23 A series of aryl
halides were reacted under standard conditions,24 and
the results are displayed in Table 1. Rapid (e5 h) and
complete reactions were observed for activated aryl
bromides (i.e., electron-withdrawing group present) as
would be expected (entries 2-6). Even electron-neutral
systems (entries 7, 8, 11) and a mildly electron-rich arene,
p-bromotoluene (run 9), were completely reacted within
5 h with assay yields >90%. The cyanation of an
N-heteroaryl bromide (entry 10) gave an 86% assay yield
after 8 h at 120 °C. High selectivity was observed for
a Reaction conditions: 1.0 equiv of aryl bromide (4-6 mmols),
0.22 equiv of K4[Fe(CN)6], 1.0 equiv of Na2CO3, 0.1-0.5 mol %
Pd(OAc)2, DMAC (0.6 M), 120 °C. b Assay yield determined by LC
vs authentic standard. c Performed with 0.18 equiv of K4[Fe(CN)6].
d Starting material was 3-bromobenzonitrile.
3-bromochlorobenzene (entry 4). In this example, the
K4[Fe(CN)6] charge was reduced to 0.18 equiv (1.07 equiv
CN). After 1.5 h at 120 °C, the reaction showed >98%
conversion and a 22:1 A% ratio of mono:bis cyanation.
The trace amount of cyanation of the aryl chloride led
us to investigate extending this methodology to aryl
chlorides.
When 0.44 equiv of K4[Fe(CN)6] was used with 3-bro-
mochlorobenzene, the reaction stalled at a 63:37 ratio
(bis:mono cyanation) after 4 h at 120 °C.25 This result
suggested that aryl chlorides are partially reacted under
these conditions. Similarly, the cyanation of 3-chloro-
benzonitrile stalled after 51% conversion. Lower yields
and/or incomplete reactions were observed with 4-bromo-
(20) Term “homeopathic quantity” refers to low loadings of ligand-
free palladium (generally 0.1 mol % or lower). See ref 18.
(21) Li, J.-H.; Liu, W.-J. Org. Lett. 2004, 6, 2809-2811. Reetz, M.
T.; Westermann, E.; Lohmer, R.; Lohmer, G. Tetrahedron Lett. 1998,
39, 8449-8452.
(22) Marcantonio, K. M.; Frey, L. F.; Liu, Y.; Chen, Y.; Strine, J.;
Phenix, B.; Wallace, D.; Chen, C. Org. Lett. 2004, 6, 3723-3725.
(23) Sodium ferrocyanide (Na4[Fe(CN)6]‚10 H2O), also known as
yellow prussiate of soda, also can be used as the cyanide source with
equal performance.
(24) Standard conditions: 0.1-0.5 mol
%
Pd(OAc)2/22 mol
%
(25) Increasing the temperature to 130 °C had no effect on further
conversion.
K4[Fe(CN)6]/1.0 equiv Na2CO3 in DMAC (0.6-0.7 M) at 120 °C.
J. Org. Chem, Vol. 70, No. 4, 2005 1509

benzaldehyde (61% assay yield), 4-bromoanisole (11%
conversion), and 3-bromothiophene.
Experimental Section
General Procedure. A 25 mL flask was charged with the
aryl bromide (6 mmols), DMAC (10 mL), K4[Fe(CN)6]‚3 H2O (557
mg; 1.32 mmols; 0.22 equiv), sodium carbonate (636 mg; 6
mmols; 1.0 equiv), and Pd(OAc)2 (0.1-0.5 mol %; either as a solid
or as a 1 mg/mL solution in DMAC). The flask was evacuated
and filled with nitrogen (two times) and heated to 120 °C.
Reaction conversion was monitored by HPLC. Upon completion,
the reaction mixture was cooled to rt and diluted with 20 mL of
EtOAc. The resulting slurry was filtered and the filtrate assayed
for content. The product can be isolated by washing the filtrate
with water (2 × 15 mL) and 5% NH4OH (1 × 15 mL). The organic
layer was dried over Na2SO4, and the volatiles were removed in
vacuo to provide the product.
The cyanation reactions proceeded efficiently with as
little as 0.1 mol % catalyst, corresponding to a catalyst
turnover number (TON) approaching 1000. An attempt
to reduce the charge to 0.01 mol % with 4-trifluoromethyl-
bromobenzene stalled at 71% conversion (TON ) 7100).
The reaction temperature can be reduced to 100 °C, as
evidenced by the >99% conversion attained within 3 h
for 4-trifluoromethyl-bromobenzene (entry 6) with 0.1 mol
% Pd(OAc)2. Elimination of the base gave minimal
reaction (3% conversion) and suggests that the base
might play a role in reducing the Pd(II)(OAc)2 to the
active Pd(0) species.26 Inherently, there is no need for
base, as several aryl cyanation methodologies use none
at all. The reaction does not proceed in the absence of
the palladium.27
We have developed a practical, ligand-free aryl cya-
nation methodology utilizing inexpensive, easy-to handle,
and nontoxic reagents. This procedure gives high yields
for a respectable variety of aryl bromides and is amenable
to large-scale work, as the reactions are rapid and require
only a modest catalyst charge. This result adds to the
growing list of metal-catalyzed reactions that can be
performed ligand-free.
Acknowledgment. The authors thank Mr. Pete
Dormer for NMR assistance, Mr. Alex Chen for particle
size analysis, and Ms. Xiujuan Jia for palladium analy-
sis.
Note Added in Revision. During the review process,
a paper appeared in print by Professor Beller and co-
workers on this topic (J. Organomet. Chem. 2004, 689
(24), 4576-4583).
Supporting Information Available: Experimental de-
tails and characterization of 2. This material is available free
of charge via the Internet at http://pubs.acs.org.
(26) Thermal decomposition of Pd(OAc)2 to Pd(0) in propylene
carbonate at 100 °C has been observed and might be a factor in this
system as well. See: Reetz, M.; Lohmer, G. Chem. Commun. 1996,
1921-1922.
(27) Use of 0.5 mol % Pd2dba3 also provides a nearly equivalent
outcome for bromoketone 1.
JO0481250
1510 J. Org. Chem., Vol. 70, No. 4, 2005