Reactions in water: Alkyl nitrile coupling reactions using Fenton's reagent

Article abstract of DOI:10.1021/jo7026905

(Chemical Equation Presented) The coupling reaction of water-soluble alkyl nitriles using Fenton's reagent (Fe(II) and H₂O₂) is described. The best metal for the reaction is iron(II), and the greatest yields are obtained when the concentration of the metal is kept low. Hydrogen-atom abstraction is selective, preferentially producing the radical α to the nitrile. In order to increase the production of dinitrile, in situ reduction of iron(III) to iron(II), using a variety of reducing agents, was investigated.

Full text of DOI:10.1021/jo7026905

SCHEME 1
Reactions in Water: Alkyl Nitrile Coupling
Reactions Using Fenton’s Reagent
Christopher L. Keller,† James D. Dalessandro,†
Richard P. Hotz,*,† and Allan R. Pinhas*,‡
Department of Chemistry, College of Mount St. Joseph,
Cincinnati, Ohio 45233-1670, and Department of Chemistry,
UniVersity of Cincinnati, Cincinnati, Ohio 45221-0172
richard_hotz@mail.msj.edu; allan.pinhas@uc.edu
SCHEME 2
ReceiVed December 18, 2007
For this reaction, it does not matter if a free radical or an
iron-oxo complex is formed. What matters is that the Fenton
chemistry generates a radical4,5 or radical-equivalent6 that can
remove a hydrogen atom from the alkyl chain of the alkyl nitrile,
and then two of these “alkyl radicals” can couple. This type of
coupling reaction was first mentioned about 50 years ago,7 but
unfortunately, the yields were low and the regiochemistry was
not investigated in detail.
In this paper, we discuss the coupling of acetonitrile and other
water-soluble alkyl nitriles.8 We have improved the reaction
yield, and in addition, we have investigated the regiochemistry
of the coupling reaction. This regiochemistry not only is
important from a synthetic perspective, but it tells us about the
energetics of hydrogen-atom removal from various positions
on the alkyl chain. We wanted to determine whether hydrogen-
atom abstraction from an alkyl nitrile is statistical or selective,
with preferential removal of a hydrogen atom from the R-carbon
of the nitrile producing a resonance-stabilized carbon radical.
In situ reduction of iron(III) to iron(II), using a variety of
reducing agents, will also be discussed.
Variations in Metal. The ability of several transition metal
ions to do “Fenton-type” chemistry with acetonitrile was
examined (Scheme 2). In these experiments, the metal ion was
the limiting reagent. Metal ions with two readily available
common oxidation states were examined, and the best yields
were obtained using iron(II). Manganese(II) and cobalt(II)
complexes each failed to produce more than a trace amount of
succinonitrile. Copper(II) and nickel(II), which are metals with
a less common higher oxidation state, were also unsuccessful
in producing succinonitrile.9,10
The coupling reaction of water-soluble alkyl nitriles using
Fenton’s reagent (Fe(II) and H2O2) is described. The best
metal for the reaction is iron(II), and the greatest yields are
obtained when the concentration of the metal is kept low.
Hydrogen-atom abstraction is selective, preferentially pro-
ducing the radical R to the nitrile. In order to increase the
production of dinitrile, in situ reduction of iron(III) to iron(II),
using a variety of reducing agents, was investigated.
For the past several years, we have been interested in
carbon-carbon bond formation by oxidative coupling of
stabilized anions1 and anion equivalents.2 The coupling of anions
R to nitriles using transition-metal dihalide complexes has been
previously reported.3 As shown in Scheme 1, nickel halides
containing diphosphine ligands were treated with cyanomethanide
anion, generated by reaction of acetonitrile with butyllithium,
to produce bis(cyanomethanide)phosphine complexes. When
these nickel(II) complexes were exposed to oxygen, reductive
elimination produced succinonitrile as the only organic product.
Among the drawbacks to this method are the need for an inert
atmosphere and the cold reaction conditions (-78 °C).
Due to our desire to form new carbon-carbon bonds under
“green reaction conditions”, a water-compatible method for
coupling nitriles at the R-position was desired. This would allow
the reaction to be run at or near room temperature and the use
of costly solvents could be avoided. Thus, we became interested
in the use of Fenton chemistry.
(4) Walling, C. Acc. Chem. Res. 1998, 31, 155.
(5) (a) See, for example: MacFaul, P. A.; Wayner, D. D. M.; Ingold, K. U.
Acc. Chem. Res. 1998, 31, 159. (b) Goldstein, S.; Meyerstein, D. Acc. Chem.
Res. 1999, 32, 547.
(6) See, for example: Sawyer, D. T.; Sobkowiak, A.; Matsushita, T. Acc.
Chem. Res. 1996, 29, 409.
† College of Mount St. Joseph.
(7) (a) Coffman, D. D.; Jenner, E. L.; Lipscomb, R. D. J. Am. Chem. Soc.
1958, 80, 2864. (b) Also see: Walling, C.; El-Taliawi, G.; M, J. Am. Chem. Soc.
1973, 95, 844.
‡ University of Cincinnati.
(1) Belletire, J. L.; Spletzer, E. G.; Pinhas, A. R. Tetrahedron Lett. 1984,
25, 5969.
(8) Attempted radical-coupling reactions of phenylacetonitrile were unsuc-
cessful due to its insolubility in water.
(2) (a) LaDuca, M. J. T.; Simunic, J. L.; Hershberger, J. W.; Pinhas, A. R.
Inorg. Chim. Acta 1994, 222, 165. (b) Pinhas, A. R.; Hershberger, J. W.
Organometallics 1990, 9, 2840.
(9) Metal sulfates are used in the reaction because use of metal halides will
result in the generation of the corresponding halogen gas. The reaction mixture
must be kept acidic to prevent occurrence of nucleophilic attack at the nitrile
carbon.
(3) Alburquerque, P. R.; Pinhas, A. R.; Krause Bauer, J. A. Inorg. Chim.
Acta 2000, 298, 239.
3616 J. Org. Chem. 2008, 73, 3616–3618
10.1021/jo7026905 CCC: $40.75 2008 American Chemical Society
Published on Web 03/26/2008

SCHEME 3
SCHEME 4
Variations in Concentration. The best yields were obtained
when the iron(II) concentration was kept low. The reported
percent yields are based on the two iron(II) ions required per
alkyl dinitrile formed. With a constant amount of iron(II) sulfate,
using a large volume of acetonitrile and water (25 mL of each),
a 70% yield of succinonitrile was obtained. Using either 5 mL
of acetonitrile and 5 mL of water, or using 1 mL of acetonitrile
and 10 mL water, the yield was only 10–20%.
Selectivity of Hydrogen Abstraction. Since the cyanomethyl
radical is the only radical that can be obtained from acetonitrile,
succinonitrile is the only dinitrile product. However, propionitrile
can form two radicals: a resonance-stabilized secondary radical
(1) formed by abstraction of an R-hydrogen atom or a primary
radical (2) by abstraction of a ꢀ-hydrogen atom. The products
formed by all possible combinations of these two radicals are
illustrated in Scheme 3.
TABLE 1. Attempts at Reducing the Iron(III) Byproduct Back to
Iron(II)
Fe(III)
reducing agent
T
(°C)
time
(hr)
succinonitrile
(mmol)
none
25
25
40
25
25
25
25
25
44
80
3
3
0.5
3
3
1
1
3
1
1.75
2.38
2.00
0.75
0.75
8.00
9.00
7.62
8.00
4.50
Mn(II)
Mn(II)
Cu(II)
Ni(II)
H2S
Fe(0)
Fe(0)
Fe(0)
Fe(0)
1
SCHEME 5
Stereoisomers of 2,3-dimethylsuccinonitrile (4) (d,l-pair and
a meso compound) are formed by the coupling of two molecules
of the secondary radical (1), while two primary radicals (2)
combine to form adiponitrile (5). Cross-coupling of these two
radicals produces 2-methylglutamonitrile (3). The observed ratio
for dinitrile isomers 3:4:5 is 50:45:5. After correcting for the
number of abstractable hydrogens at each position, it was
determined that the resonance-stabilized secondary radical (1)
and the primary radical (2) form in a 3.5:1 ratio rather than the
statistical ratio of 2:3.
Isobutyronitrile also can form two radicals: a resonance-
stabilized tertiary radical (6) formed by abstraction of an
R-hydrogen atom or a primary radical (7) by abstraction of a
ꢀ-hydrogen atom. The products formed by all possible combina-
tions of these two radicals are illustrated in Scheme 4.
Stereoisomers of 2,5-dimethyladiponitrile (10) (d,l-pair and
a meso compound) are formed by the coupling of two molecules
of the primary radical (7), while two tertiary radicals (6) combine
to form 2,2,3,3-tetramethysuccinonitrile (9). Cross-coupling of
these two radicals produces 2,2,4-trimethylglutamonitrile (8).
The observed ratio for dinitrile isomers 8:9:10 is 26:29:45. After
correcting for the number of abstractable hydrogens at each
position, it was determined that the resonance-stabilized tertiary
radical (6) and the primary radical (7) form in a 4.3:1 ratio rather
than the statistical ratio of 1:6.
Since the cyanomethyl radical is easily reduced to the cyano-
methyl anion by iron(II), with a stoichiometric amount of iron,
the lower yield is not surprising. To improve the yield, the
concentration of iron(II) should be kept low. Thus, for the large-
scale preparation of succinonitrile via Fenton’s reagent, the
reaction should be catalytic in iron(II), keeping its concentration
as low as possible.
The ideal catalyst has no abstractable hydrogens and is
capable of reducing iron(III) to iron(II) without also re-
ducing the nitrile radical to the anion. These requirements greatly
limit the number of possible reducing agents. Table 1 sum-
marizes the reactions done in the presence of a reducing agent.
A second transition-metal ion was added as a potential
reducing agent; however, only the use of manganese(II) resulted
in an improved yield. When either copper(II) or nickel(II) was
used, a notable decrease in yield was observed. An increased
yield was obtained with hydrogen sulfide, but the coproduction
of elemental sulfur made succinonitrile isolation difficult.
Iron(0) is an attractive candidate as a reducing agent because
iron(II) is the only product of the oxidation–reduction reaction
(Scheme 5).
Although the use of iron(0) creates a heterogeneous reaction
which leads to greater variability in the product yield, the
increased production of succinonitrile indicates iron(0) has an
effect on the coupling reaction by reducing iron(III) to iron(II).
The observed trend is that with longer time and higher
temperatures, the lower the yield. A possible explanation is that
as the reaction proceeds, the concentration of iron(II) increases
and the yield approaches an upper limit. In addition, at higher
reaction temperatures, reduction of the radical to the anion
becomes more competitive with the radical coupling. We also
observed that at extended reaction times, intractable materials,
Attempts at Catalysis. Although there is literature precedent
for using Fenton’s reagent to prepare succinonitrile, in those
studies, a stoichiometric amount of iron(II) was added slowly
to a H2O2-acetonitrile mixture, and an 18% yield was obtained.
(10) The insolubility of copper(I) salts in water precluded their use in this
reaction: Cotton, F. A.; Wilkinson, G.; Murillo, C. A.; Bochmann, M. AdVanced
Inorganic Chemistry, 6th ed.; Wiley-Interscience: New York, 1999.
J. Org. Chem. Vol. 73, No. 9, 2008 3617

were dried over Na2SO4, and the solvent was removed by rotary
evaporation. The residual material was analyzed by GC-mass
spectrometry.
which mostly likely arose from further reaction of the dinitrile,
were formed.
For the catalysis studies, 5–10 mmol of reducing agent was added
prior to the addition of 30% aq H2O2. The variations in reaction
time and temperature are summarized in Table 1.
Conclusion
The radical R to the nitrile is formed preferentially. The
greatest yields are obtained using a low concentration of iron(II)
to generate the radical. Initial studies indicate that it is possible
to make the reaction catalytic in iron(II); however, lower yields
are observed with longer reaction times and elevated temperatures.
General Procedure (Selectivity Studies). A 1.4 g (5.0 mmol)
portion of FeSO4 ·7H2O, 1.0 mL of nitrile [acetonitrile (0.79 g, 19
mmol), propionitrile (0.77 g, 14 mmol), or isobutyronitrile (0.76
g, 11 mmol)], and 1 drop of concd H2SO4 were dissolved in 10
mL of H2O. After the mixture was cooled in an ice bath, 2.0 mL
(∼17 mmol) of 30% aq H2O2 was added slowly over 10 min. After
the mixture was stirred at rt for 2 h, 3 g of NaCl was added, the
reaction mixture was extracted with 2 × 25 mL of CH2Cl2, extracts
were dried over Na2SO4, and the solvent was removed by rotary
evaporation. The residual material was analyzed by GC-mass
spectrometry.
Experimental Section
All materials are commercially available and were used as
received. Usually, no precautions were made to remove air from
the reagents or solvents. In those reactions in which reagents and
solvents were degassed, the experimental results were the same.
General Procedure. In an Erlenmeyer flask, 5.0 mmol metal
sulfate (1.4 g of FeSO4 ·7H2O; 0.67 g of MnSO4; 1.4 g of
CoSO4 ·7H2O; 1.2 g of CuSO4 ·5H2O; or 1.3 g of NiSO4 ·6H2O),
25 mL of acetonitrile, 25 mL of H2O, and 1.0 mL of concd H2SO4
were combined. Five milliliters (∼44 mmol) of 30% aq H2O2 was
then added slowly over 15–20 min. After the mixture was stirred
at rt for the specified amount of time, 12 g of NaCl was added, the
reaction mixture was extracted with 2 × 50 mL of CH2Cl2, extracts
Acknowledgment. We thank BP Chemicals, the University
Research Council of the University of Cincinnati, and the
College of Mount St. Joseph for financial support of this
research. C.L.K. thanks the NSF-REU program for a summer
fellowship.
JO7026905
3618 J. Org. Chem. Vol. 73, No. 9, 2008