Can. J. Chem. Downloaded from www.nrcresearchpress.com by 100.34.14.

246 on 12/06/14
For personal use only.

PREPARATION OF KETENE FROM GLACIAL ACETIC ACID.

METHYL ACETATE, AND ETHYL ACETATE1

ABSTRACT
Laboratory studies were carried out to determine whether nlethyl and ethyl
acetates could be pyrolyzed to give yields and conversions to ketene comparable to
those obtainable from acetic acid. Acetic acid was pyrolyzed to ketene and water
in the presence of 0.33% triethyl phosphate or phosphoric acid catalyst a t 680"760C. and 140-160 mm. H g absolute. Maxirnum yield obtained was 80.6y0
and maximum conversion 34.5%. Methyl acetate was cracked to ketene and
methanol a t 610-6850C. and 95-400 mm. absolute, using triethyl phosphate,
phosphoric acid, sulphuric acid, dimethyl sulphate, butyl borate, and silica gelboric acid a s catalysts. High yields could be obtained only with phosphates in
2.3% concentration. Maximum yield to ketene, including keterle recovered as
acetic acid, was 76% (conversion 31-34Yo). lklaximunl yield to nlethanol was
81.3% (conversion 33-34%). Methyl acetate containing 18% methanol was
also cracked but results were poorer. Ethyl acetate, in the presence of 0.33%
triethyl phosphate, was pyrolyzed to ketene, ethylene, and water a t 630"-685C.
and 140-190 mm. absolute. Maximum yield of ketene, includi~ig:ketene recovered as acetic acid, was 85.8y0. Maxinlum conversion to ketene was 35.4%
and to ketene plus acetic acid 85.7%. Yields to ethylene were nearly quantitative.
INTRODUCTION

The production of ketene from acetic acid and its conversion to acetic anhydride for the manufacture of cellulose acetate have been carried on industrially for many years. However, to determine whether high yields and
conversions to ketene could also be obtained from esters of acetic acid, investigations were carried out on methyl acetate and ethyl acetate, as well as
on acetic acid.
The preparation of ketene from acetic acid has already been described
somewhat extensively in the literature but was repeated here so that yields,
conversions, catalyst consumption, etc. could be compared with those obtained with methyl acetate and ethyl acetate. Industrially, ketene is produced
with a yield of 90% and a conversion of 75% by passing acetic acid, together
with 0.2% by weight of triethyl phosphate, through a stainless steel tube a t
700"-730C. and 100-450 mm. absolute (5). Our experiments were carried out
under essentially these conditions, and the ketene was passed into glacial
acetic acid t o convert it into acetic anhydride. When satisfactory yields were
obtained with acetic acid, the investigation was continued with methyl
acetate and ethyl acetate.
The pyrolysis of acetate esters to ketene has already been described.
Wilsmore (9) discovered that ethyl acetate could be pyrolyzed to ketene by a
heated wire, but his apparatus gave only low yields. Hurd (2) pyrolyzed
ethyl acetate in a vertical glass tube and obtained chiefly acetic acid and
ethylene but he also obtained traces of ketene. Loder (6) described the pyrolysis of esters of saturated aliphatic carboxylic acids in a heated copper tube
to ketenes and the corresponding alcol~ols. H e claimed a ketene yield of
Manuscript received September 19, 1951.
Contrihtio~tfrom the Research Laboratories, S~zawiniganC~zenzicalsLinzited, Shuwinigan
Falls, Que.

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 100.34.14.246 on 12/06/14

For personal use only.

87.8% from methyl acetate using silica gel impregnated with boric acid as
catalyst, but he gave no indication either of his methanol yield or his conversions. However, since this was the only prior disclosure of high yields of
ketene from acetate esters, it was decided to first try out Loder's silica gelboric acid catalyst. We were able to obtain only very low yields and conversions. In any case, it was obvious from these preliminary experi~nents
that the catalyst would probably require frequent regeneration and the regeneration was not likely to be easily accon~plished. To avoid this difficulty,
it was decided to try to find a volatile catalyst which could be added to the
feed. Triethyl phosphate and 85% orthophosphoric acid proved to be
satisfactory.
Loder (6) also indicated that ethyl acetate pyrolyzed to ketene and ethanol,
but in this investigation ketene, ethylene, and water were obtained. Since
completion of this work,* a patent (7) has been issued describing the use of
a mixture of triethyl phosphate and pyridine as catalyst and giving the
products as ketene, ethylene, and water. Optimum conditions disclosed were
essentially the same as those found in this investigation, but the yields and
conversions claimed were higher.
EXPERIMENTAL

Pyrolysis of Acetic Acid and Ethyl Acetate

The same apparatus, procedure, catalyst, feed rate, and absolute pressure
were used for acetic acid and ethyl acetate. Both materials used were manufactured by Shawinigan Chemicals Ltd. and analyzed a t least 99.5y0. The
apparatus is shown in Fig. 1. Acetic acid containing 0.33y0 by weight of
triethyl phosphate was fed a t a constant rate from a feed bottle into a stainless
steel pyrolysis tube heated by an electric furnace. The products of reaction
(i.e. ketene, water, and permanent gases) and the unconverted acetic acid
were passed through one water-cooled and two brine-cooled condensers to
condense out the water, acetic acid, and some acetic anhydride formed by
the interaction of ketene with acetic acid. The remaining ketene then passed
into two fritted glass scrubbers containing acetic acid where it was quantitatively converted to acetic anhydride. The residual permanent gas was scrubbed with 20y0 potassium hydroxide solution, dried with calcium chloride, and
exhausted by two Presso-Vac pumps operating in parallel. Its volume was
then measured by a wet-test meter. Since the reaction of ketene and acetic
acid is highly exothermic, the rate of production of ketene was followed by
observing the temperature of the cooling bath required to maintain any given
temperature within the acetic acid scrubber. The length of most runs was
five and one-half to six hours.
A constant feed of acetic acid, which is absolutely essential for good operation, was obtained by maintaining a near constant head of liquid above a
glass capillary of the proper size. Some difficulty was encountered owing to
plugging of the capillary, but this was largely overcome by introducing the
*Experirne?~talwork o n efhyl acetate was completed irz October 1945.

C.4A7ADIAN J O U R N A L 01; CHEBIISTRY.

-AUXILIARY

1'OL. 30

FEED BOTTLE

-FRITTED-GLASS

FILTER

XI T GAS
TO FUME CUPBOARD:
STAINLESS STEEL COOLER

C A L C I U M CHLORIDE
DRYING TOWER
-8

GRADUATED FEED BOTTLE

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 100.34.14.246 on 12/06/14

For personal use only.

RUBBER TUBING

{MANOMETER

FIG. 1.

YI I

'----FURNACE

Apparatus for the preparation of ketene from acetic acid or ethyl acetate.

acid into the feed bottle through a fritted glass filter and by keeping in readiness
a stand-by capillary.
T h e pyrolysis tube, which was heated by electrical resistance elements,
consisted of an S-shaped stainless steel tube, 13/32 in. I.D. and 50 in. long,
to which was welded an 8 in. length of 1 in. pipe containing a
in. O.D.
thermocouple well. The temperature was measured by means of a thermocouple connected to a Micromax recorder, and was varied by adjusting the
voltage to the resistance elements with a Variac transformer. A slow gradual
deposition of carbon occurred in the tube, and after a few weeks of operation
it was necessary t o burn i t out with a stream of air.
Three runs with acetic acid were also made in the methyl acetate cracking
apparatus. In two runs, 0.33% triethyl phosphate was used as catalyst and
in the other 0.4% orthophosphoric acid (chemically pure, 85% grade). Optimum yield conditions were not attained. T h e results are given in Table VI.
Pyrolysis of ~ W e t h y Acetate
l
Methyl acetate of 98-99% purity was obtained by acetylation of chemically
pure methanol with acetic anhydride. Operation was the same as in the
pyrolysis of acetic acid but the apparatus was slightly modified. A larger
diameter reaction tube
in. I.D.) was used so that a lower pressure drop
through the tube would result which would make possible lower absolute
pressures. The desired reaction pressure was obtained by properly throttling
a bronze globe vaIve inserted between the reaction tube and the first condenser. T h e condenser system was modified t o permit dry ice cooling. The
first condenser was lengthened from G in. to 30 in., the first 18 in. being cooled
with brine a t -5C. and the remaining 12 in. with a dry ice-methanol mixture
a t -30C. T h e second condenser consisted of a glass U-tube with a bulb

(s

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 100.34.14.246 on 12/06/14

For personal use only.

appropriately attached to store the condensate, and was cooled by a dry

ice-methanol bath a t -35OC. The third condenser was a spiral copper tube
with an attached condensate storage bulb and was cooled by a bath a t -50
to -55C.
Pyrolyses were carried out without catalyst, with one solid catalyst, i.e.,
silica gel supporting boric acid (6), and with the following conlpounds dissolved in the methyl acetate feed: triethyl phosphate (redistilled Tennessee
Eastman Corporation material), orthophosphoric acid (chemically pure, 85y0),
sulphuric acid (chemically pure, 96%), fuming sulphuric acicl, dimethyl
sulphate (Eastman Kodak, practical grade), butyl borate (Eastrnan Kodak),
hyclrocl~loricacid (chemically pure, 20% acid in water), boron fluoricle - ethyl
ether complex (redistilled, Eastman Kodak, practical grade), and pyridine
(redistilled Barrett's Grade 2A).

Pyrolysis of Methyl Acetate -Methanol Azeotropic Mixture

The same apparatus and general procedure were used as for pure nlethyl
acetate. Only triethyl phosphate and 85y0 orthophosphoric acid were used
as catalysts.
ANALYTICAL

Pyrolysis of Acetic Acid

The condensate from the first and second condensers consisted of a mixture
of acetic acid and water. The acetic acid was determined by titration with
standard X N sodium hydroxide solution using phenolphthalein as indicator.
The condensate from the third condenser, a s well as the scrubber product, consisted chiefly of acetic anhydride and acetic acid with only traces of highboiling tars. The acetic anhydride was determined by the "aniline method"
(1) and the acetic acid by difference.
Pyrolysis of Methyl Acetate
The constituents of the condensate from the first and second condensers
were determined by standard methods a s follows:Methanol: by acetylation with acetic anhydride-pyridine mixture.
Acetic acid: by titration with standard aqueous X N sodium hydroxide
solution.
Methyl acetate: by saponification with an excess of standard aqueous X N
sodium hydroxide solution followed by back-titration with standard X N
hydrochloric acicl.
Water: by titration with Karl Fischer reagent.
T h e condensate from the third condenser, as well as the scrubber product,
consisted of acetic anhydride, acetic acid, and nlethyl acetate. The anhydride
was determined by the "aniline method" (1) and the acetic acid by calculation
from the difference between the "total acidity" titer and twice the "aniline"
titer. (A negligible percentage of the nlethyl acetate was founcl to hg~clrolyze
when the anhydride was hydrolyzed with water alone a t room temperature.)
The sample, which had already been used to obtain the "total acidity" titer,

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 100.34.14.246 on 12/06/14

For personal use only.

was afterwards used to obtain the percentage of methyl acetate by saponification with an excess of standard alkali.

Pyrolysis of Ethyl Acetate

The scrubber product and condensate were analyzed a s in the previo~ls
investigations on acetic acid and methyl acetate. The permanent gases were
analyzed using a standard Burrell type apparat~lsand standard reagents.
Carbon dioxide was determined by absorption in 30% potassium 11j.droxide
solution; total in saturates by absorption in fuming sulph~iricacid; oxygen
by absorption in alkaline pyrogallol; nitrogen by multiplying the percentage
of oxygen by four; carbon nlonoxide by absorption in acidic cuprous chloride;
and hydrogen by oxidation with copper oxide. Acetylene in the unsaturates
was determined in a separate sample by absorption in alkaline silver nitrate
followed by estimation of the nitric acid formed (8).
RESULTS

Pyrolysis of Acetic Acid

The results, using 0.33% by weight of triethyl phosphate in the feed and
an absolute pressure of 140-160 nun. Hg a t the inlet end of the pyrolysis
tube are plotted in Fig. 2. I t will be noted that on increasing the feed rate
from 190 f 10 gm. per hour t o 240 f 10, the rnaximunl yield increased
somewhat, but that on further increasing i t to 315 f 10, the yield remained
approxinlately constant while the conversion decreased considerably, thus

CONVERSIONS

F E E D RATES
0

190f10 GM.PER HR.

0 240f 10 GM.PLR HR.

315 "0 GM.PER HR.

I
670

680

690 700
710
720 730 740 750
TEMPERATURE OF P Y R O L Y S I S ( ' C J

I
760

FIG.2. Pyrolysis of acetic acid: effect of temperature on yield and conversion to ketene
at various feed rates.

GEORGZEFF: K E T E N E

337

indicating t h a t the optimum capacity of t h e tube had been exceeded. T h e

maximum yieldand conversion were thus obtained a t a feed rate of 240 f 10 gnl.
per hour and a temperature of 724"C., and were 80.6% and 34.5% respectively. Of the 19.4y0 loss in yield, 14.4% was d u e t o the formation of permanent gases a n d 5.0% t o the sweeping o u t of acetic acid and anhydride
from the scrubbers by t h e permanent gases. T h e percentage of the acid
which decomposed t o permanent gases increased with increase in temperature.

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 100.34.14.246 on 12/06/14

For personal use only.

Pyrolysis of Methyl Acetate

T h e p~.rolysisof methyl acetate without a catalyst produced extremely low
yields of ketene and methanol. Of the potential catalysts tried, only triethyl
phosphate a n d orthophosphoric acid were found entirely satisfactory.
Triethyl Phosphate Catalyst
T h e maximum yields were obtained a t 670 f 10C. and 150-250 inm. H g
absolute, a n d were a s follows:t o ketene, including ketene recovered a s
acetic acid, 76.2y0 a t a conversion of 31.3y0; t o methanol, 81.3% a t a conversion of 33.4y0. A feed r a t e of 6.2 gm. per minute was used. About 2.5
t o 3.0% of phosphate catalyst, based on t h e weight of methyl acetate, was
required t o give the maximum conversion. I t was also necessary t o a d d 5-6yo
water t o the methyl acetate feed stock.
No a t t e m p t was made t o determine systematically the optimum feed r a t e
for this larger tube, a s had been done with acetic acid for a somewhat smaller
tube, since optimum yields were obtained over a fairly wide range of feed rate.
I t was merely assumed t h a t the new optimum feed rate, in moles per minute,
would be directly proportional t o the ratio of the area of the heated (inside)
surface of the larger tube t o the smaller, i.e.,
507 sq. cm./412 sq. cm. X 4/60 = 0.0820 moles per minute
= 6.08 gm. methyl acetate per minute.
T h e validity of this assumption was substantially confirmed b y the experimental results. However, i t was necessary t o determine t h e effect of t h e
following variables on the yield and conversion:(i) E f e c t of Addition of Water to the Feed

I t had already been reported (6) t h a t if up t o 870 water is added t o t h e

feed, some of the ketene which would otherwise go t o valueless by-products
would be recovered a s acetic acid. This has been confirmed. T h e addition
of 5 t o 6yo water did not increase the yield or conversion to ketene, a s such,
b u t did increase the yield a n d conversion t o ketene plus acetic acid for a n y
given concentration of catalyst in t h e feed. Some improvement in the methanol yield was also noted. T h e presence of the water tended t o depress the
formation of both permanent gas a n d paraformaldehyde. T h e experimental
results are summarized in Table I.
(ii) E f e c t of Per cent Catalyst in Feed o n Per cent Conversion
Under conditions of constant contact time a n d constant temperature in
the pyrolysis tube, i t was found t h a t the conversions t o ketene, ketene plus
acetic acid, a n d methanol increased rapidly with increase in the concentra-

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 100.34.14.246 on 12/06/14

For personal use only.

TABLE I

?!
% Conversion

% Yield
%

IZurl
No.

Et3PO4
in
feed

Feed
rate,
sm./
min.

Hz0
in
feed

A1 2
M 3
M13
M 6
ill14

0.33
0.33
0.33
0.67
0.7

6.1
5.1
6.0
6.5
5.4

0.7*
0.7*
5.2
0.6*
5.5

Temp.
of
pyrolysis,
"C.
--632 f 10
610 f 10
660 f 15
626 f 5
654 f 10

Abs.
press.,
mrn.
Hs
94
96
92
106
93

To
ketene
alone
-38.3
37.4
38.6
51.2
52.8

To
ketene
and
HOAc
40.5
38.0
45.3
52.2
57.2

II

I
To
CH3OH
-41.2
44.1
47.4
60.0
65.6

To
To
ketene
lcete~le
and
To
alone
HOAc
CH3OH
-- ----9.4
10.3
10.4
ll.G
9.7
10.0
12.1
8.8
11.6
14.9
13.0
12.5
20.2
14.8
17.5

P
A
-

2
8

E
*
%

>

9
25

.-

a
Y

r:

2
&

*A70

nualer added.

tion of triethyl p h ~ s ~ ~ l iup

a t eto
Results are shown in Fig. 3.

2y0,but

then leveled off between 2 to 3%.

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 100.34.14.246 on 12/06/14

For personal use only.

(iii) Effect of Pressure o n Conversion

The per cent conversion was found to be almost independent of the pressure
between 95 mm. Hg absolute and 325 mrn. and a t optimunl temperature, as
can be seen from Fig. 4. The highest yields and conversions, using 2% t r i e t h ~ l
phosphate as catalyst, were obtained a t about 150 to 250 mm. Hg absolute.
Above 325 mm., yields decreased rather rapidly and conversions decreased
somewhat less rapidly, owing to the formation of larger percentages of per-

80

70

YIELDS

I- 60-

z
W

cZ 50-

e
40

CONVERSIONS

Y I E L D S AND

CONVERSIONS

TO KETENE
o TO K E T E N E C H ~ C O O H
0 TO METHANOL

0.0

I
I
I
I
1.5
2.0
2.5
3.0
3.5
P E R C E N T TRIETHYL PHOSPHATE CATALYST
I

0.5

1.0

FIG.3. Pyr01ysis of methyl acetate: effect of percentage of triethyl phosphate catalyst

in feed on yield and conversion. Temperature = 663 Z!Z 10C.; absolute pressure = 230270 mm. Hg; feed rate = 6 gm. per minute; per cent water in feed = 5.3.

manent gases. The optimum temperature apparently decreased from

662 f 5C. a t 95 mm. absolute to 626 f 10C. a t 402 mm.

Orthophosphoric A c i d a s Catalyst
Yields and conversions with orthophosphoric acid (85%) were approximately the same as those obtained with triethyl phosphate for the sanle
percentage, by weight, of catalyst in the feed, Some difficulty was experienced
in starting up the runs, but once started, operation was smooth. The general
composition of the condensate, as well as the by-products, e.g., paraformalde-

hyde, were also similar. Table I1 gives a comparison between triethyl phosphate and orthophosphoric acid as catalysts.
Other Catalysts
Sulphuric acid (96%) and fuming sulphuric acid also catalyzed the reaction
to ketene and methanol but the highest yields obtained were only 4540%
and the highest conversions 8-l0Y0. Increasing the percentage of catalyst
above 0.67y0 did not appear to improve the conversion. Sulphuric acid
appeared to be easily reduced, as evidenced by the strong odor of hydrogen
sulphide and organic sulphur compounds in the condensate.
Dimethyl sulphate had only a slight catalytic effect. .
Can. J. Chem. Downloaded from www.nrcresearchpress.com by 100.34.14.246 on 12/06/14
For personal use only.

YIELDS A N D CONVERSIONS
TO K E T E N E
TO KETENE + C H ~ C O O H
0 TO M E T H A N O L

m
YIELDS

40k

50

FIG.4.
sioris.

CONVERSIONS

100
150
200
250
300
350
400
450
ABSOLUTE PRESSURE IN MM. HG AT INLET OF TUBE

P y r o l y s i s of m e t h y l a c e t a t e : e f f e c t
P e r c e n t t r i e t h y l p h o s p h a t e in f e e d = 2.0.

of

a b s o l u t e p r e s s u r e 011 y i e l d s

and

conver-

Butyl borate was also a poor catalyst. I t was better than dimethyl sulphate but poorer than sulphuric acid. Much larger quantities of paraformaldehyde were formed with the above three catalysts than with the phosphates.
The largest quantity was formed with dimethyl sulphate.
The only solid catalyst tried was silica gel impregnated with boric acid (6).
Even with a fresh catalyst, very low yields and conversions were obtained.
Hydrochloric acid (20% solution in water), boron trifluoride-ethyl ether
complex, and pyridine were found to have no catalytic effect a t all.
Pyrolysis of Methyl Acetate - Methunol Azeotropic klixture
Methyl acetate containing 18% methanol was also pyrolyzed, but conversions
were only 38% to 56% of those obtained with pure methyl acetate for the

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 100.34.14.246 on 12/06/14

For personal use only.

TABLE I1

COMPARISON
BETWEEN

PYROLYSIS
OF METHYL ACETATE
TRIETHYL PHOSPIIATE AND PHOSPHORIC ACID AS CATALYSTS
-

Run
NO.

EtaP01

HaPo,
(85%)

M 40
M29
M 41
M 16
M 42
M 22

0.0
0.70
0.0
1.33
0.0
2.0

0.67
0.0
1.33
0.0
2.0
0.0

Feed
rate,
gm./
rnin.

Hz0
r in
feed

Temp.
of
pyrolysis,
"C.

6.0
6.1
5.7
5.6
5.3
6.1

5.2
5.2
5.2
5.6
5.0
5.1

660 f 10
658 f 5
666 f 5
660 f 10
644 f 10
638 =t10

% Yield

% Conversion

To
ketene
alone

To
ketene
and
HOAc

11.8
11.9
15.0
18.5
19.4
21.0

15.9
16.2
21.6
22.5
28.1
27.7

Abs.
press.,
mm.
Hg

To
ketene
alone

To
ketene
and
HOAc

228
275
242
94
232
198

48.1
48.7
52.9
58.2
68.3
64.0

55.7
55.4
61.8
62.8
75.6
70.2

--

To
CHaOH

72.0
70.2
73.6
76.8
75.5
78.2

To
CHDH

20.5
20.4
25.7
24.5
28.0
30.8

th
7
h

2
y

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 100.34.14.246 on 12/06/14

For personal use only.

342

C A N A D I A N J O U R N A L OF C H E d l i S T i 7 Y .

VOL. 30

same percentage of triethyl phosphate or ortl~ophospl~oric

acid in the feed,
all other conditions being essentially the same. (For details see Table 111.)
Increasing the percentage of catalyst increased the conversion, b u t conversio~ls
as high as the maximum obtained with pure methyl acetate could not be
obtained with up'to 5% catalyst. When the dry azeotrope was cracked, very
large amounts of paraformaldehyde and permanent gases were formed.
T h e very much lower conversions obtained with the azeotrope as compared
with pure methyl acetate for any given percentage of catalyst, would appear
to be due t o a t least three factors. Firstly, the feed stock already contains
33.7 mole yoof methanol and this exerts an unfavorable influence on the final
equilibrium. Secondly, methanol appears t o react with phosphate catalysts
reducing them to catalytically ineffective compounds. Hence, the higher
concentration of methanol present when the azeotrope is cracked is responsible
for the higher coilcentration of catalyst required to produce any given conversion. (Further details are given in the section on "Side-Reactions".) And
thirdly, there is probably a greater loss of ketene by recombination with
methanol in the condenser system.
Side Reactions
T h e major loss in yield in the pyrolysis of methyl acetate and methyl
acetate- methanol azeotropic mixture was due to the formation of permanent
gases, e.g,, carbon monoxide, hydrogen, ethylene, etc. T h e average nlolecular
weight of the gas was determined for many runs by dividing the weight of
methyl acetate not recovered a s liquid or solid by the total number of molecular
volunles of gas formed. Molecular weights varied from run to run, b u t most
values fell between 25 and 28 with an over-all average of 26.6.
Other important side reactions were the reduction of the phosphate catalysts
t o catalytically ineffective compounds and the oxidation of methanol t o formaldehyde with its subsequent polymerization to paraformaldehyde. Several
factors indicate that the methanol was the chief reducing agent of phosphate.
Firstly, acetic acid pyrolyzes to ketene and water, and ethyl acetate to ketene,
ethylene, and water and both require only 0.2-0.33% phosphate to give the
maximum conversion. On the other hand methyl acetate which pyrolyzes to
ketene and methanol requires 2.5-3.0yo. Hence, since ketene and ester are
not responsible for the higher consumption of catalyst, methanol probably is.
Secondly, since a much higher concentration of phosphate is required to
bring about the maximum or near maximum conversion with methyl acetatemethanol azeotrope than with pure methyl acetate, i t would appear t h a t the
higher concentration of methanol present in the former is responsible for the
higher consumption of catalyst. Thirdly, the formation of significantly large
quantities of paraformaldehyde and water, along with the disappearance of
phosphate (as described below), add further weight to the hypothesis of
simultaneous oxidation of methanol and reduction of phosphate.
Some paraformaldehyde and water (over and above that added to the feed)
were obtained in the condensate in all runs, including those in which catalysts
other than phosphates were used. T h e paraformaldehyde was identified by

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 100.34.14.246 on 12/06/14

For personal use only.

GEORGIEFF: K E T E X E

2
z

a,

2-0

,a

$ 6
V)

2 2
Ez$z;g
*

w o w o w w
a o a a , a a
o w o w o o
N LS. U
0
NL S
. aN) aNI

1 <&<LC<

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 100.34.14.246 on 12/06/14

For personal use only.

344

C A N A D I A N IOL'RNAL OF C H E M I S T R Y . I'OL. 30

treating it with dilute sulphuric acid to convert it to formaldehyde and then

applying the standard qualitative tests, i.e., fuchsin, chron~otro~ic
acid,
resorcinol condensation, and odor. I t was noted that the addition of water
to the feed reduced the amount of paraformaldehyde formed, and increased
the yield of methanol somewhat. I t was also noted that the amount of water
recovered in the condensate, over and above that added to the feed, increased
somewhat with increase in conversion.
Ailalyses of condensates of two runs for phosphate by precipitation as inagnesium ammonium phosphate and ignition to magnesium pyrophosphate (4)
showed only 0.20% to 0.24% phosphate, calculated as phosphoric acid even
though 2.5 to 4y0 phosphate was used in the feed. This indicates that the
phosphate phosphorus may have been reduced to highly volatile compounds,
perhaps to its hydrides or perhaps even to elemental pl~osphorus. In the
pyrolysis of acetic acid, when the first condenser was talcen down for cleaning,
small amounts of what may have been elemental phosphorus were observed
to ignite with a bright flash when given a sharp blow with a metallic object.
Pyrolysis of Ethyl Acetate
Ethyl acetate was pyrolyzed to ketene, ethylene, and water in the presence
of 0.33y0 by weight of triethyl phosphate. The maximum yield to ketene,
including lcetene recovered as acetic acid, was obtained a t 644* 5C. and
154 mm. Hg absolute, and was 85.8%. The maximunl conversion to ketene
was 35.4y0 and to ketene plus acetic acid 85.770. A feed rate of 4.2 gm. per
minute was used and all the ethyl acetate was cracked in a single pass through
the tube a t temperatures above 628C. The yield of ethylene was determined
for only Run No. E7 and was found to be 104% of theory, assuming that all
the ethylene was derived from the ethyl radical. However, it is known that
ketene also decomposes t o ethylene, carbon monoxide, and other gases, and
if it is assumed that all the acetate radical which is not recovered as ketene
or acetic acid is converted quantitatively to ethylene, carbon monoxide, and
water according to the equations,

//

2CH3 - C - OC2Hs -+2CH2

+ 2H20 + 2C2H4,

TABLE IV

feed
in

Et3P04

Run
No.
E4
E5
E6
E7

Feed
rate,
mln.
gm-'

H2 0
in
feed

'

Temp.

pyrolysis,
OC.

'%;1

97, Conversion
Abs.
press.,

1
1

% Yield to
ketene and
acetic acid

kz:ne
alone

To
btene
and
HOAc

GEORGZEFF: K E T E N E

3-15

then the yield to ethylene from the ethyl radical becoilles 90%. Of the 10%
loss, about j.3y0 was recovered as acetylene. A summary of the results is
given in Tables IV and V.
TABLE V

PERMANENT
GASES

Spot gas analysis*

Run
No.

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 100.34.14.246 on 12/06/14

For personal use only.

FORMED I N THE PYROLYSIS OF ETHYL ACETATE

%6H:

% CO

(
I

%HZ

%OI

%N?

* T l ~ egas analysis given for R u n No. E7 i s tlze average for the nohole run. All others are only
typical analyses and are not necessarily the average.

DISCUSSION

The highest yield of ketene obtained from acetic acid in this investigation
was comparable to that obtained industrially, but the highest conversion
was a little less than half as great. The low conversion would appear to be
due to recombination of ketene and water to acetic acid and acetic anhydride.
I11 several experiments, the presence of a substantial quantity of acetic anhydride in the aqueous condensate from the first and second condensers was
established by cooling the condensate to -lOC. to retard the hydrolysis of
the anhydride until it could be determined qualltitatively by the "aniline"
method (1). Results are shown in Table VI. Since the anhydride in the
aqueous condensate could not be easily isolated by distillation it was considered, for yield purposes, as recovered acetic acid, and in most runs was
allowed to hydrolyze to acetic acid. Further evidence of recombination of
lcetene and water was obtained in another investigation on a much larger
scale, details of which are not given here. Yields of 90-92% and coilversions
of 70-72y0 were obtained by using linear velocities 10 to 12 times those used
in this investigation, and by adding a catalyst neutralizer (gaseous ammonia)
a t the inlet or outlet of the pyrolysis tube. The higher linear velocity permitted a much more rapid separation of ketene and water in the condenser
system, and the ammonia probably neutralized the phosphate catalyst which
otherwise would have catalyzed the hydrolysis of ketene. In the present
small-scale investigation no attempt was made to add any neutralizer owing
to the difficulty of metering in such a small flow.
The highest conversion obtained in the pyrolysis of methyl acetate was
also much lower than that obtained in the industrial pyrolysis of acetic acid
but was of the same order a s that obtained in our laboratory pyrolysis of
acetic acid. I t would appear probable that the conversion could be greatly
increased by employing the same expedients as with acetic acid.
In the pyrolysis of methyl acetate - methanol azeotropic mixture, conversions
were lower and catalyst consumption was much higher than for pure methyl

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 100.34.14.246 on 12/06/14

For personal use only.

6
2

TABLE VI

"
,

Ir

%
Run
No.

Catalyst
used

Catalyst
in Feed

Feed
rate,
gm, per
min.

Temp.
of
pyrolysis,
"C.

Abs.
press.,
mm.
Hg

% Conversion to ketene

yield
to
ketene

Conv.
to
ketene

if AcnO in first and

second condensates is
included

NOTE:-The three above experinrents (only) were carried out in the metlryl acetate apparatzrs.

3c

17
L

2
r\

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 100.34.14.246 on 12/06/14

For personal use only.

CEORCIEFF: K E T E N E

347

acetate. I t would thus appear advisable to obtain first substantially pure

rnethyl acetate from the azeotropic mixture.
In the pyrolysis of ethyl acetate, much higher conversions to ketene than
those obtained in this investigation are also undoubtedly possible. I n fact,
conversions of 85% and quantitative yields have already been claimed in the
patent literature (7). T h e optirnurn temperature of pyrolysis was found t o
be about 60" t o 80C. lower than for pure acetic acid, which would indicate
that the primary products of the pyrolysis are not acetic acid and ethylene, followed by dehydration of the acetic acid. Instead it would seem
probable that the pyrolysis is exactly analogous t o that of methyl acetate but
that, in this case, the alcohol (ethanol) is unstable under the conditions of
pyrolysis and dehydrates to ethylene and water, according to the equations,
0

//

+
+

CHI - C - OCzH5 -+ CHz = C = 0

CzHsOH
+CH2 = C = 0
CzH4
HzO.
Some of the ketene then recombines with water giving a mixture of ketene
and acetic acid. No ethanol was isolated, which is not surprising, since it is
well Icnown that ethanol dehydrates in the presence of catalysts such as alkali
and alkaline-earth metal phosphates containing free phosphoric acid a t
temperatures of 480C. (3).

ACKNOWLEDGMENT

The author wishes t o thank Dr. W. E. Barker for his collaboration on the
work on acetic acid, and also Shawinigan Chemicals Limited for permission
to publish this paper.
1.
2.
3.
4.
5.
6.
7.
8.
9.

REFERENCES
W. D. et al. Ind. Eng. Chem. 18: 636. 1926.
COLLINS,
F. H. J. Am. Chem. Soc. 60: 2419. 1938.
HURD,C. D. and BLUNK,
I. G. FARBENINDUSTR~E
A.-G. Brit. Patent 326,185. 1930.
KOLTHOFF,
I. M. and SANDELL,
E. B. Textbook of quantitative inorganic analysis.
The Macmillan Co., New York. 1936. p. 367.
LOASBY,
G. and BROWN,
F. S. Brit. Ministry of Supply and U.S. Tech. Ind. Intelligence
Comm. 'FIAT Final Rept. No. 145. 1945.
LODER,D. J. U.S. Patent 2,175,811. 1939. Brit. Patent 504,626. 1939.
LONZAELEKTRIZIT;~TSWERKE
and CHEMISCHE
FABRIKEN
A.-G. Swiss Patent 233,176.
1944. Chern. Abstracts, 43: 6643. 1949.
Ross, W. H. and TRUMBULL,
H. L. J. Am. Chem. Soc. 41: 1180. 1919.
\ ? T 1 ~ ~N.~T.O M.
~ ~ J., Chem. SOC.91: 1938. 1907.