THE KINETICS OF THE OXIDATION OF GASEOUS PROPIONALDEHY D E E. W. R. STEACIE, W. H. HATCHER,
AND
S. ROSENBERG
Laboratory o j Physical Chemistry, McGill University, 31ontrea1, Canada
Received April 26, 1936 INTRODCCTION
Aldehydes are of considerable importance in the mechanism of the oxidation of most gaseous organic compounds. As a result a considerable amount of work has been done on the oxidation of gaseous acetaldehyde (1, 3, 4). The acetaldehyde oxidation, however, is complicated and the mechanism is by no means clear. It is therefore of interest to investigate the oxidation of other aliphatic aldehydes where some of these complications might be absent. The present paper deals with the kinetics of the oxidation of gaseous propionaldehyde. EXPERIMEYTAL
The reaction was investigated by introducing propionaldehyde and oxygen separately into a heated vessel and observing the rate of change of pressure as the reaction proceeded. The apparatus was similar to that used in a previous investigation (3). The required temperatures were obtained by means of an electric furnace, the reaction vessel, a 200-cc. Pyrex bulb, being packed into the furnace with coarse iron filings to ensure uniformity of temperature. The temperature was measured with a standardized mercury thermometer, and could be maintained constant to within 1°C. Propionaldehyde from three sources was used, viz., Kahlbaum, Eastman Kodak, and Eimer and -4mend. In each case it was purified by fractional distillation. The three samples gave identical results. Commercial oxygen was obtained from cylinders. It contained 6.4 per cent nitrogen, and was purified merely by drying with phosphorus pentoxide. EXPERTRIEKTAL RESULTS
The reaction was investigated from 120°C. to 170°C. E o appreciable effect due to the aging of the reaction vessel was noticed after the first ten or fifteen runs. It u as also found that neither the degree of evacuation nor the temperature a t which this was carried out had any appreciable effect on the succeeding run. 1189
1190
E . W. R. STEACIE, W7. H. HATCHER, A S D S. ROSENBERG
I n all runs reported here, oxygen-aldehyde mixtures were made up by introducing the reactants separately into the reaction bulb. (The order of admission had no effect.) This was done because it was found that the maximum pressure decrease associated with the reaction was diminished about 15 per cent, and the rate of reaction was diminished by about 40 per cent, if the reactants were mixed outside the reaction bulb in the presence of a mercury manometer. Tests showed the presence of peroxides under these conditions.
The pressure change accompanying the reaction As the reaction proceeded the pressure diminished. The maximum decrease amounted to about 73 per cent of the initial partial pressure of the TABLE 1 Maximum pressure decrease TEMPERATURE
XYGEN-ALDEHYDE
degrees C .
RATIC
INITIAL PARTIAL ALDEHYDE PRESSURE
M A X I M U M PRESSUR& DECREASE
cm.
per cent
150.8 150.8 150.8 150.8 150.8 150.8 150.8 150.8 150.8
2.52 1.05 4.39 1.36 1.01 1.06 0.90 0.76 1.23
20.00 20.00 12.00 12.00 8.85 9.40 8.85 8.75 5.65
73 69 74 71 70 68 63 52 63
139.1 139.1
2.86 3.54
20.00 12.00
75 71
125.3 125.3
2.87 1.96
20.00 20.00
I
74 76
aldehyde. The pressure then slowly rose again. The maximuni pressure decreases obtained in a number of typical cases are shown in table 1. As shown in table 1, the maximum pressure decrease is constant and independent of the experimental conditions, provided that (a) the partial aldehyde pressure is not much below 8 cm., and (b) the oxygen-aldehyde ratio is not much below 1-00. In as much as the pressure decrease is followed by a pressure increase, it is evident that the reaction proceeds in stages. Since in most cases the maximum decrease is constant, however, there is no doubt that it is justifiable to use the rate of pressure change as a measure of the rate of reaction. In the cases where the full pressure decrease is not reached, it is probably justifiable to employ the same
OXIDATION O F GASEOUS PROPIONALDEHYDE
1191
criterion of reaction if we consider only the eztrlypart of the process. In consequence we have used the time for the pressure to change from a 15 per cent decrease to a 30 per cent decrease as a measure of the rate of reaction.
The rate of reaction Complete data for some typical runs are given in table 2, and are plotted in figure 1. It will be seen that as is usual in gaseous oxidation reactions, there is a small induction period. Table 3 gives values for T 3 0- T15 under various conditions.
TIME FIG. 1. TYPICAL PRESSURE-TIME CURVES Temperature, 150.8°C. Curve 1: C*H&HO, 20.0 cm.; 0 2 ,50.4 cm. Curve 2: CzHs CH3, 9.7 cm.; 02, 38.8cm.
The efect of pressure For any given oxygen-aldehyde ratio, the effect of the total pressure on the rate corresponds to a n order of about 2.0. The effect of the oxygen concentration on the rate of the reaction is shown by figure 2, in which T 3 0- Tl5 is plotted against the oxygen-aldehyde ratio for various constant initial partial pressures of aldehyde. It will be seen that the rate is virtually independent of the oxygen concentration if the oxygen-aldehyde ratio is above 2.0. When the oxygen concentration is reduced below this value, the rate is somewhat diminished.
1192
E.
w.
R. STEACIE,
w.
H. HATCHER, AXD
s.
ROSENBERG
TABLE 2 Data for typical runs at 150.8'C. 'I Time
-AP
minutes
cm.
0 0.5 1 1.5 3 4.5 7.25 12.5 18.5 28.5 73.5 1680
0 0.12 0.28 0.45 1.34 2.05 3.24 4.65 5.70 6.50 7.28 5 05
-AP
Per cent AP
cm.
0 1 2 2 9 4.6 13 8 21 1 33 4
FIG. 2. EFFECT OF
0 0 5 1 1 5 2 3 4
0 0.60 1.80 3.85 4.60 6.65 8.15 12 00 14 50 14.85 14,96 14,60
0 3 .O 9.0 19.3 23.0 33.3 40.8 60.0 72.5 74.3 74.8 73.0
OXYGEN CONCENTR.4TIOS
Temper at ure, 150.8OC.
Figure 3 shows the effect of varying the aldehyde concentration while the oxygen concentration is kept constant. The values are given in the form of a log-log plot. From the slopes of the lines the order with respect to
1193
OXIDATIOK OF GASEOUS PROPIONALDEHYDE
aldehyde is found to be betn-een 1.8 and 2.1. Hence, provided the oxygen concentration is not too low, the rate can be approximately expressed by
-
d(C2HsCHO) dt
TABLE 3 Reaction rate data OXYGEN
- Ti5
TEMPERATURE
PARTIAL ALDEHYDE PRESSURE
degrees C.
cm.
150.8 150.8 150.8 150.8 150.8 150.8 150.8 150.8 150.8 150.8 150.8 150.8 150.8 150.8 150.8
20.00 20.00 20 .oo 12.00 12.00 12.00 12.00 8.00 8.00 8.00 8.00 8.00 4.00 4.00 4.00
2.52 1.68 1.05 4.39 4.37 2.58 1.36 5.46 5.75 3.17 2.10 1.32 5.64 3.02 1.86
85 80 110 155 156 157 184 260 260 270 256 297 502 502 546
167.4 167.4
12.00 8.00
3.00 4.00
76 128
139.1 139,l 139.1 139.1 139.1 139.1
20.00 20.00 20.00 12.00 12.00 8.00
2.86 1.62 1.62 3.54 2.42 3.20
134 144 142 280 282 508
125.3 125.3
20.00 20.00
1.96 2.87
260 266
150.8 150.8
8.00 20.00
2.58* 1.17'
320 111
ALDEHYDE
T30
seconds
* Nitrogen added in an amount equal to t h a t of the oxygen. The temperature coeficient The temperature coefficient of the reaction has been inferred from three series of runs at constant aldehyde pressures of 20, 12, and 8 cm., using
1194
E.
w.
R. STEACIE,
w.
H. HATCHER, AND
s.
ROSEXBEHG
FIG. 3 . EFFECTO F PROPIOS.41,DEHI.DE C O S C E S T R A T I O N Temperature, l5C.S0C. Curve 1, Poi = 20 em. Curve 2 . Poi = 40 em.
I/T x IO5 FIG.4. TENPERATCRE COEFFICIEST
oxygen-aldehyde ratios in the region in which the rate is almost independent of oxygen. The results are shown in figure 4, in which log
1195
OXIDATIOS OF GASEOCS PROPIONALDEHYDE
(T30 - T I S )is plotted against the reciprocal of the absolute temperature. From the slopes of the lines we obtain the following values for the apparent heat of activation: E i n calmies per gram-mole
raldehlde
15,700 16,200 14,400 Mean = 15,400
20 em. 12 cm. 8 cm.
The products of the reaction Some analyses of the gaseous products of the reaction are given in table 4. In order to refer the quantities of gaseous products back to the partial pressures which they exerted in the reaction vessel, runs were made with the TABLE 4 Thc gaseous p r o d u c t s o j the reaction I
OX'IGEKALDEHYDE RATIO
degrees c'.
I
CO1
pcr r s n t
cm.
152 153 150 151
20 20 20 20
139
30 2
I
0 0 0 0
1.20 2.05 1.16 1.35
27 9 27 17
1.60 1.54*
13 34
co
coz CO
per cent
13 5
-~
I I 1
l;
5 4
2.1 1.8 2.1 2.4 2.6 8.5
* Reaction mixture allowed t o stand a t 139°C. for 15 hours. addition of definite amounts of nitrogen. Thus t n o runs a t 150.8"C. with oxygen-aldehyde ratios of 1.16 and 1.35 gave the following results:
+ 0 980, 0 11C02 + 0 05CO + 1 00 Condensable + 0 8702 = U 13COr + 0 05CO + 0 94 Condensable
(a) 1C2H5CH0 (b) 1C2H,CH0
=
Hence, from a material balance, we find that the condensable products consist of (a) C z 84H60z.69, and (b) Cz ~2H602.4~.On chemical grounds the most probable product is propionic acid. On the basis of the carbon above, this has the formula CZ g2H5e 4 0 1 8 8 . This leaves unaccounted for (a) HO3 6 0 ~81, and (b) HO3 6 0 0 5 5 . If a small amount of complete oxidation occurred, it would result in an amount of water equal to the CO + CO,, i.e., 0.16 and 0.18 H20,respectively. This is in exact agreement m-ith the hydrogen unaccounted for above. K e still have unaccounted for (a) 006 3 ) and (b) Oo 3 7 . This can obviously be accounted for by assuming the formation of some perpropionic acid. This would have the formula T H E JOURX.4L O F P H T S I C A L C H E M I S T R Y , VOL.
38,
NO.
9
1196
E. W. R. STEACIE,
W.
H. HATCHER, A X D S. ROSENBERG
Cz,82H5.6402 If the excess oxygen were used in this way, n-e would have (a) 67 per cent perpropionic acid, and (1)) 40 per cent perpropionic acid. The overall reaction is therefore
+ 0 05CO + 0 16Hr0 + 0 56C2HjCOOOH + 0 28C?H,COOH (b) 1CgHsCHO + 0 8702 = 0 13COJ + 0 05CO + 0 18HZO + 0 30CzH,COOOH + 0 46C2HsCOOH
(a) 1C2HsCH0-k 0 9802 = 0 11CO2
The presence of acids in the condensable products was confirmed by carrying out a run a t 150.8"C. and removing the reaction bulb. The bulb was then thoroughly washed 13 ith distilled water and the washings titrated with 0.05 X sodium hydroxide. The results indicated approximately 1 molecule of acid for each molecule of aldehyde disappearing, as required by the above scheme. A similar experiment n-as then carried out, the washings being analyzed for per acids using a neutral solution of potassium iodide and titrating the liberated iodine n i t h 0.05 S sodiuni thiosulfate. Approximately 50 per cent of the total acid n-as found to be per acid, again confirming the above scheme. In addition to the above products, traces of esters were detected by their odor. I t has previously been shown ( 2 ) that ethyl propionate is a product of the decomposition of perpropionic acid. The formation of this compound would be accompanied by the formation of carbon oxides and water in equal quantities, and hence necessitates no modification in the above calculations. Hence the predominant reaction under investigation is C,H,CHO
+ 02 = CzHsCOOOH
followed by the formation of propionic acid from the per acid, or from a reaction between the per acid and the aldehyde.
The eflect of surface Runs were also made in a bulb packed with short lengths of Pyrex tubing, so as to increase tlie surface-volume ratio to about six times its former value. I t was found that the reaction process was completely altered in the packed bulb. Aging effects were noticed, but after the surface had reached a steady condition it was found that the reaction now led to a pressure increase of about 15 per cent. It therefore appears that the chain process in the enipty bulb has been wholly or partially replaced by a heterogeneous reaction. It has previously been slion-n that in the enipty liulb the pressure after its initial decrease then slowly rose again. As shown by table 4, this increase in pressure was accompanied by tlie production of a large amount of
OXIDATION O F GASEOUS PROPIOXALDEHYDE
1197
carbon dioxide. Analyses of the products from runs in the packed bulb, where there was also a pressure increase, showed the same preponderance of carbon dioxide. I t thus appears probable that the secondary process in the empty bulb is also a surface reaction.
Explosions There was no evidence of any critical explosion limits. Increased temperature and aldehyde concentration caused a progressive increase in the rate without any sign of a discontinuity. DISCUSSION
As shown above, the main reaction is the oxidation of propionaldehyde to the per acid, followed by the forination of propionic acid from the per acid. The formation of the per acid would lead to a pressure decrease of 100 per cent, while the formation of propionic acid would give 50 per cent. Both reactions must thus occur, since the observed pressure decrease is about 75 per cent, but they are not necessarily independent processes. This point will be referred to again later. There is no doubt that the main process involved is a chain reaction. This is shown by analogy with other gaseous oxidation reactions, and also by the characteristic form of the differential equation expressing the rate. The heat of activation is 15,400 calories. This is considerably lower than that corresponding to a bimolecular reaction proceeding at an equal rate in the same temperature range. This, however, is a quite common characteristic of a chain reaction, since the heat of activation is a composite one and includes the temperature coefficient of the chain length. One corninon criterion of a chain process is the suppression of the reaction by packing. I n this case little inforniation can be obtained from the results in the packed vessel, since the course of the reaction is completely altered. I t appears, however, that the deactivating effect of the walls is not very pronounced. This is shown by the reproducibility of the results, and by the fact that inert gases, including excess oxygen, have little influence on the rate. The mechanism of the reaction -4s we have seen the rate of the reaction is given by
On the basis of the Bodenstein schenie (1) for the oxidation of acetaldehyde we have A*
+
-1 0 2
= =
P* =
A* P*
(1)
+
;i*
(2) 0 2
(3)
1198
E. W. R. STEACIE, TV. H. HATCHER, AND S. ROSESBERG
+
P* A = P + A * P*+K = P +K* P* 0 2 = P 0 2 P* = P (wall)
+
+
where A represents aldehyde, P per acid, and E( a negative catalyst. In the absence of a negative catalyst, this leads to d(N
KiK4 (-4)'
-
+~ ~ ~ 5 2
~ ~ ( 0 ~ )
dt
This equation disagrees with the experimental results in that it ascribes a retarding effect to oxygen. The only way to avoid this would be to assume that K & is very much greater than K 5 ( 0 2 ) . This, however, is not here a valid assumption since it has been found that the wall effect (K6) is small. Any attempt to get over this difficulty by niaking minor modifications in the scheme is impossible. On the Bodenstein scheme the only fate of A* is to react with oxygen. Hence oxygen exerts no favorable effect on the chain propagation to balance its deactivating effect, and thus it will always appear in the denominator. T o avoid this difficulty it is necessary to make a fundamental alteration in the initiation step, as follows : A P* A* P*
+ = P* + A = P + .4* + = P* + 0, = P t 0 2
0 2
0 2
P*
=
P (wall)
The alteration made here is to abandon the assumption of independent activation of the aldehyde, and substitute the more usual idea of an initial bimolecular step. Whence n-e have
Evaluating the concentrations of unstable intermediate substances by equating their rates of formation and destruction in the steady state, we obtain
-~ d(A) dt
=
KIM) ( 0 2 1
+ KK :, (KO?s()A ) *(02) K5 +
Neglecting K g ,since the n-all effect is sniall, and dropping the first term, since the chain-starting step is negligible conipared to the chain-carrying step, we have finally
which agrees with the experimental results.
OXIDATION O F GASEOUS PROPIONALDEHYDE
1199
It has been previously mentioned that propionic acid is also formed. If this results from a secondary decomposition of the per acid, the formation of the per acid will be the rate-determining step, and the above equation will still hold. I t is, hotvever, possible that propionic acid arises from the reaction of aldehyde and per acid, as an integral part of the chain process. If so, we can account for the observed concentration relationships by the addition to the above scheme of the following steps:
+
+
P* A = S* S S * + A = d +A*
where S represents propionic acid. This leads to the final equation
which is again of the right form. It seems likely that this is the way in which propionic acid arises, since a straight decomposition of perpropionic acid does not normally yield propionic acid (2).
T h e oxidatzon of acetaldehyde
As mentioned at the outset, it was lioped that this work might throw some light on the oxidation of acetaldehyde. It turns out in fact that the propionaldehyde oxidation is essentially similar to that of acetaldehyde, but lacks most of its complications. Thus both give per acids as the primary product. Both are accompanied by a pressure decrease, followed by an increase in the later stages. Both reactions have a low apparent activation energy, and there is no doubt that they proceed by chain mechanisms. In addition the concentration terms in the expression for the rate are strikingly similar for both. Thus the propionaldehyde reaction rate is proportional to the square of the aldehyde concentration and independent of that of oxygen, while the acetaldehyde reaction is proportional to (aldehyde)' i, and independent of oxygen. The objections to the Bodenstein mechanism outlined above apply with equal force to the acetaldehyde oxidation. In vien- of the great similarity of the two reactions there is little doubt that they proceed by analogous mechanisms. It is therefore suggested that the scheme outlined above applies also to the main chain process in the acetaldehf.de oxidation. The acetaldehyde oxidation, however, is complicated by a pronounced induction period. A similar effect is noticed in tlie propionaldehyde oxidation in the first few runs in a new reaction vessel. When the ~ ~ a lhave ls aged (presumably owing to irreversible adsorption of products), the induction period practically disappears. I t is apparent therefore that the
1200
E . W. R . STEACIE, W , H . HATCHER, AND 6 . ROSESBERG
induction period with acetaldehyde may be explained by assuming that poisoning of the walls is a necessary preliminary to the propagation of chains. SUMMARY
The kinetics of the oxidation of gaseous propionaldehyde have been investigated from 120°C. to 170°C. by a static method. The reaction is a chain process and is similar to the oxidation of acetaldehyde. The rate is proportional to the square of the aldehyde concentration, and independent of that of oxygen. The apparent heat of activation is 15,400 calories per gram-molecule. The following mechanism is suggested for the oxidation of aldehydes: RCHO RCOOOH* RCHO* RCOOOH* RCOOOH* RCOOOH* RCOOH*
=RCOOOH*
+ 0 2
+ RCHO = RCOOOH + RCHO* =RCOOOH* + 02 = RCOOOH + = RCOOOH (wall) + RCHO RCOOH* + RCOOH + RCHO = RCOOH + RCHO* $ 0 2
0 2
=
(1) (2) (3) (4) (5) (6)
(7)
In the absence of deactivation a t the wall) this leads to
-
d(RCH0) - 2KIKe,(RCHO)' dt K,
in agreement with experiment. REFEREKCES (1) BODESSIEIN:Sitzber. preuss. Akad. Kiss. 111, 73 (1931); Z. physik. Chem. 12B, 141 (1931). (2) FICHTER . ~ X DKRCMMENACHER: Helv. Chim. .Seta 1, 146 (1918). (3) HATCHER. STEACIE, . ~ N DHOWLAYD: Can. J. Research 5, 648 (1931); 7 , 149 (1932). (4) PEASE: J -4m. Chem. SOC.55, 2753 (1933).
