Cooperative catalyst effects in palladium-mediated cyanation reactions of aryl halides and triflates

Article abstract of DOI:10.1021/jo9808674

Substoichiometric quantities of copper or zinc species dramatically improve both conversion rate and efficiency of Pd(0)-catalyzed cyanation reactions. The optimum reaction conditions involve the use of a nitrile solvent, NaCN, and a catalyst system employing the combination of cuprous iodide with tetrakis(triphenylphosphine)palladium(0), [Pd(PPh₃)₄]. Beneficial effects were observed for the conversion of aryl halides, aryl triflates, and a vinyl bromide to the corresponding nitriles. The process was demonstrated on preparative scale with a broad range of aromatic and heteroaromatic substrates containing diverse functionality. A dual catalytic cycle is proposed to explain the profound influences of the cocatalyst system.

Full text of DOI:10.1021/jo9808674

8224
J . Org. Chem. 1998, 63, 8224-8228
Coop er a tive Ca ta lyst Effects in P a lla d iu m -Med ia ted Cya n a tion
Rea ction s of Ar yl Ha lid es a n d Tr ifla tes
Benjamin A. Anderson,* Edward C. Bell,1 Francis O. Ginah, Nancy K. Harn,
Lisa M. Pagh, and J ames P. Wepsiec
Lilly Research Laboratories, A Division of Eli Lilly and Company,
Chemical Process Research and Development Division, Indianapolis, Indiana 46285
Received May 7, 1998
Substoichiometric quantities of copper or zinc species dramatically improve both conversion rate
and efficiency of Pd(0)-catalyzed cyanation reactions. The optimum reaction conditions involve
the use of a nitrile solvent, NaCN, and a catalyst system employing the combination of cuprous
iodide with tetrakis(triphenylphosphine)palladium(0), [Pd(PPh3)4]. Beneficial effects were observed
for the conversion of aryl halides, aryl triflates, and a vinyl bromide to the corresponding nitriles.
The process was demonstrated on preparative scale with a broad range of aromatic and
heteroaromatic substrates containing diverse functionality. A dual catalytic cycle is proposed to
explain the profound influences of the cocatalyst system.
In tr od u ction
have proven beneficial in the specific examples for which
they were developed, an expedient procedure which yields
the palladium-catalyzed cyanation reaction generally
applicable to a wide variety of substrates has not been
demonstrated. In this paper, we report a solution to this
troublesome but important synthetic problem.
The palladium-catalyzed displacement of aryl halides
and triflates with cyanide ion to afford the corresponding
aryl nitrile has been reported.2,3 The transition-metal-
catalyzed process offers practical improvements to the
Rosemund-von Braun reaction for which the same
overall transformation necessitates both stoichiometric
quantities of copper cyanide and high reaction temper-
atures. Product isolations are also complicated by sepa-
ration of the resulting copper halide.4 Tetrakis(triphen-
ylphosphine)palladium(0) [Pd(PPh3)4], a reactive and
readily available Pd(0) source,5 has been used most
extensively in the catalyzed reaction. However, the Pd-
(PPh3)4-mediated process is recognized as being notori-
ously unreliable, thus prompting reports of several modi-
fications to the original protocol. These variations are
categorized generally by the application of either modified
ligand sets,2f,6 the use of covalent5,7,8 and alumina im-
pregnated cyanide sources,9 or the inclusion of crown
ether additives.10 Additionally, conducting the reaction
in the presence of electrochemical reducing potential has
been shown to minimize the persistent problem of
catalyst deactivation.11 Although these modifications
Our synthesis of a tricyclic partial ergot alkaloid drug
candidate required the preparation of the aryl nitrile
derivative 2 from 1a which was accomplished through
the use of a transition metal-catalyzed process (eq 1).12
Standard Pd(PPh3)4-catalyzed cyanation protocols involv-
ing KCN and aryl iodide 1a in THF at reflux were
ineffective.2b Ultimately, our discovery that the Pd(0)-
mediated reaction could be dramatically accelerated by
addition of a catalytic amount of copper(I) iodide allowed
efficient production of 2.12 We detail herein a systematic
study of this reaction and apply the new and experimen-
tally expedient method to a wide range of structurally
diverse substrates. The convenient process affords a
dependable catalyst system which allows the facile
conversion of aryl iodides, bromides and triflates to the
corresponding nitrile derivatives.
Resu lts a n d Discu ssion
(1) Eli Lilly Summer Research Intern, Tougaloo College, J ackson,
MS.
Cyanation of iodide 1a to 2 required the presence of
both Pd(PPh3)4 and CuI. The cocatalyzed cyanation
reaction of 1a was complete within 2.5 h in deoxygenated
THF at reflux (Table 1). Without CuI, similar conditions
provided only 10% conversion to 2 after 18 h. The use of
catalytic CuI without a Pd(0) catalyst did not generate
any detectable amounts of aryl nitrile.13
Evaluation of several reaction parameters furnished
an improved protocol. The relative reaction rates for the
Pd/Cu-catalyzed cyanation of 1a conducted in varied
(2) (a) Takagi, K.; Okamoto, T.; Sakakibara, Y.; Oka, S. Chem. Lett.
1973, 471. (b) Sekiya, A.; Ishikawa, N. Chem. Lett. 1975, 277. (c)
Takagi, K.; Okamoto. T.; Sakakibara, Y.; Ohno, A.; Oka, S.; Hayama,
N. Bull. Chem. Soc. J pn. 1975, 48, 3298. (d) Takagi, K.; Okamoto. T.;
Sakakibara, Y.; Ohno, A.; Oka, S.; Hayama, N. Bull. Chem. Soc. J pn.
1976, 49, 3177. (e) Akita, Y.; Shimazaki, M.; Ohta, A. Synthesis 1981,
974. (f) Takagi, K.; Sasaki, K.; Sakakibara, Y. Bull. Chem. Soc. J pn.
1991, 64, 1118.
(3) Farina, V. In Comprehensive Organometallic Chemistry II; Abel,
E. W., Stone, F. G. A., Wilkinson, G., Ed.; Pergamon: Oxford, UK,
1995; pp 225-226.
(4) Ellis, G. P.; Ronmey-Alexander, T. M. Chem. Rev. 1987, 87,
779.
(5) Amatore, C.; J utand, A.; Khalil, F.; Mohamed, A. M.; Mottier,
L. Organometallics 1993, 12, 3168.
(6) Gunderson, L. L. Acta Chem. Scand. 1996, 50, 58.
(7) (a) Tschaen, D. M.; Desmond, R.; King, A. O.; Fortin, M. C.; Pipik,
B.; King, S.; Verhoeven, T. R. Synth. Commun. 1994, 24, 887. (b)
Selnick, H. G.; Smith, G. R.; Tebben, A. J . Synth. Commun. 1995, 25,
3255.
(10) (a) Yamamura, K.; Muranhashi, S. Tetrahedron Lett. 1977, 18,
4429. (b) Sato, N.; Suzuki, M. J . Heterocycl. Chem. 1987, 24, 1371. (c)
Piers, E.; Fleming, F. F. Can. J . Chem. 1993, 71, 1867.
(11) Davison, J . B.; Peerce-Landers, P. J .; J asinski, R. J . J . Elec-
trochem. Soc. 1983, 130, 1862.
(12) Anderson, B. A.; Becke, L. M.; Booher, R. N.; Flaugh, M. E.;
Harn, N. K.; Kress, T. J .; Varie, D. L.; Wepsiec, J . P. J . Org. Chem.
1997, 611, 8634.
(8) Chatani, N.; Hanafusa, T. J . Org. Chem. 1986, 51, 4714.
(9) Dalton, J . R.; Regen, S. R. J . Org. Chem. 1979, 44, 4443.
10.1021/jo9808674 CCC: $15.00 © 1998 American Chemical Society
Published on Web 10/21/1998

Palladium-Mediated Cyanation Reactions
J . Org. Chem., Vol. 63, No. 23, 1998 8225
Ta ble 1. P d (P P h 3)4-Ca ta lyzed Cya n a tion of 1a
Ta ble 2. Ad d itive Effects on th e P d (P P h 3)4-Ca ta lyzed
Cya n a tion of 3a a
entry
addend
time, h
%, convb
entry substrate solvent
M
mol % CuI time, h %, convb
1
2
3
4
5
6
7
8
9
10
11
12
13
14
none
0
1
1
2
2
4
3
3
6
1
2.5
1
3.5
2
7
100
100
100
100
94
1
2
3
4
5
6
7
8
9
1a
1a
1a
1a
1a
1a
1a
1a
1a
1a
1a
1b
THF
THF
THF
THF
THF
MeCN
MeCN Na
EtOAc
EtOAc Na
DMF Na
MeOH Na
EtCN Na
K
K
K
Na
Bu4N
K
0
10
10c
10
10
10
10
10
10
10
10
10
18
3
3
0.5
72
1.5
0.5
4
0.5
24
19
4
10
100
100
100
3
99
99
96
10 mol % CuI
10 mol %CuIc
5 mol % CuI
10 mol % CuCN
10 mol % CuOTf
10 mol % Cu(OTf)2
10 mol % CuBr
10 mol % CuBr2
10 mol % CuCl
10 mol % ZnI2
10 mol % KI
57
100
100
4
100
1
K
98
10
11
12
25d
3d
99d
10 mol % TMSCN
10 mol % B(Ph)3
>98
100
a
Conditions: 2 equiv of MCN, 5 mol% Pd(PPh3)4, 0.4 M 1 in
b
deoxygenated solvent under N2, 65 °C. Conversion determined
a
Conditions: 2 equiv of KCN, 5 mol % Pd(PPh3)4, 0.4 M 3a in
by HPLC. c 10 mol
reflux.
% PPh3 added. Reaction conducted at
d
b
THF. Conversion determined by HPLC. c 10 mol % PPh3 added.
Ta ble 3. P d (0)-Ca ta lyzed Cya n a tion Rea ction s of Ar yl
Br om id e 3ba
media at a constant temperature (65 °C) increased in the
order: EtOAc < THF < MeCN (Table 1, entries 2, 6, and
8). Both DMF and methanol were poor solvents for the
reaction and provided only modest conversion to the
desired product (entries 10, 11). Deoxygenating solvents
by subsurface nitrogen sparge enhanced reproducibility
and conversion yields.12,14 NaCN consistently afforded
more rapid conversions than KCN; however, both were
suitable reagents. Use of high quality, finely powdered
dry cyanide sources also improved reliability.15 Tetra-
n-butylammonium cyanide (Bu4NCN) was ineffective and
led to only 3% conversion of 1a to 2 after 72 h (entry 5).
The optimal reaction conditions involve the combination
of a nitrile solvent with NaCN.
entry
1
Pd source
mol %, CuI solvent
time, h
7
% convb
61c
Pd2(dba)3
4 DPPF
Pd2(dba)3
4 DPPF
Pd2(dba)3
4 DPPF
Pd2(dba)3
4 DPPF
0
10
0
NMP
NMP
EtCN
EtCN
2
3
4
7
20
6
69c
22
10
97
5
6
7
Pd(PPh3)4
Pd(PPh3)4
Pd(PPh3)4
0
10
10
EtCN
EtCN
MeCN
8
4
7
12
97
97
a
Conditions: 5 mol % Pd, 2 equiv of NaCN, 1 M solution of 3b
at reflux temperature. Conversion determined by HPLC. c 60 °C.
b
A study of reactions involving naphthalene derivatives
3 established the requisite characteristics of the catalyst
system (eq 2, Table 2). As observed for 1, the cyanation
of 3a employing KCN (2 equiv) in deoxygenated THF and
Pd(PPh3)4 (5 mol %) provided negligible production of 4
(7%, 6 h). Conversely, the presence of copper salts
resulted in efficient conversion of 3 to 4 (entry 2). Cop-
per(I) iodide effected the most pronounced acceleration.
A 2:1 Cu:Pd ratio was most often employed. Although a
1:1 metal ratio was sufficient to accomplish the desired
conversion, the reaction proceeded at a proportionately
slower rate (entry 4). CuI, CuOTf‚0.5 C6 H6 and CuBr
were also competent addends. CuCl, however, did not
impact the Pd(0)-catalyzed reaction (entry 10).
Importantly, the accelerating effect was not limited to
cyanation reactions catalyzed by Pd(PPh3)4 as a similar
influence was demonstrated for a Pd2(dba)3-mediated
process. [Pd2(dba)3]‚CHCl3 in the presence of chelating
phosphines (1,1′-bis(diphenylphosphino)ferrocene, DPPF)
in NMP has been reported to be a useful system for
catalysis of cyanide addition to aryl iodides and triflates.2f
However, like other palladium-catalyzed cyanation pro-
tocols, aryl bromides were shown to be far less reactive
substrates, and minimal conversion of bromobenzene to
benzonitrile was reported after extended reaction time
(12%, 12 h, 60 °C).2f While no beneficial effects of added
copper(I) iodide was observed for the cyanation reaction
of 3b conducted in NMP (Table 3, entries 1 and 2), CuI
(10 mol %) efficiently promoted the conversion of 3b in
The reaction proved sensitive to the oxidation state of
the copper species. Although Cu(OTf)2 (entry 7) and
CuBr2 (entry 9) were effective cocatalysts, reactions em-
ploying the corresponding copper(I) species provided more
rapid conversion.
(13) For a high-temperature copper-mediated cyanation reaction,
see: Carr, R. M.; Cable, K. M.; Wells, G. N.; Sutherland, D. R. J .
Labeled Compd. Radiopharm. 1994, 34, 887.
The accelerating effect was not limited to copper
cocatalysts. Zinc iodide (10 mol %) in THF also facilitated
the palladium-catalyzed cyanation of 3a (entry 11).16
Other transition metal addends, including MgBr2, Fe-
(CN)3, and Cr(CO)6, did not accelerate or improve the Pd-
(PPh3)4-catalyzed conversion of 3a to 4 under the condi-
tions studied.
(14) Andersson, Y.; Långstro¨m, B. J . Chem. Soc., Perkin Trans. 1
1994, 1395.
(15) Commercially available NaCN (99.40%) routinely provided
excellent results whereas a lower purity source (95%) has resulted in
poor conversion.
(16) A failed attempt to effect Pd(PPh3)4-catalyzed cyanation of an
aryl bromide employing KCN and catalytic Zn(CN)2 has been reported;
see ref 7a. We have found that DMF is a poor reaction medium for the
cocatalyzed process (Table 1).

8226 J . Org. Chem., Vol. 63, No. 23, 1998
Anderson et al.
Ta ble 4. Su bstitu en t Effects on P d (0)-Ca ta lyzed
Cya n a t1ion of 5
Ta ble 5. Cu I/P d (P P h 3)4-Ca ta lyzed Cya n a tion Rea ction s
of Ar yl a n d Vin yl Ha lid es a n d Tr ifla tesa
entry substrate
Y
addend
time %, conva
1
2
3
4
5
6
7
8
5a
5a
5b
5b
5c
5c
5d
5d
OCH3
10 mol % CuI
none
10 mol % CuI
none
4
4
1
5
1
4
5
1
97
14
H
100b
8b
COCH3 10 mol % CuI
none
NO2
100
47
48
10 mol % CuI
none
98
a
b
Conversion determined by GC. Conversion determined by
HPLC.
propionitrile with [Pd2(dba)3]‚CHCl3-(4 DPPF) (5 mol %)
(entry 4). Complete conversion of the bromide 3b to 4
was observed after 6 h in propionitrile while only 22%
conversion occurred after 20 h in the absence of the CuI
additive.
Consistent with kinetic studies which have docu-
mented the relative reactivity of the two Pd(0) catalysts,5
the Pd(PPh3)4 and CuI-catalyzed reaction proved more
rapid than the analogous Pd2(dba)3-mediated process, and
conversion of 3b was complete within 4 h (Table 3, entry
6). The corresponding reaction conducted at lower tem-
perature in acetonitrile at reflux provided 97% conversion
after 7 h (Table 3, entry 7). Furthermore, the cyanation
reaction of aryl bromide 1b in EtCN was also efficiently
accomplished within 4 h (Table 1, entry 12). Although
several favorable catalyst systems were identified, the
application of CuI and Pd(PPh3)4 provided the optimal
catalyst combination studied.
Application of the new methodology to a broad range
of substrates was an important focus of this study. A
preliminary evaluation of electronic effects confirmed
that the cocatalyst system positively influenced the
conversion for a variety of simple aromatic iodides 5 to
the corresponding nitriles 6 (eq 3). Copper-induced accel-
eration was most pronounced for the electron-rich sub-
strates (Table 4). The impact was moderated for electron-
poor substrates which are typically more reactive sub-
strates for the palladium-catalyzed processes. For exam-
ple, 5d was an exceptionally reactive substrate and com-
plete conversion to 6 in the absence of CuI was accomp-
lished within 1 h. Interestingly, conversion in the pres-
ence of 10 mol % CuI stalled after only 48% conversion.
Evaluation of the process employing substrates con-
taining diverse functionality was conducted in a series
of preparative experiments (Table 5). In addition to
simple substituted aryl halides, several iodo- and bromo-
substituted heterocycles were successfully converted to
the corresponding nitriles. The facile cyanation method
illustrated in Tables 3 and 5 for the conversion of aryl
bromides has particular practical significance since the
substrates are often more readily available than the
corresponding aryl iodides.
a
Conditions: 2 equiv of NaCN, 5 mol % Pd(PPh3)4, 10 mol %
CuI, nitrile solvent at reflux, MeCN (ArI, ArOTf); EtCN (ArBr).
Isolated yield. c Reaction conducted in valeronitrile, 115 °C. 10
b
d
mol % Pd(PPh3)4, 20 mol % CuI, MeCN, 50 °C.
process proved broad and accommodated a wide range
of substrates including species containing Lewis basic
functionality.
Modification of the standard protocol was required in
some special cases. Only trace levels of conversion of
5-bromoindole were observed when the reaction was
conducted in propionitrile (97 °C, 16 h). The use of
valeronitrile as solvent allowed a higher reaction tem-
perature (110 °C) and resulted in the production of useful
yields of the desired nitrile (entry 8). The successful
cyanation of a vinyl halide was also enabled by the
addition of CuI (entry 9). Addition of up to 50 mol %
Pd(PPh3)4 failed to efficiently catalyze the conversion of
this substrate in the absence of the cocatalyst (<20%
yield). The sensitivity of the vinyl bromide required a
reduction of the standard reaction temperature to 50 °C.
An increased catalyst loading of 10 mol % Pd(PPh3)4 and
20 mol % CuI offset the rate-retarding effect of the lower
reaction temperature.
Analogous to the corresponding halide derivatives 3a
and 3b, the Pd(PPh3)4-catalyzed conversion of aryl triflate
3c to 4 failed to go to completion. The standard Pd/Cu-
catalyzed protocol established above provided complete
conversion within 1 h. In general, the scope of the
The operative mechanism of the bimetallic catalyst sys-
tem has not been established. However, several observa-
tions provide insight. Combinations of catalytic quanti-
ties of Pd(PPh3)4 and KI (Table 2, entry 12) or LiI (not
shown) failed to promote the cyanation reaction, and only

Palladium-Mediated Cyanation Reactions
J . Org. Chem., Vol. 63, No. 23, 1998 8227
Sch em e 1
trace conversion of either 3a or 3b was observed after 6
h. These results discount the possibility that the metal
addends serve as halide sources which might in turn be
responsible for the observed effects.2c,17
Alternatively, the proposed M(CN)n intermediate might
facilitate the reaction by providing a species which func-
tions more efficiently in the transmetalation step with
the Pd(II) complex than does either NaCN or KCN.22
A
Interestingly, the addition of PPh3 had minimal effect
on the rate of the copper/palladium-catalyzed cyanation
reaction. Conducting the cocatalyzed cyanation reaction
of 3a in the presence of up to 40 mol % triphenylphos-
phine afforded complete conversion to the nitrile at a rate
nearly indistinguishable from that achieved with the
bimetallic catalyst system. Similarly, the CuI-promoted
cyanation reaction of 1a remained operational when
conducted in the presence of 10 mol % PPh3 and catalytic
Pd(PPh3)4 (entries 2 and 3, Table 1). In contrast, addition
of PPh3 completely suppresses the copper-accelerating
effect associated with the palladium-catalyzed Stille
cross-coupling reaction.18,19
A two-cycle process is proposed to explain the role of
the addend.20 The transition metal additive might serve
as a vehicle which transfers the cyanide ion between the
poorly soluble stoichiometric source (KCN or NaCN) and
the Pd(II) intermediate (Scheme 1). This second cycle
(II) could act in concert with the primary process which
involves activation of the substrate through the Pd(0)-
mediated oxidative addition reaction followed by trans-
metalation and reductive elimination.
reaction was initiated in the absence of a copper cocata-
lyst in which 3a was treated with 2 equiv KCN and 5
mol % Pd(PPh3)4 (THF). As expected, production of the
nitrile peaked at only 10% conversion in 1 h. Addition
of 10 mol % CuI to this reaction mixture led to 89%
conversion after 3.5 h. The rate at which 4 was produced
following addition of the cocatalyst was much slower
when compared to reactions in which CuI was present
at the outset (complete, 1 h). This suggests that some
degree of catalyst poisoning had occurred, and although
facilitating the transmetalation step may be a contribut-
ing factor, it is not the exclusive role of CuI.
While the exact nature of the proposed catalytically
active metal cyanide species has not been determined,
the feasibility of such an entity to be involved in the
transmetalation step was demonstrated. Both CuCN and
KCu(CN)2 (1 equiv) proved to be efficient agents for the
cyanation of 1a in the presence of Pd(0) (5 mol %). High
yields of the cyanoindoline 2 were obtained after reaction
in THF at reflux with rates similar to the CuI-catalyzed
process (3-4 h). The Pd(0) catalyst was required in both
of these examples to produce 2.
Two pathways in which the proposed M(CN)n inter-
mediate might influence the reaction are considered.
High concentration of alkali metal cyanide (KCN or
NaCN) is known to inhibit the palladium-catalyzed
cyanation reaction.2c,d Catalyst inhibition might be mini-
mized by conversion of cyanide ion to the more covalent
transition metal species (i.e. M(CN)n). Limited solubility
of KCN and NaCN allows application of only a catalytic
amount of added copper or zinc salts.21 Reaction condi-
tions which allow ionic cyanide concentrations to over-
whelm the intercepting catalytic addend are therefore
ineffective. For example, no benefit of a cocatalyst was
realized when a soluble cyanide species (e.g. n-Bu4NCN)
was employed (Table 1, entry 5) or when the reaction
medium significantly improved the solubility of the alkali
cyanide (Table 1, entries 10, 11).
The proposed mechanism requires that an effective
cocatalyst must have the capacity to react with NaCN
or KCN to form a covalently bound cyanide species which
can be regenerated upon transmetalation with the Pd-
(II) intermediate. Additive candidates are therefore not
limited to transition metals. This analysis prompted our
evaluation of trimethylsilyl cyanide (TMSCN)23,24 and
triphenylborane as cocatalysts.25 Indeed, the Pd(Ph3)4-
catalyzed conversion of 3a to 4 was accomplished in 3.5
h when conducted in the presence of 10 mol % TMSCN
(Table 2, entry 13) and within 2 h with added B(Ph)3
(Table 2, entry 14). These results are consistent with
(21) The solubilities of sodium and potassium cyanide in several
solvents at 60 °C are: DMF, 15.0 mmol dm-3; THF, 2.4 mmol dm-3
;
acetonitrile, 1.9 mmol dm-3. Sakakibara, Y.; Ido, Y.; Sasaki, K,; Sakai,
M.; Uchino, N. Bull. Chem. Soc. J pn. 1993, 66, 2776.
(22) L2Pd(CN)2 complexes have been prepared from NaCN and the
corresponding L2PdCl2 species, see Vertuyft, A. W.; Cary, L. W.; Nelson,
J . H. Inorg. Chem. 1976, 15, 3161.
(23) Trimethylsilyl cyanide is an effective stoichiometric cyanide
source for Pd(0)-catalyzed processes when the reaction is conducted
in triethylamine as solvent; see ref 8.
(17) Forster, D. J . Am. Chem. Soc. 1975, 97, 951.
(18) Liebeskind, L. S.; Fengl, R. W. J . Org. Chem. 1990, 55,
5359.
(19) Farina, V.; Kapadia, S.; Krishnan, B.; Wang, C.; Liebeskind,
L. J . Org. Chem. 1994, 59, 5905.
(20) A related mechanism was proposed to interpret kinetic data
derived from Pd(II)-catalyzed cyanation reactions; see ref 2d.
(24) For leading references describing the conversion of TMS-X to
TMS-CN, see Sukata, K. Bull. Chem. Soc. J pn. 1987, 60, 2257.

8228 J . Org. Chem., Vol. 63, No. 23, 1998
Anderson et al.
8-Qu in olin ecar bon itr ile (9). 8-Quinolinyltrifluoromethane-
sulfonate (0.50 g, 1.8 mmol), CuI (0.034 g, 0.18 mmol), NaCN
(0.18 g, 3.6 mmol), and Pd(PPh3)4(0.10 g, 0.087 mmol) were
charged to a round-bottom flask. Acetonitrile (5 mL) was
added, and the mixture was heated to reflux for 1 h. Reaction
workup was accomplished employing the standard method.
The product was purified by silica gel chromatography (30%
ethyl acetate in hexanes) to give 0.24 g of the desired product
(87%). Mp 87.5-88.3 °C; TLC Rf ) 0.3 (30% ethyl acetate in
hexanes); IR (CHCl3) 2232 cm-1; 1H NMR (CDCl3) δ 8.94 (dd,
1H, J ) 1.2, 4.2 Hz), 8.15 (dd, 1H, J ) 1.0, 8.5 Hz), 8.0-7.96
(m, 2H), 7.52-7.42 (m, 2H); 13C NMR (CDCl3) δ 152.0, 146.9,
136.2, 135.1, 132.6, 127.6, 125.5, 122.4, 116.9, 112.5; mass
spectrum, m/z (FIA, M + 1) 155. Anal. Calcd for C10H6N2:
C, 77.91; H, 3.92; N, 18.17. Found: C, 77.94; H, 4.08; N, 18.06.
8-Cyan o-4,10b-dim eth yl-1,2,3,4,4a,5,6,10b-octah ydr oben -
zo[f]qu in olin -3-on e (10). The aryl bromide31 (1.07 g, 3.47
mmol), CuI (0.07 g, 0.4 mmol), NaCN (0.36 g, 7.3 mmol), and
Pd(PPh3)4 (0.21 g, 0.18 mmol) were charged to a round-bottom
flask. Propionitrile (5 mL) was added, and the mixture was
heated to reflux for 2h. Reaction workup was accomplished
employing the standard method. The product was purified by
silica gel chromatography (ethyl acetate) to give 0.73 g of the
desired product (83%). Mp 146-147.4 °C; TLC Rf ) 0.38 (ethyl
acetate); IR (CHCl3) 2231, 1630 cm-1; 1H NMR (CDCl3) δ 7.33-
7.28 (m, 3H), 3.40 (dd, 1H, J ) 3, 13 Hz), 2.95-2.88 (m, 5H),
2.53-2.48 (m, 2H), 2.29-2.22 (m, 2H), 1.81-1.71 (m, 2H), 1.04
(s, 3H); 13C NMR (CDCl3) δ 170.4, 149.4, 135.3, 132.7, 129.7,
126.0, 118.8, 110.2, 61.1, 37.8, 32.0, 29.5, 29.1, 27.9, 21.9; mass
spectrum, m/z (FIA, M + 1) 255. Anal. Calcd for C16H18N2 O:
C, 75.56; H, 7.13; N 11.01. Found: C, 75.89; H, 7.21; N, 10.97.
(Z)-6-[2-Cya n o-1-(3-flu or op h en yl)eth en yl]-1-[(1-m eth -
yleth yl)su lfon yl]-1H-ben zim id a zol-2-a m in e (13). The cor-
responding vinyl bromide32 (2.00 g, 46 mmole), CuI (0.16 g,
0.84 mmol), KCN (0.60 g, 9.2 mmol), and Pd(PPh3)4 (0.52 g,
0.45 mmol) were charged into the reaction flask. Acetonitrile
(40 mL) was added, and the resulting mixture was heated to
50 °C for 90 min. The reaction mixture was cooled to room
temperature. The mixture was filtered through a plug of silica
gel, and the solid cake was washed with ethyl acetate followed
by a wash with methylene chloride until the eluting solution
was colorless. The filtrate was concentrated to a solid. The
solid was stirred in ethyl acetate and washed with water
followed by a wash with brine. The organic phase was dried
over anhydrous Na2SO4 and filtered. The filtrate was con-
centrated to 1.5 g of an amorphous white solid (85% yield). 1H
NMR (DMSO) δ 7.9 (s, 1H), 7.7-7.0 (m, 6H), 5.9 (s, 2H), 5.5
(s, 1H), 3.8 (m, 1H), 1.4 (d, 6H); IR (CHCl3) 2215, 1640, 1608
cm-1, MS m/z (FID) 384. Anal. Calcd for C19H17N4O2SF: C,
59.36; H, 4.46; N, 14.57. Found: C 59.41; H 4.57; N 14.03.
the proposed mechanism albeit the reactions were mark-
edly slower when compared to the more optimal Pd-
(PPh3)4/CuI catalyst system.
Although these experiments establish the proposed
mechanism to have predictive value, other possibilities
remain viable. For example, Lewis acids are known to
accelerate the nickel-catalyzed diene hydrocyanation
reaction.26 If operative in the present case, a Lewis acid
effect is not general nor does it correlate with the acidity
of the addends. For example, copper (I) iodide is a more
effective adjutant than copper(II) triflate. Moreover, the
presence of other Lewis acids such as MgBr2 does not
afford appreciable acceleration of the Pd(0)-catalyzed
conversion of 3a to 4 (5% conv, 3 h, THF).
Con clu sion
A novel cocatalyzed cyanation reaction involving Pd-
(0) catalysts has been developed that provides efficient
and reproducible access to aryl nitriles. The process,
studied in greatest detail for a CuI/Pd(PPh3)4 catalyst
system, appears to be mechanistically distinct from the
copper effect involved in the palladium-catalyzed cross-
coupling reaction.18 The new process allows efficient con-
version of aryl and vinyl halides as well as aryl triflates
to the corresponding nitrile derivatives utilizing potas-
sium or sodium cyanide. Palladium-catalyzed cyanation
reactions were demonstrated to be uniformly sluggish or
to stall completely in the absence of the cocatalyst.
Several different copper sources and zinc(II) iodide were
shown to accelerate the palladium(0)-catalyzed reaction.
The copper cocatalyst extended the effectiveness of both
Pd(PPh3)4 and [Pd2(dba)3]‚CHCl3-DPPF mediated pro-
cesses. This convenient and reliable protocol eliminates
the harsh reaction conditions and excessive metal waste
which are characteristic of previous existing methods.
Exp er im en ta l Section
All solvents were deoxygenated by a subsurface N2 purge
for 30-45 min immediately prior to use. High-pressure liquid
chromatography (HPLC) was performed using a Zorbax RX-
C8 column (4.6 mm i.d. × 25 cm) with diode array UV
detection. Similar absorption coefficients (>95%) were estab-
lished for the starting material and product at the monitored
wavelength. Gas chromatography (GC) was performed using
a Fisher Scientific DB-1 capillary column (30 m; film thickness
0.25 mm). Products were compared to authentic samples ob-
tained commercially for compounds 4, 6a -c, 11, and 12. Spec-
tral data for 727 and 828 were compared with reported values.
Pd(PPh3)4 was prepared as previously described.29
Commercially available CuI (Aldrich, 99.999%) was typically
purified;30 however, experiments in which the reagent was
used directly provided similar results. All other catalysts were
purchased and used without further purification. NaCN
(Mallinckrodt, 99.40%) and KCN (Fisher, g98%) were finely
ground and stored in a vacuum oven (100-150 °C) prior to
use.
Ack n ow led gm en t. The authors thank Dr. Leland
Weigel and Mr. David Varie for helpful discussions and
the supply of selected substrates employed in this study.
Mr. J oseph Pawlak is also gratefully acknowledged for
helpful discussions.
J O9808674
(25) The adduct of KCN with B(Ph)3 is well-known, see Brehm, E.;
Haag, A.; Hesse, G. Liebigs Ann. Chem. 1970, 737, 80.
(26) Kreutzer, K. A.; Tam, W. WO 9611182 A1.
(27) Randall, W. C.; Streeter, K. B.; Cresson, E. L.; Schwam, H.;
Michelson, S. R.; Anderson, P. S.; Cragoe, E. J ., J r.; Williams, H. W.
R.; Eichler, E.; Rooney, C. S. J . Med. Chem. 1979, 22, 608.
(28) Liu, Y.; Hong, Y..; Svensson, B. E.; Cortizo, L.; Lewander, T.;
Hacksell, U. J . Med. Chem. 1993, 36, 4221.
Rep r esen ta tive P r oced u r e. The substrate (3.94 mmol),
KCN (0.51 g, 7.88 mmol), Pd(PPh3)4 (0.228 g, 0.197 mmol), and
CuI (0.394 mmol) were placed in a flask which was flushed
with N2. Solvent (5 mL) was added via syringe. The resulting
mixture was heated to reflux under N2 with vigorous agitation
by a magnetic stirrer and sampled at 15 min intervals. The
mixture was cooled to room temperature, diluted with ethyl
acetate (30 mL), and then filtered through Celite. The filtrate
was washed with water and brine, dried over anhydrous
MgSO4, and concentrated by rotary evaporation. Purification
was accomplished by silica gel chromatography.
(29) Coulson, D. R. Inorg. Synth. 1972, 15,121.
(30) Kauffman, G. B.; Teter, L. A. Inorg. Synth. 1963, 7, 9.
(31) (a) Audia, J . E.; Neubauer, B. L. US 95-408745 950321; E. P.
733365 A2 960925. (b) Wiegel, L. O. Chiral 97, Boston, MA. May 13,
1997.
(32) Werner, J . A.; Dunlap, S. E.; Frank, S. A.; Dunigan, J . M.; Copp,
J . D.; Ginah, F. O. Presented at the 215th ACS National Meeting,
ORGN 239, Dallas, TX, March 29, 1998.