Catalytic isomerization of cyanoolefins involved in the adiponitrile process. C-CN bond cleavage and structure of the nickel π-allvl cyanide complex Ni(η3-1-Me-C3H4)(CN)(dppb)

Article abstract of DOI:10.1021/om049706i

Isomerization of 2-methyl-3-butenenitrile to 3-pentenenitrile is catalyzed by nickel π-allyl cyanide complexes. The mechanism is supported by in situ NMR monitoring and DFT studies.

Full text of DOI:10.1021/om049706i

© Copyright 2004
Volume 23, Number 14, July 5, 2004
American Chemical Society
Communications
Ca ta lytic Isom er iza tion of Cya n oolefin s In volved in th e
Ad ip on itr ile P r ocess. C-CN Bon d Clea va ge a n d
Str u ctu r e of th e Nick el π-Allyl Cya n id e Com p lex
Ni(η3-1-Me-C3H4)(CN)(d p p b)
Alexandra Chaumonnot,† Franck Lamy,† Sylviane Sabo-Etienne,*,†
Bruno Donnadieu,† Bruno Chaudret,*,† J ean-Claude Barthelat,‡ and
J ean-Christophe Galland§
Laboratoire de Chimie de Coordination, CNRS, 205 route de Narbonne,
31077 Toulouse Cedex 04, France, Laboratoire de Physique Quantique, IRSAMC,
Universite´ Paul Sabatier, 118 route de Narbonne, 31062 Toulouse Cedex 04, France, and
Rhodia Recherches, 85 rue des fre`res Perret, 69192 Saint Fons Cedex, France
Received April 23, 2004
Summary: Isomerization of 2-methyl-3-butenenitrile to
3-pentenenitrile is catalyzed by nickel π-allyl cyanide
complexes. The mechanism is supported by in situ NMR
monitoring and DFT studies.
typical batch experiment leads to a 1:1.5 ratio of 2M3BN
and 3PN.4 Herein we report an effective catalytic
isomerization of 2M3BN into 3PN using the zerovalent
Ni(COD)2 precursor in the absence of any Lewis acid.
The mechanism is established by in situ NMR catalytic
monitoring and isolation of the key π-allyl nickel cyanide
intermediate as well as by DFT theoretical studies.
The optimal catalytic experiment was carried out by
adding Ni(COD)2 (1 equiv, COD ) 1,5-cyclooctadiene)
to a mixture of 2M3BN (110 equiv) and PPh3 (10 equiv).
The mixture was stirred at 100 °C for 3 h. GC analysis
shows a 96% conversion and the formation of 3PN in
81% yield (see Scheme 1).
The synthesis of adiponitrile (AdN) based on a nickel-
catalyzed double hydrocyanation of butadiene is a major
industrial success for homogeneous catalysis. In the first
step, hydrocyanation leads to a mixture of the desired
3-pentenenitrile (3PN) and the undesired 2-methyl-3-
butenenitrile (2M3BN). The branched isomer needs to
be isomerized to the linear 3PN. The second hydrocya-
nation step produces AdN together with several byprod-
ucts. The reaction is assisted by Lewis acid cocatalysts.
Extensive mechanistic investigations, in particular by
the DuPont group, appeared almost 20 years ago, and
the role of a π-allyl nickel cyanide complex has been
outlined in the first hydrocyanation step.1-3
A similar experiment was performed in an NMR tube,
and the reaction was monitored directly in the NMR
probe at 100 °C using C7D8 as solvent. Introduction of
toluene reduces the conversion to 85% but allows a good
2M3BN production represents a major drawback of
the first catalytic hydrocyanation step. For example, a
(1) (a) Tolman, C. A.; McKinney, R. J .; Seidel, W. C.; Druliner, J .
D.; Stevens, W. R. Adv. Catal. 1985, 33, 1 and references therein. (b)
Tolman, C. A. J . Chem. Educ. 1986, 63, 199. (c) McKinney, R. J . In
Homogeneous Catalysis; Parshall, G. W., Ed.; Wiley: New York, 1992;
p 42.
(2) Backvall, J . E.; Andell, O. S. Organometallics 1986, 5, 2350.
(3) Druliner, J . D. Organometallics 1984, 3, 205.
* To whom correspondence should be addressed. E-mail: sabo@
lcc-toulouse.fr (S.S.-E.).
† LCC-CNRS, Toulouse.
‡ IRSAMC, Toulouse.
§ Rhodia, St Fons.
(4) Drinkard, W. C.; Lindsey, R. V., J r. U. S. Patent 3,496,215, 1970.
10.1021/om049706i CCC: $27.50 © 2004 American Chemical Society
Publication on Web 06/10/2004

3364 Organometallics, Vol. 23, No. 14, 2004
Communications
Sch em e 2
All our attempts to isolate 1 have failed, but using a
stabilizing diphosphine such as Ph2P(CH2)4PPh2 (dppb)
allowed us to isolate the analogous complex Ni(η3-1-Me-
C3H4)(CN)(dppb) (2) (see Scheme 2).5
The molecular structure is shown in Figure 2.6 The
structure can be described as a trigonal pyramid with
a methylallyl unit in the apical position and confirms
the C-CN bond breaking from the initial 2M3BN. The
Ni-CN distance of 1.893(5) Å is analogous to that
observed for other cyanide complexes7 but shorter than
F igu r e 1. 31P{1H} NMR spectra of a Ni(COD)2/2M3BN/
PPh3 mixture in a 1/16/8 ratio at various temperatures.
From the top to the bottom T ) 25, -60, -70, -80, and
-90 °C. The sharp signals present in very small amounts
in the region between 28 and 33 ppm are due to impurities.
i
in Ni(η3-C3H5)(CN)(dippe) (1.994(3) Å, dippe ) Pr2P-
(CH2)2PiPr2) very recently reported.8 It is worth noting
that a square-pyramidal geometry with a CN in an
apical position is observed in this latter compound.8
A catalytic test was also performed by using a Ni-
(COD)2/2M3BN/dppb mixture in a 1/110/1.3 ratio. GC
analysis shows a 96% conversion and the formation of
3PN in 90% yield (2-methyl-2-butenenitrile, 3%; 4-pen-
tenenitrile, 3%). A similar experiment was performed
using 2 as catalyst precursor, in the absence of any
added phosphine. Remarkably, a similar conversion
(97%) was obtained after heating (3 h at 100 °C) a
mixture of 2/2M3BN in a 1/110 ratio, and GC analysis
showed a 3PN selectivity of 83%.
On the basis of our DFT calculations (B3LYP) carried
out using PH3 as a model,9 we propose the mechanism
depicted in Scheme 3 for the isomerization of 2M3BN
into 3PN, involving the following steps: (i) coordination
of 2M3BN via the CdC bond; (ii) C-CN bond breaking
and formation of a σ-allyl species; (iii) isomerization to
the π-allyl species; (iv) σ-allyl formation; (v) C-CN bond
coupling to 3PN coordination. The B3LYP calculated
F igu r e 2. ORTEP drawing of compound 2. Selected bond
lengths (Å): Ni-C(100), 1.893(6); Ni-P(1), 2.2609(16); Ni-
P(2), 2.1937(16); Ni-C(1), 2.064(6); Ni-C(2), 1.970(6); Ni-
C(3), 2.130(6); C(1)-C(2), 1.374(8); C(2)-C(3), 1.359(9);
C(3)-C(4), 1.495(8); C(100)-N(100), 1.173(6). Selected bond
angles (deg): P(1)-Ni-P(2), 102.18(6); P(1)-Ni-C(100),
102.37(16); P(2)-Ni-C(100), 92.16(17).
Sch em e 1
(5) Selected NMR data are as follows. For 1 (from TOCSY, COSY
1H-1H, and HMQC 1H-13C experiments): 1H NMR (400 MHz, C7D8,
273 K) δ 4.63 (m, H2, 1H), 3.42 (qd, J H
H3, 1H), 2.25 (br, H1, 1H), 1.65 (d, J H
) 6.2 Hz, J H
3Me
3H2
) 12.5 Hz,
NMR analysis. The organic products are clearly identi-
fied by 1H NMR, whereas the 31P NMR spectrum shows
at 100 °C a broad signal at 4.3 ppm indicative of an
exchange. To better visualize the nickel species, we
reduced the initial quantities of reactants. A C7D8
mixture of Ni(COD)2/2M3BN/PPh3 in a 1/16/8 ratio was
heated for 30 min at 100 °C and then analyzed by NMR
at various temperatures. The 31P NMR spectrum shows
at room temperature a very broad signal at 2 ppm (see
Figure 1). Decoalescence leads at -90 °C to a sharp
signal at -4.6 ppm, assigned to free PPh3, and to an
AB signal at 32.6 and 23.9 ppm with a J P-P value of
103 Hz corresponding to the allyl cyanide complex
Ni(η3-1-Me-C3H4)(CN)(PPh3)2 (1) on the basis of several
1D and 2D NMR experiments.5
3Me
) 6.2 Hz, Me, 3H), 1.45 (br,
H1′, 1H); 13C NMR (100.6 MHz, C7D8, 273 K) δ 108.7 (C2), 86.5 (C3),
51.7 (C1), 19.6 (Me). For 2: 1H NMR (400 MHz, C7D8, 293 K) δ 4.72
(m, H2, 1H), 3.44 (br, H3, 1H), 2.24 (br, H1, 1H), 1.73 (brd, J H
) 6
3Me
Hz, Me, 3H), 1.25 (br, H1′, 1H); 31P NMR (161.97 MHz, C7D8, 293 K) δ
21 (br). extremely broad signals down to 183 K; 13C{1H} NMR (100.6
MHz, thf-d8, 263 K) δ 152.2 (CN), 96.2 (C2), 81.6 (C3), 44.8 (C1), 20.5
(Me). See Scheme 2 for atom labeling.
(6) Crystal data for 2‚C7H8: C40H43NP2Ni, Mr ) 658.40, triclinic,
space group P1h, T ) 160(2) K, a ) 10.095(2) Å, b ) 11.301(2) Å, c )
15.754(3) Å, R ) 80.07(3)°, â ) 86.08(3)°, γ ) 74.07(3)°, V ) 1701.9(6)
Å3, Z ) 2, µ ) 1.285 mm-1, 14 884/5457 reflections collected/unique,
R1 ) 0.0536, wR2 ) 0.1112, GOF ) 0.938.
(7) (a) Garcia, J . J .; J ones, W. D. Organometallics 2000, 19, 5544.
(b) Goertz, W.; Keim, W.; Vogt, D.; Englert, U.; Boele, M. D. K.; van
der Veen, L. A.; Kamer, P. C. J .; van Leeuwen, P. W. N. M. J . Chem.
Soc., Dalton Trans. 1998, 2981.
(8) Brunkan, N. M.; J ones, W. D. J . Organomet. Chem. 2003, 683,
77.
(9) For details of the calculations, see the Supporting Information.
Geometrical data (Å) for C: Ni-CN ) 1.892, Ni-C1 ) 2.044,
Ni-C2 ) 2.045, Ni-C3 ) 2.182, C1-C2 ) 1.428, C2-C3 ) 1.402,
C3-C4 ) 1.507. The anti isomer is found 15.5 kJ mol-1 above the syn
isomer. We have ruled out a heterolytic C-CN cleavage, as this process
leads to a cationic nickel species and CN- would require a very high
energy of 565 kJ mol-1. As suggested by a reviewer, a concerted process
involving C-CN bond cleavage and direct formation of C from A
without the intermediacy of the σ-species B (and D) cannot be excluded.
Further calculations are in progress.

Communications
Organometallics, Vol. 23, No. 14, 2004 3365
Sch em e 3. B3LYP -Op tim ized Str u ctu r es a n d
En er gies of Isom er s A-E In volved in th e
Isom er iza tion of 2M3BN to 3P N
such a conversion has been very recently evidenced in
related nickel species.8,10,11 An alternative mechanism
involving the intermediacy of a hydrido nickel cyanide
complex and elimination of butadiene has been ruled
out, being too high in energy. Such a mechanism has
been proposed to operate on the basis of isotope ex-
change reactions in the presence of Lewis acid and with
Ni[P(O-p-tolyl)3]4 as precursor.3
We have demonstrated here that the π-allyl nickel
cyanide complexes 1 and 2 result from C-CN bond
cleavage12 from 2M3BN and that they are active pre-
cursors in the catalytic isomerization of 2M3BN into
3PN. 2 is the only complex so far reported which is
active for the 2M3BN/3PN isomerization in the absence
of any extra ligand and of any Lewis acid. Moreover,
preliminary DFT studies show that a complex similar
to 1 or 2 but replacing Ni by Pd cannot be optimized.
All our attempts led to decoordination of one phosphine
and formation of a square-planar palladium complex.
This work highlights the unique properties of the nickel
system. Full details will be reported in due course.
Ack n ow led gm en t. We acknowledge the support of
the CNRS and Rhodia.
Su p p or tin g In for m a tion Ava ila ble: Text and tables
detailing the catalytic experiments and the synthesis of 2, full
details of the crystal structure analysis of 2 (CCDC Deposition
No. 227317), and DFT calculation details. This material is
available free of charge via the Internet at http://pubs.acs.org.
energy associated with the isomerization of 2M3BN to
3PN is -14.6 kJ mol-1. Relative energies of the opti-
mized geometries are indicated in Scheme 3. Isomers
A and E, with 2M3BN and 3PN respectively coordinated
through the double bond, are degenerate. The π-allyl
species C corresponds to the syn isomer, and the
geometrical data compare well with those obtained for
the X-ray structure of compound 2.9 We propose that
the key steps of C-CN bond cleavage and formation,
through the π-allyl species, most probably involves σ-π
allyl conversion (species B and D). The importance of
OM049706I
(10) Brunkan, N. M.; Brestensky, D. M.; J ones, W. D. J . Am. Chem.
Soc. 2004, 126, 3627.
(11) Carbo´, J . J .; Bo, C.; Poblet, J . M.; Moreto´, J . M. Organometallics
2000, 19, 3516.
(12) For recent data on oxidative addition of nitriles, see ref 10
and: (a) Yamamoto, T.; Yamaguchi, I.; Abla, M. J . Organomet. Chem.
2003, 671, 179. (b) Garcia, J . J .; Brunkan, N. M.; J ones, W. D. J . Am.
Chem. Soc. 2002 124, 9547. (c) Nakazawa, H.; Kawasaki, T.; Miyoshi,
K.; Suresh, C. H.; Koga, N. Organometallics 2004, 23, 117. (d) Taw, F.
L.; White, P. S.; Bergman, R. G.; Brookhart, M. J . Am. Chem. Soc.
2002, 124, 4192. (e) Churchill, D.; Shin, J . H.; Hascall, T.; Hahn J .
M.; Bridgewater, B. M.; Parkin, G. Organometallics 1999, 18, 2403.