J . Org. Chem.

1981,46, 754-757

754

than the fragmentation of the carbon-halogen bonds.

Experimental Section
The four 6-halogeno-2H,3H-ben[b]~ophene-2,3-d
were
prepared by cyclization of the appropriate 3-halogenophenylthioacetic acid12J3according to the general procedure described
by Werner et al.14 and by subsequent oxidation of the 6halogen0-3-hydroxybenzo[b]thiophene'~*~~
by the method of El
Shanta et al.:16 6-F, mp 108-111 "C (from methanol); 643, mp
134-135 O C (from methanol); 6-Br, mp 169-170 "C (from methanol); 6-1, mp 197-198 O C (from benzene). Satisfactory elemental
(12) Pasto, D. J.; McMillan, D.; Murphy, T. J. Org. Chem. 1965, 30,

2688.

(13) Das, A. K.; Sinha, A. K. J . Indian Chem. SOC.

1972, 49, 993.
(14) Werner, L. H.;
Shroeder, D. C.;Ricca, S. J.Am. Chem. SOC.1957,
79, 1675.
(15) Das, A. K.; Sinha, A. K. J. Indian Chem. SOC.1967, 44, 933.
(16) El Shanta, M. S.; Scrowston, R. M.; Twigg, M. V. J. Chem. SOC.
C 1967, 2364.

analyses were obtained. Carbonyl adsorptions were recorded at

1725 cm-' (CS2). Solid potassium tert-butoxide and hydroxide
were commercial materials while sodium tert-butoxide was prepared by a standard method. Reagent grade solventswere utilized
without further purification with the exception of dimethyl
sulfoxide which was distilled under reduced pressure from calcium
hydride and stored over molecular sieves.
Radical anions, detected with a Varian E-109 spectrometer,
were simply generated in capillary glass tubes by adding the
solvent to the solid reagenta. Degassing of the solutions resulted
neither in any improvement of the quality of the ESR spectra
nor in any change of the observed reactivity.

Acknowledgment. Financial support from CNR

(Rome) is gratefully acknowledged.
Registry NO.I (X = F), 73655-11-5;I (X = Cl), 50789-28-1;I (X
= Br), 73655-14-8;I (X = I), 73655-16-0; I (X= H), 493-57-2;I1 (X
= F) K+, 75750-94-6; I1 (Y = F) Na+, 75716-07-3; I1 (Y = C1) K+,
75716-08-4; I1 (Y = Br) K+, 75716-09-5; 11 (Y = I) K+, 75716-10-8.

Gas-Phase Synthesis of Nitriles

Roger J. Card and J. Lawrence Schmitt*
Chemical Research Division, American Cyanamid Company, Stamford, Connecticut 06904

Received August 26, 1980

n-Octanenitrile is obtained in 87% yield from the copper-catalyzed, gas-phase reaction of 1-octanolwith ammonia
at 325 "C and 1atm. Similar high yields of alkyl or aryl nitriles are obtained from the reaction of ammonia with
primary alcohols, aldehydes, primary amines, secondary amines, and esters. The effect of reaction variables is
examined: highest yields of nitriles are obtained when the reaction is carried out at ca. 300 O C and at ammonia
to substrate ratios ?f greater than 101. The 15% Cu/A1203 m e was examined most extensively, but vanadium,
iron, and nickel aL show selectivity to nitriles while manganese and antimony give only very small amounts
of these products.
I

A number of methods have been developed for the

synthesis of organic nitriles. Although it is often possible
to convert a specific functional group into a nitrile group,
no convenient, general method of synthesis has been reported.
Approaches to nitrile synthesis may be divided into two
distinct groups. The first generally involves the use of
cyanide as a nucleophile and results in the formation of
an organic nitrile containing one more carbon atom than
the starting material. Several modifications in this
methodology have appeared recently. The reaction of
potassium cyanide with alkyl halides is facilitated by the
use of crown ether in acetonitrile solvent' or by the use
of hexamethylphosphoramide solvent with or without
added crown etherq2 The reaction of n-akyl bromidzs with
sodium cyanide impregnated onto alumina allows the
small-scale synthesis of aliphatic nitriles in hydrocarbon
~ o l v e n tand
, ~ aryl nitriles are formed from the reaction of
aryl halides with sodium cyanide in the presence of catalytic amounts of nickel phosphine4 or palladium phosphine5 complexes.

(6) Kryukov, Y. B.; Baehkirov, A. N.; Zakirov, N. S.; Novak, F. 1. Dokl.

Akad. Nauk SSSR 1966,170,852.
(7) Cathala, M.; Perrard, A,; Gennain, J. E.Bull. SOC.Chim. Fr. 1979,

(1) Cook, F. L.; Bowers, C. W.; Liotta, C. L. J. Org. Chem. 1974, 39,
3416.
(2) Shaw,J. E.; Hsia,D. Y.; Parries, G. S.;Sawyer, T. K. J. Org. Chem.
1978,43, 1017.
(3) Regen, S. L.; Quici, S.; Liaw, S.J. J. Org. Chem. 1979, 44, 2029.
Dalton,J. R.; Regen, S.L. Ibid. 1979, 44, 4443.
(4) (a) Cassar, L., J. Organomet. Chem. 1973,54, C57. (b) Cassar, L.;
Foa, M.; Montanari, F.; Marinelli, G. P. Ibid. 1979, 173, 335.
(5) Sekiya, A,; Ishikawa, N., Chem. Lett. 1975, 277.

(8)See for example: Burrington, J. D.; Graeaelli, R. K. J. Catal. 1979,

59, 79; Forni, L.; Stanga, M. Ibid. 1979,59, 148.
(9) Nakagawa, K.; Mieo, S.;Kawamura, S.;Horikawa, M.; Tokumoto,
T.; Mori, 0. Synth. Commun. 1979, 9, 529.
(10) Campagna, F.; Carotti, A.; Casini, G. Tetrahedron Lett. 1977,
1813.
(11) Olah,G. A.; Vankar, Y. D. Synthesis 1978, 702.
(12) Carotti, A.; Campagna, F.; Ballini, R. Synthesis 1979, 56.
(13) Olah,G. A.; Vankar, Y. D.; Gupta, B. G. B. Synthesis 1979,36.
(14) Furukawa, N.; Fukumura, M.; Akasaka, T.; Yoshimura,T.; Ode,
S. Tetrahedron Lett. 1980, 761.

0022-3263/81/1946-0754$01.00/0

An alternative approach involves the conversion of an

organic compound into a nitrile containing the same total
number of carbon atoms. One example involves the conversion of aliphatic primary alcohols to nitriles over a fused
iron catalyst6(eq 1). Nitriles have also been obtained from
n-C4H90H+ NH3

"Fe"

n-C3H7C=N (1)
a5%
the ammoxidation of aldehydes, ketones,' and benzylic or
allylic methyl groups? but these reactions require severe
conditions (usually >400 "C). Aryl nitriles are obtained
from the reaction of aryl aldehydes with ammonia and an
excess of nickel p e r ~ x i d e .Methods
~
have also appeared
for the conversion of amides to nitriles,lDthe dehydration
of aldoximes,"J2 the synthesis of nitriles from nitro comp o u n d ~ ,and
' ~ the synthesis of nitriles via the reaction of
sulfimide with aldehydes.14

173.

0 1981 American Chemical Society

Gas-Phase Synthesis of Nitriles

J. Org. Chem., Vol. 46, No. 4, 1981 755

Table I. Yields of Nitriles Obtained from the Reaction of

Ammonia with a Variety of Organic Compoundsa
starting mat1

product

yield,

temp, "C

benzaldehyde
benzonitrile
benzyl alcohol
benzylamine
p-methoxybenz- p-methoxybenzoaldehyde
nitrile

95
96
95 (89)
95

300
300
300
280

ethanol
ethyl acetate

acetonitrile

87
95c

325
325

n-butanol

butyronitrile

96

325

n-octanol
n-octanal
n-octy lamine
di-n-0ctylamine

n-octanenitrile

90 (87)
92
92
95c

325
325
325
325

The catalyst used was 15% Cu/Al,O,; NHJstarting

material ratio was - 2 O : l .
Yields were determined by
GLC. Yields in parentheses are isolated yields. Based
on formation of 2 mol of nitrile/mol of starting material.
a

180

200

220

?\*-, *.q

240 260

280

320

300

340

360

TEMPERATURE ("C)

Each of these routes works in reasonable-to-excellent

yield for a limited group of compounds. We felt that it
would be possible to synthesize nitriles from the vaporphase reaction of ammonia with a wider variety of organic
compounds than allowed by many of the current methods.
In addition, this approach should eliminate the need for
costly solvents and minimize the workup. We present the
results of this study here.

Results and Discussion

The conversion of 1-octanol to n-octanenitrile occurs in
good yield and illustrates the general method. 1-Octanol
was fed by syringe into a flowing stream of ammonia to
produce a mixture containing ca. 5 mol % of octanol. This
mixture was heated to 325 "C and passed through a catalyst bed containing 3 cm3of 15% Cu/A120p The pressure
in the system was ca. 1 atm, and the normal residence time
was ca. 2 s. The product stream was cooled, collected, and
analyzed. In this case, GLC analysis indicated a 90% yield
of n-octanenitrile. n-Octanenitrile was isolated in 87%
yield after drying and distillation of the crude product.
Similar yields of n-octanenitrile have been obtained from
the reaction of ammonia with n-octanal, n-octylamine and
di-n-octylamine. Other examples in Table I demonstrate
the generality of this process: alkyl and aryl nitriles are
prepared from a relatively wide variety of functional groups
including primary alcohols, aldehydes, primary amines,
secondary amines, and esters. In the case of dialkylaminea
and esters, 1mol of nitrile is formed from each "half" of
the molecule.
The conditions described above were the best found for
nitrile preparation and resulted from a careful study of the
reaction variables.
The most important variable is the reaction temperature. This is illustrated by the data in Figure 1. At 200
"C, the reaction of octanol with ammonia leads to noctylamine as the major product. In agreement with
predictions based on the calculated thermodynamica of the
system, the product ratio changes as the temperature is
increased, and octanenitrile quickly becomes the dominant
species formed. Other workers have observed a similar
temperature dependence in the amination of alcohols.6
Qualitatively, although enthalpy favors the amine, entropy
favors the nitrile and becomes the dominant fador above
-220 "C.
Small amounts of octane are observed at >320 "C.
Presumably, this is due to alumina-catalyzed dehydration
of 1-octanol followed by metal-catalyzed hydrogenation of

Figure 1. Product yields obtained from the reaction of 1-octanol

with ammonia over 15% Cu/&Os as a function of temperature:
NH3/H2/ROH = 2 0 5 1 ; octanenitrile, e; octylamine, A; octane,
1

Scheme I
R-CEN

4 2

II

H2

R-CH=NH

R-CHz-NHz

R-CH=O

-H2[1H2

R-CH-N-CHzR

the olefin intermediate. Octane is not observed under

these conditions when a silica support is used.
Formation of amines as the major products from the
reaction of alcohols with either ammonia or low molecular
weight amines is well documented and is used commer~ially.'~J~
The currently accepted mechanism for this process is
illustrated in Scheme I. The data in Table I1 demonstrate
that most of these reactions are reversible.
At low temperatures, the reaction is kinetically controlled, and different product ratios are obtained from each
starting material (Table 11). Thermodynamic equilibrium
(15) Baiker, A; Richarz, W. Ind. Eng. Chem. Prod. Res. Dev. 1977,16,
281. Baiker, A.; Richarz, W. Hela Chim. Acta 1978, 61, 1169.
(16) Schweizer, A. E.; Fowlkes, R. L.; McMakin, J. H.; Whyte, T. E.
Kirk-Othmer Encycl. Chem. Technol., 3rd Ed. 1978,2, 272-283. The

Commercial procesees normally require moderate pressure ~100-500pig)

and produce mixtures of mono-, di-, and trialkylamines.

756

Card and Schmitt

J. Org. Chem., Vol. 46, No. 4, 1981

Table 11. Yields of Products Obtained from the Reaction

of Ammonia with Octane Derivativesa
product yields, %b,c
reaction
temp,
matlb
"C RCH,OH RCH,NH, RC=N otherd
RCH,OH
RCHO
RCH,NH,
(RCH,),NH

185e
185e
185e
185e

20
2
0
0

50
56
70
54

10
15
9
11

21
27
20
31

RCH,OH
RCH,NH,
RCN
RCH,OH
RCH,NH,
(RCH,),NH
RCN

200
200
200
250
240
255
24 0

21
0
0
0
0
0
0

42
62
31
12
15
13
11

28
19
64
83
68
83
82

9
18
4
6
17
4
7

a Conditions, unless otherwise noted: 15%Cu/Al,O,

catalyst; NH,/H,/starting material ratio of 2 0 : 5 : 1 ;1 atm;
contact time ca. 2 s.
R = n-C,H,,.
Determined by
GLC.
Almost exclusively (RCH,),NH. e NH,/H,/starting material ratio of 50:15:1.

is reached at -250 "C, and above this temperature, nearly

identical product ratios are obtained from the various
starting materials. (See, for example, the data in Table
I.)
Consistent with the sequence of equilibria outlined in
Scheme I, the molar ratio of hydrogen, ammonia, and
starting material is also important in determining the
products obtained (Table 111). For the reaction of l-octan01 with ammonia, low hydrogen ratios result in maximum yields of nitrile, high ammonia/substrate ratios favor
monoalkylamine formation, and low ammonia/substrate
ratios favor dialkylamine formation. Similar trends have
been qualitatively observed in other systems and appear
to be general. Since hydrogen is detrimental to nitrile
synthesis, no hydrogen was used in the preparative runs.
Also, since a high ammonia to substrate ratio gave the best
results, this ratio was maintained at 120:l for the preparative work.
The mechanism outlined in Scheme I contains two
different types of reactions: dehydrogenation and net
nucleophilic addition to unsaturated intermediates, e.g.,
1 and 3. An effective catalyst should be able to facilitate
both processes.
A number of catalysts were screened for nitrile synthesis.
Table IV presents a summary of the data. In general, the
catalysts were examined over a temperature range of
200-340 O C . The data in Table IV present the maximum
yield of nitrile obtained over each catalyst and the minimum temperature at which that yield was obtained. The
ammonia/hydrogen/substrate ratio was maintained at
20:5:1.
This work was intended to show whether other supported catalysts would perform as well as 15% Cu/A1203
and whether they might, in addition, allow the use of lower
Table 111. Effect of the A"onia/Hydrogen

Ratios o n the Reaction of 1-Octanol with Ammonia over 15% Cu/AI,O,

product yield, %a,b

NH,/H,/RCH,OH
molar ratio

temp, "C

RCH,OH

20:5:1
15:lO:l
10:15:1
5:20:1
20:5:1
15:lO:l
10: 15:1
5:20:1

220
220
220
220
250
250
2 50
250

2
2
0

Determined by GLC.

R = n-C,H, s .

metal loadings or low reaction temperatures or both. From

the data, it appears that effective catalysts could be prepared from vanadium, iron, and nickel as well as copper,
but none of these systems appears to offer significant improvement.
For the copper/alumina system, a large decrease in reaction temperature is observed on increase of the metal
loading from 5% to 15%. A further increase to 30% results in only a minor additional decrease in the optimal
reaction temperature.
It is interesting that manganese and antimony displayed
very minor activity for nitrile formation. Reaction of 1octanol over these catalysts resulted in relatively clean
formation of mixtures of n-octylamine and di-n-odylamine.
Apparently, the manganese and antimony systems facilitate nucleophilic addition to an imine-type intermediate
more efficiently than they catalyze the corresponding
dehydrogenation.
Our results suggest that although many metah exhibit
catalytic activity for producing nitriles, careful control of
metal loading and choice of support must be exercised to
obtained optimum nitrile yields. Although we have not
investigated catalyst stability with extended runs, we have
observed no significant activity decline with 15% Cu/A1203
over a period of several days.
Some limitations to this process have been identified.
Neopentyl glycol, tetrahydrofuran, ethylene glycol, N,Ndimethylethanolamine, and 1,3-propanediol were not
converted into mono- or dinitriles under any conditions
studied. In each case, a complex product mixture was
obtained, and no more than minor amounts of nitrile
products could be found.
Halides also present difficulties. Reaction of benzyl
chloride with ammonia over fresh 15% Cu/Al2O3 catalyst
at 300 "C produced a 60% yield of benzonitrile. Other
products included bibenzyl(35%) and toluene (10%); but
deactivation of the catalyst was rapid, possibly due to the
poisoning effect of chlorine on copper or formation of
NHICl which may have clogged the pores of the supported
catalyst.
o-Chlorobenzaldehyde reacted with ammonia over 15%
Cu/A1203 at 320 "C to give o-chlorobenzonitrile in only
25% yield. This may have been due to steric problems.
(0-Chlorotoluene was recovered unchanged under these
conditions.)
In conclusion, we have demonstrated that the vaporphase reaction of ammonia with primary alcohols, esters,
amines, and aldehydes can produce nitriles in very good
yield. The reaction can be used for the synthesis of both
aryl and aliphatic nitriles, does not use a solvent, and
requires only a minor workup. The reaction is sensitive
to the presence of additional functional groups and works
best for relatively simple systems. All reaction conditions
are important, but careful control of temperature and
substrate/ammonia mole ratio is essential.

0
0
0
0

43
50
49
41
12
25
41
37

45
28
12
5
83
72
46
28

9
18
40
54
6
3
13
36

Gas-Phase Synthesis of Nitriles

Table IV. Effect of Catalyst on the Efficiency of
Production of n-Octanenitrile from 1-Octanol
yield of
catalyst
temp, "C RCsN, %
5% V/Al,O,
340
25
5% Mn/5-A sieves
0
340
5% Mn/Al,O,
340
0
5% Fe/Al,O,
340
13
5% Fe/C
34 0
52
5% Fe/SiO,
0
34 0
5% Ni/C
27 5
71
4% Ni/4-A sieves
280
63
5% cu/Al,o,
34 0
9
1 5 % cu/Al,o,
280
97
30% Cu/Al,O,
260
96
5% Sn/Al,O,
340
4
340
5% Sb/A1,0,
<1
340
<1
5% Sb/C
340
5% Sb/5-A sieves
<1
15% Sb/Al,O,
340
8
5% Cu, 5% Ni/Al,O,
28 0
85
5% Cu, 5% Zn/Al,O,
300
97
6% Cu, 5% Mn/AI,O,
350
71

Experimental Section
AU chemicals were reagent grade and were used without further
purifcation. The alumina catalyst support was Cyanamid HCFX
7-Al2O3in the form of 0.050-in. TRILOBE extrudates. It was
calcined at 600 "C for at least 1 h prior to use. Physical properties
surface area 304 m2/g (N2BET area); pore volume 0.64 cm3/g.
The carbon is a Cyanamid proprietary material with a surface
area of 550 m2/g and a pore volume of 0.85 cm3/g. Grade 01silica
was obtained from Davison: pore volume 0.43 cm3/g; surface area
695 m2/g. Molecular sieves were obtained from Linde (4 A) or
Aldrich (5 A).
NMR spectra were obtained on a Varian A-60 or EM-360A
nuclear magnetic resonance spectrometer. Analytical gas chromatography was carried out on a Hewlett-Packard 5750 instrument with a flame-ionization detector, complete with a Hewlett-Packard 7128A recorder with disk integration, or a Hewlett-Packard 7128A instrument with a flame ionization detector,
complete with an HP-7127A recorder with electronic integration.
All GLC analyses were carried out on a 1.3 m X 6.5 mm glass
column packed with 28% Pennwalt 223 plus 4% KOH on Gas
Chrom R (Applied Science Laboratories).
Reactor. The reactor consisted of a 13 mm (0.d.) X 270 mm
Pyrex glass tube terminating in a 10/30 ground-glass joint. A
coarse-fritted glass disk was located 30-40 mm above this joint.
A side arm (9 mm 0.d. X 25 mm) was located 25 mm from the
top of the tube for the introduction of the various reactants. The
top of the tube was fitted with a rubber serum cap penetrated
by a 3-mm (0.d.) Pyrex glass thermowell. A chromel-alumel
thermocouple (Omega Engineering Inc.) was adjusted so that its
tip was located in the center of the catalyst bed.
The reactor was placed in a tube furnace with a 2 X 22 cm core.
The position of the tube was adjusted so that the catalyst bed
was located at the center of the furnace.
The hydrogen and ammonia were introduced through separate
flow meters (Lab Crest) which were calibrated for the respective
gases. The organic substrate was introduced by syringe by using
a Harvard Apparatus compact infusion pump which was calibrated

J. Org. Chem., Vol. 46, No. 4, 1981

757

for each substrate used. (For isolated yield determinations, the

amount of starting material fed into the system was determined
by weight.)
A number of collection system were used. These were uswlly
based on a straight tube collector fitted at the top with a female
10/30 joint. The product could then be collected by bubbling
the product gas through methanol or by collection in a cold trap.
Catalysts. AU catalysta used were prepared by the pore volume
saturation ("incipient wetness") method. Compositionsare given
as the percent metal on the support regardlees of the form of metal
present. For example, the 15% Cu/AlZO3 catalyst was prepared
by the addition of a solution containing 14.1 g of Cu(N0&3H20
dissolved in 12.8 mL of water to 20.0 g of AlzO3. After thorough
mixing, the moist solid was dried for several hours at 120 "C and
then calcined for 2 h at 500 "C in flowing air.
General Reaction Procedure. In a typical run,1-odanol was
fed by syringe (0.012 g/min, 0.1 mmol/min) into a flowing stream
of ammonia (45 mL/min at room temperature, 2 mmol/min) or
ammonia plus Hz (NH3/Hz ratio of 5, total flow 55 ml/min). The
gas mixture was passed through a vaporization and preheating
section packed with 3-mm glass beads into a hot zone held at 325
"ccontaining 3 cm3of 15% Cu/&03. The product stream passed
through the glass frit and was collected in methanol. The system
was run for 20 min (2 mmol of octanol consumed), and then the
alcoholic product solution was diluted to a total volume of 100
mL and analyzed by GLC using an external reference. Products
were identified by retention time and peak enhancement with
known samples.
In an alternative method, the product was collected in an
ice-cooled, dry receiver flask. The crude reaction product may
be analyzed by N M R and/or GLC, or it may be dried (anhydrous
MgSO,) and distilled to give pure nitrile.
n-Octanenitrile. In the manner outlined above, 5.0 g (0.04
mol) of n-octanol was fed over 15% Cu/A&O3catalyst at a rate
of 0.02 mL/min (0.1 mmol/min). Ammonia was introduced at
100 mL/min (4 mmol/min). The catalyst was maintained at 325
OC. The product was collected in a trap in an ice-water bath,
taken up in acetone, dried (anhydrous MgSO,), and distilled to
yield 4.2 g (87%) of n-octanenitrile: NMR (DCClJ 6 2.3 (m, 2
H), 1.4 (m, 10 H), 0.9 (m, 3 H); this is identical with that of
commercial material. The product contained only trace quantities
of amine by GLC analysis.
Benzonitrile. Benzylamine (5.2 g, 0.05 mol) was allowed to
react with ammonia under conditions identical with those above
except that the catalyst was maintained at 300 "C. The product
was collected at icewater temperature, taken up in methylene
chloride, dried (anhydrous MgS04), and distilled to give 4.5 g
(89%) of benzonitrile. The product contained no more than trace
impurities by GLC, and ita N M R (DCC1,) spectrum was identical
with that of commercial material, 6 7.55 (m).

Acknowledgment. We thank E. Battistelli and D.

Tabak for experimental assistance and Drs. A. Zweig, R.
L. Potter, and R. Proverb for helpful discussions.
Reg-istry No. Benzaldehyde, 100-52-7; benzyl alcohol, 100-51-6;
benzylamine, 100-46-9;p-methoxybenzaldehyde,123-11-5; ethanol,
64-17-5; ethyl acetate, 141-78-6;butanol, 71-36-3; octanol, 111-87-5;
octanal, 124-13-0; octylamine, 111-86-4; dioctylamine, 1120-48-5;
benzonitrile, 100-47-0; p-methoxybenzonitrile, 87490-8; acetonitrile,
75-05-8; butyronitrile, 109-74-0; octanenitrile, 124-12-9; ammonia,
7664-41-7;octane, 111-65-9.