2120 G. D. VAN EPPS AND E. EMMET REID.

FROM THE CHEMICALLABORATORY

[CONTRIBUTION OF JOHNS HOPKINS ]
UNIVERSITY.
STUDIES IN THE PREPARATION OF NITRILES. 11. THE PREP-
ARATION OF ALIPHATIC NITRILES.
BY G. D VAN EPPS A N D E. EMMET
REID.
Received August 2, 1916
Introduction.'
For the preparation of nitriles we have had the old method of Letts'
as improved by K r i i s ~which
,~ was t o heat an acid, such as benzoic acid,
with lead thiocyanate. This method was easy to carry out and gave
fair yields. Several years ago,3in this laboratory, this method was studied
and i t was found that far better yields could be obtained by substituting
the zinc salt of the acid for the free acid, but only aromatic acids were
experimented with. The present work was taken up to test the ap-
plicability of this improved method to aliphatic nitriles and to find the
best conditions for their preparation.
The preparation of acetonitrile has been studied most extensively on
account of the importance and the availability of acetic acid. Various
methods of heating have been experimented with. Zinc salts of acetic,
propionic, butyric, isovaleric, stearic, phenylacetic, phenylpropionic,
cinnamic, and terephthalic acids, and the lead salts of some of these acids,
have been heated with lead thiocyanate. Several other thiocyanates
have been used and the results compared with those obtained with the
lead salt. Calcium, copper, stannous, manganous, nickel, cobalt, mag-
nesium, strontium, barium and basic ferric acetates have been tried.
Historical.
E. A. Letts4 of London, in 1872, studied the action of potassium thio-
cyanate on acids. In case of acetic acid, he could not obtain the nitrile,
but with isobutyric acid and valeric acid he found both amide and nitrile,
the amide being the main product. With benzoic acid the nitrile was
almost the sole product, only traces of the amide being found.
Pfankuch6 heated barium benzoate with excess of barium thiocyanate
and got benzonitrile but mixed with large quantities of other products.
Barium acetate and barium thiocyanate gave similar results, the other
products being in excess of the acetonitrile.
KekulP reviewed the work up t o his time and speculated as to the
mechanism of the reaction, but made no advance in its application.
G. Kriiss7 complained of the yield in Letts' process and improved the
Ber., 5 , 669 (1872).
* Ibid., 17,1766 (1884).
A m . Chem. J., 43, 152 (1910).
1 Ber., 5, 669 (1872).
J . prakt. Chem., [ 2 ] 6, 97 ( 1 8 7 3 ) .
e Ber., 6 , 111 (1873).
' Ibid., 17, 1766 (18841.
 STUDIES IN THE PREPARXTIOK OF NITRILES. 11. 2121

method by substituting lead thiocyanate for the potassium salt. Krdss

gives the equation
+
z C ~ H ~ C O O H PbJCNS)2 = 2CbHsCK + PbS+ H2S f zCO?.
He reported a yield of 60 to 35 g. of benzonitrile from IOO g. of benzoic
acid, or double what he was able to obtain by the method of Letts.
Reid' studied this reaction quantitatively and showed that it is not
represented correctly by the equation given. He concluded that the
nitrile is formed by interaction between the lead thiocyanate and the lead
salt of the organic acid, this salt being formed by reaction between the
free acid put in and a part of the lead thiocyanate. This led to a sub-
stitution of the previously prepared salt of the organic acid for the free
acid, and to extended experiments with different salts, the conclusion
being that the best yield is obtained when the zinc salt of the organic
acid is used. Reid's method left little to be desired in the preparation
of benzonitrile and yther aromatic nitriles, both as to yield and con-
venience of work, but was not extended to aliphatic.
Materials.-Lead thiocyanate was prepared by precipitation from lead
nitrate and ammonium thiocyanate. The precipitated salt was well
washed with water, and twice with alcohol. The air-dried salt was
further dried in an air-bath a t 120'. The salt is pure white but turns
somewhat yellow on long keeping. I t blackens somewhat when dried
in the air-bath.
Zinc thiocyanate was prepared by boiling ammonium thiocyanate
with excess of zinc hydroxide till no more ammonia came off. The solu-
tion was filtered hot and evaporated to crystallization. The crystals
were centrifuged. I t is difficult to dry this salt as it fuses easily and
begins to decompose a t low temperatures. The salt is somewhat hygro-
scopic.
Cuprous thiocyanate was prepared by mixing copper sulfate and ammo-
nium thiocyanate in the presence of sulfur dioxide. The gray salt was
well washed with water and dried in an air-bath. It turns yellow.
Zinc acetate, calciinln acetate, zinc butyrate, zinc propionate, and zinc
isovalerate were prepared by boiling water solutions of the acids with
excess of the carbonate of the metal. The solutions were filtered hot and
evaporated to crystallization. The latter three salts were dried ox-er
sulfuric acid in a vacuum, the others were dried in an air-bath.
Zinc stearate, lead phenylacetate, zinc phenylacetate, lead cinnamate,
zinc cinnamate, zinc hydrocinnamate, lead benzoate, and zinc terephthalate
were prepared by double decomposition from the neutral ammonium or
sodium salts of the acids and lead nitrate or zinc sulfate. All of these
salts were dried in air-bath a t 120'.
Potassium thiocyanate, lead acetate, copper acetate, barium acetate, liasic
l A m . Chem J , 43, 162 (1910).
2122 G. D. VAN EPPS AND E. EMMET REID.

ferric acetate, stannous acetate, manganous acetate, nickel acetate, cobalt

acetate, magnesium acetate, and strontium acetate were good commercial
preparations.
Method of Work.-In each experiment the salts were thoroughly
dried and powdered. The weighed amounts were thoroughly mixed
before the heating was begun. As the heat is applied the mixtures melt
and blacken and bubbles of gas and vapors come from the pasty mass.
Usually the nitrile begins to 'drop from the retort neck in a very short
time. The reaction proceeds regularly and spreads through the whole
mass, which becomes more or less liquid throughout. At the end of the
reaction, the nitrile ceases to come off and the residue sets to a hard mass
of the sulfides of the metals.
Five methods of heating the mixtures were tried. In the tables the
experiments are marked with the letters, a , b, c, d , and e t o show in which
of these ways the operation was carried out.
a. Heating in an iron retort under reduced pressure. The retort used
was such as is sold for the distillation of mercury and was of 500 cc.
capacity. The lid was luted on with a ring of asbestos paper wet with
glycerine. Reduced pressure was mainkained by a water pump. The
heating was so regulated that the pressure was kept j o to 100 mm.
b. Heating in the iron retort under atmospheric pressure. I n this
the same retort was used. In both cases the retort was heated by a
ring burner which supplied the heat uniformly to the walls of the retort,
additional heat being supplied by a Bunsen burner to complete the re-
action. The distillates from the iron retort were somewhat more colored
than those from glass, owing to small amounts of iron dissolved.
c. Heating in a glass retort. The retorts used were of 2 5 0 and 500 cc.
capacity without tubulature. The heating was done by means of a Bunsen
burner held in the hand, and the heat was applied a t first one spot and
then another on the side and finally a t the bottom. The retort may be
cleaned with nitric acid and used repeatedly. With the glass retort,
the course of the reaction is readily followed and it is readily kept under
control.
d. Heating in a horizontal glass tube. The tube used was of hard
glass, 175 mm. long and 2 2 mm. in diameter, and closed a t one end. The
open end was closed by a cork through which a narrower tube passed
to the condenser. The heating was done by a burner held in the hand.
This was used for small scale experiments. It has the advantage that
the nitrile does not condense and run back and does not remain so long
in contact with the highly heated material.
e. Heating in a horizontal glass tube surrounded by a brass block.
I n this case the tube was of soft glass, 300 mm. long and 2 2 mm. in diam-
eter, and closed a t one end. The block was of cast brass and was j o
 STUDIES IN THE PREPARATION OF KITRILES. 11. 2123

mm. square by 300 mm. long. A centrally located hole 25 mm. in diam-
eter extended from one end t o the other. The open end of the glass tube
was closed by a cork carrying a smaller tube as in d. The end carrying
the cork projected far enough out of the brass block to avoid burning the
cork. The advantage in this method of heating is that the heat is ap-
plied regularly throughout the whole length of the tube and the intense
local heating, caused by the direct application of the flame, is avoided.
As all parts of the tube are above the boiling point of the nitrile, there
is no condensation and running back of the nitrile. The brass block was
heated by two Bunsen burners properly spaced.
In no case was much trouble experienced with foaming except with the
mixture containing zinc thiocyanate. The last two methods gave much
purer products, but the yields are much more uncertain, owing to the
difficulty of working up the small amounts of the products. The losses
of nitrile are proportionately greater than with the larger amounts.
With these tubes only small amounts of materials can be used.
Working up Products.--The methods of separating the nitriles from the
crude distillates were those commonly used in organic preparations.
The distillate containing acetonitrile or butyronitrile was diluted with
half its volume of water, treated with solid potassium carbonate or ammo-
nia gas to saturation, the upper layer separated and fractioned. I n some
cases, the total distillate containing acetonitrile was fractioned up to
85 and the distillate put through this treatment.
Propionitrile was separated by calcium chloride and the upper layer
fractioned. Isobutyl cyanide was separated by fractionation.
I n the cases of stearic, cinnamic, hydrocinnamic, and phenylacetic
nitriles, the crude distillates were treated with a little water and a slight
excess of ammonia and steam distilled. The nitrile was separated and
the water layer extracted with ether. Stearic nitrile was also separated
from the crude distillate by fractionation in a vacuum. It was never
obtained pure. I

The crude terephthalic nitrile was purified by sublimation. This

gave beautiful short white needles melting a t about zoo’.
Results.-The results of the various experiments are brought together
below in tables. Under “Ratio” is given the ratio of amount of the
thiocyanate used to the calculated amount. Under “Prod.” is given
the weight of the crude product obtained by the distillation of the mixture
of salts, and, in the next column, the weight of the nitrile isolated from
such crude product. In the last column is given the yield in per cent.
of the calculated.
Discussion of Results.
From Table I, in which are given experiments in which acetates of
various metals were heated with lead thiocyanate, it appears that the
2124 G. D. VAN EPPS AND E. EMMET REID.

acetates of the most of these metals give considerable yields of the nitrile.
Nickel acetate gave little and magnesium acetate none that could be
isolated. The distillates in Experiments 7 and g from mixtures con-
taining magnesium and barium acetates had fearful, suffocating odors
which may have been due to the presence of isocyanates. I n an experi-
ment, not included in the table, 75 g. of potassium thiocyanate were heated
with 2 5 g. calcium acetate and only 1 . 2 g. of crude distillate was ob-
tained. This had a similar bad odor.
TABLEI.
Acetates of Various Metals Heated with Lead Thiocyanate.
Ace- Ace-
No. Method. tate. Pb(CNS)z. tate. Ratio. Time. Prod. Nitrile. Calc. Yield.
I d Zn 4.0g. 2.08. I. 14 30 4 . 0 g. 0 . 2 g. 0 . 9 g. 22 %
2 d Zn 11.0 5.0 I .25 75 1.4 0.56 2.2 25
3 d Mn 20.0 9.0 I .20 90 3.0 3.0 4.3 70
4 d h?n 10.0 4.2 I .28 45 1.1 0.5 2 .o 25
j d Ni 10.0 5.0 1.37 45 1.5 0. I 2.3 4
6 d Co 4.0 2.0 1.37 30 0.5 0 . 2 0.9 22

7 d Mg 7.0 3.0 I ,03 30 0.6 ? ... ?

8 d Sr 6.0 3 4 1.13 45 0.3 0.3 1.4 21

9 d Ba 6.5 5.0 I .08 45 0.8 0.3 I .6 19

IO 1. cu 90 50 I .oo 60 25 6.0 22.6 27
11 c Pb 120 I20 I .oo 2 40 28 5 .O 30.3 16
12 c Pb 85 83 I .oo 90 22 6.0 20.9 29
13 c Pb 95 85 1.12 150 21 6.0 21.5 28
TABLE11.
Lead, Cuprous, and Zinc Thiocyanates with Zinc Acetate.
No. Method. Metal. 'M(CNS)z. Zn(Ac0)t. Ratio. Time Prod. Nitrile. Calc. Yield.
I d Pb 4.og. 2 . 0 9 . 1.14 30 4.og. 0 . 2 g . 0.9g. 22%
2 d Pb 11.0 5.0 1.26 75 1.4 0.56 2.2 25
14 d Cu 3.0 4.0 0.57 30 0.8 0.4 1.0 40
I j e Zn 2.0 3.0 0.99 180 0.7 0.4 1.3 31
TABLE111.
Lead Thiocyanate and Zinc Acetate in Different Proportions and with Different
Methods of Heating.
No. Method. Pb(CNS)$.Zn(Ac0)i. Ratio. Time. Prod. Nitrile. Calc. Yield.
16 c 90 g. 50 g. 1.03 210 2 3 . 0 g. 9 . 0 g. 2 2 . 4 g. 40%,
17 C 72 20 2.05 120 11.0 4.0 8.9 45
18 c 115 60 1.09 195 28.0 15.5 26.8 57
I9 c 50 25 1.13 150 7.0 1.4 11.2 12
20 spec. 18 IO 1.02 180 3.0 LO 4.5 22
I d 4.0 2.0 1.14 30 4.0 0.2 0.9 22
2 d 11.0 5.0 1.26 75 1.4 0.56 2.2 z j
21 e 6.0 3.0 1.13 60 1.5 0.8 1.3 61
22 c 6.0 3.0 1.13 60 1.2 0.3 1.3 24
23 e 6.0 3.0 1.13 180
0.6 0.3 1.3 24
24 b 40 20 1.13 180
7.4 4.1 8.9 46
25 b 100 50 1.13 19.8
420 11.3 22.4 jo
26 b 140 70 1.13 420
26.5 17.8 31.3 57
M = Pb, Zn, or CUI:! as indicated in 3rd column.
 STUDIES IN THE PREPARATION OF NITRILES. 11. 212j

TABLE IV.
Zinc Salts of Propionic, Butyric, Isovaleric, Stearic, Phenylacetic, Cinnamic, Hydro-
cinnamic, and Terephthalic Acids with Lead Thiocyanate.
No. Meth. Acid. Pb(CNS)z. ZnAz.2 Ratio. Time Prod. Nitrile. Calc. Yield.
27 e propionic 5.og. 3.0g. I .oy I20 I . 5 g. L I ~ .1.6g. 71%
28 C butyric 35.0 18.0 1.44 150 10.0 3.0 10.4 29
29 e butyric 4.5 3.0 0.97 90 1.j 1.3 2.1 61
30 b butyric 30.0 15.0 I .30 120 8.9 5.3 8.6 62
31 e i.valeric 5.0 3.0 I .38 I20 I .3 1.0 1.86 54
32 e i.valeric 4.0 1.5 2.21 I20 0.9 0.8 0.93 86
33 a stearic 51 IO0 I .oo 90 52 16 84 I9
34 (1 stearic 60 60 2 .m I20 50 5 50 10
35 c stearic 15 25 I .26 45 22 20 21 95?
36 c phthalic 45 40 I . 16 90 21 15 30 50
37 U cinnamic 75 50 .I ,67 2 40 .... I4 44 32
38 U cinnamic 7 5 25 3.33 2 40 .... 8 22 36
39 C cinnamic 6.5 25 2.97 I 80 8 6 22 27
40 C hydrocin. 35 35 1.12 90 21 18 25 72
41 C hydrocin. 5 0 25 . 2.20 60 19 12 18 67
42 d terephth. 5.0 3.0 I. 18 30 I .o I .o 1.7 59
TABLEV.
Lead Salts of Phenyl Acetic and Cinnamic Acid with Lead Thiocyanate.
No. Method. Acid. Pb(CNS)z PbAg.1 Ratio. Time. Prod. Nitrile. Calc. Yield.
43 a ph. ac. 35 g. jog. 1.03 120 15 g. 6 . 0 g . 24 g. 25%
44 a cin. 100 75 2.04 Ijo . ... 8.0 37 21

In most of the first g experiments the amounts of materials used were

small and the amounts of acetonitrile worked with were quite small.
On this account the losses in handling the nitrile were proportionately
great and the results given, representing the amounts isolated, are prob-
ably much too low. It appears from these experiments that manganese
and zinc acetates are the most suitable for the preparation of the nitrile.
The manganese acetate gave a particularly high yield and a clean pure prod-
uct, but as zinc salts are so much more readily obtained than those of
manganese, the zinc salts are to be preferred for practical work.
Experiments IO, 11, 12, and 13 with copper and lead acetates, in glass
retorts using larger amounts of materials, are to be compared with Ex-
periments 15, 17, and 18 of Table I11 in which zinc acetate was used under
the same conditions. It thus appears that zinc acetate gives nearly
twice the yield that lead acetate does. The stability, fusibility, and
mutual solubilities of the fused salts are probably the determining factors
in the matter of the yields from various mixtures.
From the results in Table 11, it appears that thiocyanates of other
heavy metals can be used. In Experiment 14 an excellent yield was
obtained from cuprous thiocyanate. In this case the yield is calculated
from the thiocyanate present instead of from the acetate as in all the
A stands for one equivalent of the acid indicated by abbreviation in 3rd column.
2126 G. D. VAN EPPS AND E. EMMET REID.

other experiments. The results obtained with zinc thiocyanate suggest

that it might be substituted for the corresponding lead salt t o advantage,
but its physical properties make against its use. Lead thiocyanate,
being insoluble in water, is readily obtained by precipitation, and is
easily dehydrated. Zinc thiocyanate is very soluble in water and hence
inconvenient t o prepare and is difficult to dehydrate. For these reasons
the lead salt is better for practical work.
From the results given in Table 111, it appears that much better re-
sults are obtained by methods c and b in which larger quantities of ma-
terials are used in glass or iron retorts. Contrasting Experiment 16 with
Experiment 17 in which double the calculated amount of the lead sulfo-
cyanate was used, it appears that little or nothing is gained by the use
of this large excess.
From the experiments in Table IV, it appears that the yields are better
with the higher molecular weight acids than with acetic, isovaleric giving
86y0 of the calculated amount of nitrile. This may be only apparent,
since the higher nitriles, on account of their being less volatile and less
soluble in water, are much easier to isolate than is acetonitrile. The
results with stearic acid are not regarded as reliable as no good method
is available for the isolation of this nitrile.
Comparing Experiments 37 t o 39 with 40 and 31,we see the effect of
the ethylene group in cinnamic acid. A much larger yield of hydro-
cinnamic nitrile is obtained than of the cinnamic. The use of diminished
pressure during the distillation in Experiments 37 and 38 did not ma-
terially increase the yield. The high yield of hydrocinnamic nitrile,
when the mixture of salts is heated in a plain glass retort, is surprising.
This shows that the method is available for the preparation of nitriles
of high molecular weight, provided there are no interfering groups in the
molecule.
I n this work, the mixed aromatic-aliphatic acids in which the carboxyl
group is not in the ring are considered as aliphatic.
Experiment 42, with terephthalic acid, an aromatic acid, does not
properly belong to this study, but was tried to test the applicability of the
method for di-nitriles.
The experiments in Table V, with lead salts, are to be compared with
those with the corresponding zinc salts in Table IV. The com-
parison is to the advantage of the zinc salts as i t was with the acetates in
Table I.
According to Liebig and Rammelsburg' and also Voelckel,2 thiocyanic
acid decomposes when heated, in presence of water, into carbon dioxide,
carbon disulfide, and ammonia.
1 Pogg. Ann., 56, 94 (1842).
Zbid., 58, 135 (1843);Ann., 43, 74 (1842).
 STUDIES IN THE PREPARATION OF NITRILES. 11. 2127

It was thought that possibly some of these decomposition products

might act on the lead salt of an acid to give the nitrile. Accordingly
an experiment was performed to determine the effect of the decomposition
products of thiocyanic acid on the lead salt of benzoic acid. Dry ammonia
was passed through carbon disulfide which was sufficiently warmed to
give it the proper vapor tension. The mixture of the vapors of ammonia
and carbon disulfide was passed over lead benzoate (39 g.), which was
heated in an oil bath to a temperature of 220'. Results: a mixture of
benzoic acid and benzonitrile condensed on the cooler part of the lead
benzoate flask and also in the condenser. A few drops of a pale yellow
liquid passed through the condenser. This had an odor of carbon di-
sulfide. All the lead benzoate was converted into a black residue of lead
sulfide, with black mirror of same around the upper part of the flask.
The amount of benzonitrile formed was insignificant. This shows that
ready formed CS2 and NH3 do not produce the same results as the lead
thiocyanate.
In another experiment, a mixture of ammonia and carbon dioxide
was passed over a mixture of 3j g. of lead benzoate and 5 g. of sulfur,
heated in an oil bath to a temperature of 240'. Results: Solid ammo-
nium carbonate was deposited on the walls of the lead benzoate flask
and in the condenser and also in the receiver. Sulfur sublimed and de-
posited on the upper part of the flask. The odor of benzonitrile was ap-
parent but the amount formed was insignificant.
In Experiment 2 0 , to determine whether the nitrile is decomposed by
side reactions, etc., into the free acid, the mixture of zinc acetate and
lead thiocyanate was heated in an oil bath with reflux condenser. The
blackening began between I j o and 170' and the reaction began a t about
2.50'. The mixture was kept a t this temperature for a considerable time
and then the bath was removed and the heating completed with a luminous
flame. The nitrile was finally distilled off. The yield was less than
usual, showing that prolonged contact of the nitrile with the heated mass
lowers the yield.
Stannous and basic ferric acetates were heated with lead thiocyanate
but the mixtures did not fuse and gave no nitrile. In .the case of the
ferric salt, a white solid was deposited in the condenser. This proved
to be ammonium carbonate.
In several experiments, the issuing gases were measured and examined.
KO hydrogen sulfide was found. ,4 small amount of carbon monoxide
was found. The amount of carbon dioxide corresponded closely t o that
calculated from the reaction given below.
2128 G. D. VAN EPPS AND E. EMMET REID.

Conclusions.
I. Aliphatic nitriles, as well as aromatic, are readily prepared, with
good yields, by distilling mixtures of well dried heavy metal salts of the
acids with slight excess of lead, or other thiocyanate.
2 . ,411 things considered, the best mixture is the zinc salt of the organic
acid with lead thiocyanate.
3. The reaction takes place, in the main, according t o
(RC00)zZn +
Pb (CNS)2+ 2RCN PbS ZnS + 2C02. + +
4. Excess of the thiocyanate and use of vacuum do not materially
increase the yields.
5 . Preparation of Acetonitrile.-A mixture of one part well dried zinc
acetate and two parts, by weight, of lead thiocyanate is heated in a plain
glass retort till no more distillate comes over. The crude distillate is
mixed with half its volume of water, saturated with solid potassium
carbonate or ammonia gas and the nitrile separated and fractioned.
BALTIMORE,
MD.

[CONTRIBUTION
FROM THE LABORATORY
CHEMICAL OF JOHNSHOPKINSUNIVERSITY.
]
STUDIES IN THE PREPARATION OF NITRILES. 111. THE CATA-
LYTIC PREPARATION OF NITRILES.
BY G. D. VAN EPPS AND E. EMMET REID.
Received August 2, 1916.
Introduction.
Numerous experiments have recently been described, in which sub-
stances have been made t o react, with elimination of water, by passing
them, in vapor form, over certain metallic oxides, notably alumina and
thoria. Thus Senderens' prepared ethers by passing vapors of the
alcohols over alumina a t 240-260'. Sabatier and Mailhe2 found that
alcohols yield ethylene hydrocarbons and water with thoria and alumina
a t 300-350'~ and that various phenols3 when mixed with methyl alcohol
give the methyl ethers of the phenols with thoiia a t 390-420°, or the di-
phenyl ethers4 when the alcohol is left om. The same authors6 found
that, when a mixture of the vapors of an acid and an alcohol is passed over
titania or thoria a t 280°, esterification takes place with great rapidity,
the limit of esterification being reached in minutes instead of the hours
required a t 154' or 200°, without catalyst.
Since the formation of a nitrile from the arid and ammonia
CHsCOOH +
NHa + C H X N ~Hz0 +
1 Compt. rend., 148,227-8 (1909)
Ibid., 150,823-6 (1910).
Ibid., 151, 359-62 ( 1 9 1 0 ) .
Ibid., 151,492-4 ( 1 9 1 0 ) .
Ibid., 152, 494-497, 1044-47 ( I ~ I I ) ,Chem. Ztg., 35, 193 ( 1 9 1 1 ) .