The Decomposition of Benzoyl Peroxide. II. The Rates of

Press, Baltimore 0935). (25) Rest. Seeboi.d, and ... but changes gradually as the peroxide concentration is increased, and finally assumes the limitin...
3 downloads 0 Views 648KB Size
('eHs('0

f

C's13t;

0 ('nH,C'O

+

CGH~

+

('eHs('OH

II

II

I/

+

C'eH5

(3)

0 --*

('sHjC'OCoHs

(4)

0

.it finit? coiicentrations. ho\vever, n i'ornially second-order side reaction is observed, which predominates above i i certain critical peroxide conrentration. The stoichiometry of this higher-order reaction is initially ('sHsC'OOCC:sH,

/I

0

/I

0

+

CsHs

-+

C',HsCOH

/I

0

+

('eHsCOCsH5

II

0

(5)

DECOMPOSITION OF BEiiZOYL PEROXIDE.

I1

943

but changes gradually as the peroxide concentration is increased, and finally assumes the limiting form:

The work is extended herein to a variety of other solvents and t o mixtures of certain solvents. The object of the study was to determine whether the first-order decomposition prevails in all cases at low enough peroxide concentrations, and in passing from solvent to solvent to follow the changes in stoichiometry, as reflected in the amounts of carbon dioxide and free acid formed. In all cases, a first-order decomposition proceeds at Ion peroxide concentrations. Particular attention was given t o the effect on the decomposition rate of the presence in the solvent of a double bond or a weakly held hydrogen atom, since in accordance with a proposal of Price 113) such solvents niay be expected to cause a n increase in the rate of decomposition owing to their greater disturbance of an assumed equilibrium between the peroxide and benzoate radicals. No such effect was observed; instead, in agreement with the literature ( 5 , 12), the highest decomposition rates were observed in certain highly polar, associated solvents, such as alcohols and acids. On the other hand, solvents containing a double bond or a ireakly held hydrogen atom may affect the stoichiometry strongly. I n general such solvents cause the evolution of appreciably less carbon dioside and the correspondingly increased formation of benzoic acid or a benzoate-olefin radical. I n the presence of osygen, and in some cases even in an air atmosphere, the decomposing peroxide may .sensitize the formation of a relatively stable hydroperoxide of the solvent. 11. EXPE:KIMEST.AL METHODS

The decompositions \yere carried out in the all-glass apparatus described in the preceding paper (1). The total acidity \\-as determined by titration with dilute standard sodium hydroxide. I n most eases the analysis for oxidizing power \vas carried out by the method previously described, in ivhich the peroxide liherated iodine from 1 solution of potassium iodide in acetic acid irhile the misture lras refluxed on the steam bath for 15 min. Certain sollmts-for example. allyl dcohol-were found to react with the iodine so formed, however, nnd for these it \vas necessary to use a modification of the method of Gelissen ond Hermans (6). In this modification a weighed sample of the reaction misture \vas washed into thc titration f l x k with cn. 10 c c . of acetone, and the flask flushed nith carbon dioside. Tirenty-five cubic centimeters of a saturated solution of sodium iodide in acetone were then added and the mixture shaken for 2-3 min. hefore titrating n-ith'standard thiosulfate.

JII. RATESOF L)Ecowosrrrox IS \*.~RIOVS SOI.VESTS .\T Low PEROXIDE Coxnm.\~rIoss Decomposition rates were measured in a variety of solvents a t relatiwly low peroxide concentrations whwe the main courre is a first-order reaction. The

944

BENJAMIN BARVETT AND WILLIAM E. V.4UGHAK

data are sunmarized in table 1 in which, for comparison, are included some data for benzene solutions given in the preceding paper. In some of the solventsfor example, benzene, ethylbenzene, cumene, chlorobenzene, benzaldehyde, n-hept,ane, and the ketones-the rates appear to be roughly dependent on their internal pressures (9), as might be expected on the basis of the theory of absolute reaction rates (8),but for most of the solvents no such simple correlation seems to hold. The a- and O-values in columns 5 and 6 of table 1 are respectively the ratios: moles carbon dioxide evolved moles free acid formed and moles peroxide decomposed moles peroxide decomposed Comparison of the data in table 1 for the solvents benzene, ethylbenzene, cumene, and benzaldehyde indicates that the availability of more weakly bonded hydrogen at,oms in the last three solvents does not increase the rate of decomposition. I n fact), the constants a t 80OC. for ethylbenzene and cumene are lower than the constant for benzene, and the specific rate in benzaldehyde is only 25 per cent faster than t'hat in benzene. This result appears to weigh against the existence of s n equilibrium between peroxide and benzoate radicals which might be rate determining, since solvents wit>hmore Tveakly held hydrogen atoms might be expected to remove the benzoate radicals more rapidly by reaction 3. The higher values of p in such solvents as cumene and ethylbenzene show that this reaction is, in fact, accelerated in these solvents, while the lower values of a show that reaction 2, by which carbon dioxide is evolved, is correspondingly depressed. The values of a obtained for benzene solutions are plotted in figure 1 against the temperature. The plot includes the data a t intermediate temperatures of McClurc, Robertson, and Cuthbertson (11). These authors conclude that a can vary only between 1 and 2, corresponding to the limiting stoichiometries:

This, however, would exclude the possibility of independent reaction, in accordance n-ith equations 2, 3, and 4, of the radicals formed in the basic firstorder decomposition of the peroxide. Figure 1 s h o w , in fact, that a does fall below unity a t temperatures below ca. 65°C. The apparent limiting values of l and 2 for a were found by McClure, Robertson, and Cuthbertson because they worked at temperatures above 65"C., where these limits happen to hold in benzene. The greater variability of a reported in the present work for benzene solutions is consistent with low a-values found in other solvents. As demonstrated in the preceding paper (1) the first-order decomposition of

DECOMPOSITIOS OF Bth'ZOYL 1"dROXIDE.

945

I1

TABLE 1 Decomposition of benzoyl peroxide Specific decomposition rates in various solvents I PO

kt (06s

molrr kt

Benzene ~

Bthyl benzene

Culllellc Chlorobeniene (a)$ Benzaldehyde (a) Phenol (a) Nitrobenzene (a1 Cvclohe\ane biethylcycloheuane Tetialin (a) Decalin (a) %-Heptane (a) Dioxane (a) Carbon tetiachioride (a) Methyl ethyl ketone Methyl isobutyl ketone n-Butyl alcohol . . , n-timy1 alcohol (a) 1 Benzyl alcohol (a) Allyl alcohol (a). . . . , Formic acidm (a). Acetic arid Propionic acid (a) ~

55 60 70 75 80 75 yo 85 00 80 85 90 80 80 90 80 80 80 80 h0

so

80 80 75 80 80 80 80

80 80 80 75 80

hr

0.07112 0 06699 0 06792 0 07031 I 0.06460 0.07203 0.08740 ' 0.07727 0.06792 0.08710 0.08260 0.06607 0.05834 0.06412 0,07380 0,00965 0.0s14 0.03733 0.030iO 0,06152 0.02i35 0.03548 0 04820 0 01161 0.07586 ' I 0.08570 0.02005 0.04477 0.06655 ' 0.05200 0.01118 0.04980 0.06830 ~

~

1

~

I

' ~

~

~

I

81

LI.

__-

_ _ _ _I

-

-1

0 00411 f 0 00007 0 00996 i.0 00031 0 0413 i 0 0016 0 OS34 i 0 0042 0.160 f 0.002 0.0650 f 0.0038 0.120 f 0.008 0.200 f 0.019 0.363 f 0.023 0.133 f 0.011 0.230 i 0.012 0.429 f 0.019 0.167 f 0.005 0.198 i.0.010 0.615 f 0.020 2.25 f 0.38 0.165 f 0.012 0.278 i.0.006 0.189 i 0.014 0.171 f 0.014 0.812 f 0.2 0.112 0.004 2 42 f 0 16 0 0384 i.0 0012 0.167 i.0.009 0.154 i 0.003 2.18 f 0.30 0.534 0.030 1.60 f 0.08 1.37 0.11 2.50 hO.15 0.271 & 0.016 0.115 i 0.003

*

+

~

0.72 0.86 1.21 1.42 1.55

1.14 0 81

0.86 0.92 1.21 0.76 0.88 1.06 0.95 0.20 0.21 0.04 0.71 0.91 0. 79 0.28

0.45 0 42

0 09

0.54 0.38 0 28

0.91 0.57 0.92 0.32

0.93 0.28 2.41 1 1.66 , 0.83 0.50 I 0.17 ~

~

'

*

,

i ~

0.06 0.75 0.79 1.03

~

0.70 0.32 0.13 0.23 1.42 0.93 1.71 1.23

' .-

moles carbon dioxide evolved moles peroxide decomposed . moles free arid formed t s = moles peroxide .decomposed' f Atmosphere nitrogen except for cases marked (a) (air). 5 The gas evolved from carbon tetrachloride solution precipitated silver chloride from solutions of silver nitrate. 7 The c . e . formic acid used contained ca. 10 per cent water. =

benzoyl peroxide in benzene is accompaxiied a t all finite concentrations by a higher-order decomposition. In certain cyclic hydrocarbons, such as cyclo-

2

I

tj

70

60

80

Temp. ‘C.

FICA1. Decomposition of benzoyl peroside in benzene. Variation of a (= moles carbon dioxide evolved/moles peroxide decomposed). 0 , present work; Q, reference 11.

27

2.8

2.9

3.0

+x IO’

FIG. 2. Decomposition of benzoyl peroxide. Temperature coefficient in various solvents. 0 , present work, PO= 0.064-0.071; 6 , reference 11, PO= 0.125;0, reference 10,PO= 0.05; V,reference 3,PO= 0.26. 916

DECOYPOSITIOS O F IILSZOYI.

PEROXIDE.

947

I1

hexane, tetralin, and decalin, t.he decomposition shows deviation from firstorder kinetics much earlier than in benzene. Presumably, in the former solvents the higher-order reactions are more rapid, and, consequently participate to a greater extent in the decomposition even a t low concentrations. This effect is also evident in the aliphatic acids. The data for hen'zene, ethylbenzene, cumene, and benzaldehyde permit estimation of the activation energies in these solvents. From conventional plots of --log, k, against l/Y' (see figure 2) the values given in table 2 h a l e been obtained. For clarity of presentat,ion the plots in figure 1have been displaced an amount a whose value is indicated in each case. Activation energies of from 30 to 33 kg.-cal. per mole have been reported for narrower temperature ranges for the decomposition in benzene (10, 11). All these values appear to approach the bond energy for the 0-0 linkage in the peroxide (14).l TABLE 2 Actioation energy of the decomposition of benzoyl perozide E

SOLYENT

kg cal h o l c

32

Benzene

Iv. THERATE O F DECOhfPOSITION O F BENZOYL PEROXIDE IN

MIXED SOLVENTS

A. DECOMPOSITION RATE I N BENZENE, A S D CUMENE COKTAINISG

20

ETHYLBEKZEKE,

PER CEXT BY VOLUME O F STYRENE

h previous comparison (1) of the available data (4, 11) indicated that a t 64°C. the presence of 20 per cent by volume of styrene probably had little effect on the decomposition rate of benzoyl peroxide in benzene. However, Price and Tate (13) have recently concluded that a t 80°C. the presence in benzene of this amount of styrene raises the rate of decomposition of tribromobenzoyl peroside almost sixfold over its value in pure benzene a t this temperature. T o determine the effect of this proportion of styrene on the rate at higher temperatures, a series of experiments were carried out, not only in benzene, but also in ethylRecently Cass (.J. Am. Chem. SOC.68, 1976 (1916))has published on the decomposition + 0.2O in a variety of solvents ranging from aromatic hydrocarbons t o polar straight-chain derivatives, such as acetone and butyraldehyde. He also finds a considerable variation in rate and departure from simple first-order kinetics. The activation energy of ca. 25 kg.-cal. found by Cass was obtained by combination of his own data at W'C. with the results of other investigators at higher temperatures. This author has also found that the ratio- depends strongly on the solvent, and may be less than unity. 1

of benzoyl peroxide a t 30°C.

948

REKJA?dIK RARKbTT A S D WILLIAM E. VAUOHAh'

benzene and in cumene, over a range of temperatures. The benzoyl peroxide concentration used was approximately the same as that of the tribromobenroyl peroxide in the experiments of Price and Tate. The data appear in table 3, which also includes values of 01 and 6. The styrene used was prepared just before starting each experiment in which the monomer was to be added. The crude styrene, a Dow product, was washed with 10 per cent sodium hydroxide and then distilled a t about 45'C. under a pressure of 20 mm. of mercury. TABLE 3

Decomposition of benzoyl peroxide Effect of the presence of 20 per cent by volume styrene on the decomposition rate in certain solvents SOLVEKT

__ molrrlkg.

"C.

hr.-c

0.0679 0.0703 0.0646

70

0.0413 f 0.0016 0.0834 f 0.0042 0.160 f 0 . 0 0 2

1.19 1.37

0.0720

io

0.0714 0.0594

75 80

0.0418 f 0.0014 0.0832 f 0.0024 0.160 f 0.002

0.039 0.53 0.92

0.0720 0.0679

75 90

0.0656 f 0.0038 0.363 f 0.023

1.21

0.0814 0.0631

75 90

0.0668 f 0.0028 0.466 f 0.023

0.48 0.82

.

0.0661

90

0.429

f 0.019

1.06

+ styrene.. . . . . . . .

0.0619

Ix)

0.422 f 0.017

0.73

Benzene, . , . , . . . . . , . . . , . . .

Benzene

+ styrene.,. . . . . . .

Ethylbenzene , . . , . . , . . . . . . . Ethylbenzene

+ styrene.. .

Cumene.. . . . . . . . . . . . . Cumene

,

..

,

75 80

1.55

0.09

Between 70" and 8OoC., as a t 64'C., the results show that the presence of styrene in the solvents used has no sensible effect on the rate of decomposition of benzoyl peroxide. Price and Tate, in carrying out their experiments a t SO'C., state that their solutions were refluxed a t this temperature. It is evident, however, that their initial temperature must have been considerably higher. h boiling point of ca. 93°C.is to be expected for a 20 per cent by volume solution of styrene in benzene, but actually such a solution has been found to boil a t 86.2"C.a t 762 mm., if boiling chips are present. Otherwise it can overheat to maintain as high a temperature as 93-94°C. On adding an amount of benzoyl peroxide sufficient to give the molar concentration of peroxide used by Price and Tate the boiling point rose to 86.7"C., and thereafter fell off as polymerization progressed:

DECOMPOSITION OF BENZOYL PEROXIDE.

I

TILIE

949

I1

BOILING P O N AT 762 MX.

'C.

hwrs

0 1.50

86.7 85.4

2.50 4.25 5.25 16.5 23.5

84.8 84.4

84.2 82.2 82.1

Consequently, depending on whether overheating occurred, the decomposition rates observed by Price and Tate for their tribromobenzoyl peroxide may have been obtained a t temperatures from 6-13OC. higher than that measured in pure benzene a t 80°C. On this basis, and assuming the same activation energy for the decomposition of tribromobenzoyl peroxide as for that of benzoyl peroxide, this temperature effect could account for a rate increase of from 2.5- to 4.5-fold in the presence of the added styrene The increase in the values of CY in each solvent with increasing temperature s h o w the increasing tendency of benzoate radicals t o decompose to carbon dioxide and phenyl radicals by reaction 2. The effect is clearly evident also in the presence of styrene, but such solutions the values of a are much lower, at the same temperature. than In the pure solvents. Thus, although the monomer does not sensibly affect the rate of decomposition of benzoyl peroxide, it enters intimately into the folloTving rewtions of the benzoate radical by rapidly removing it, presumably in accord with the reaction.

k

1

CsH,C=O

I CnHbC=O

which must be faster, particularly a t the loner temperatures, than reaction 2. At 70°C., in benzene containing styrene, equation 8 represents the major path for the removal of benzoate radicals, and even at 80"C., in this mixed solvent, the sum of a /3 is still less than the value of a alone in pure benzene a t this temperature. This result appears to comprise the first evidence for the entrance of the benzoate radical itself into the polymerization of styrene, and is in agreement nith the finding of Bartlett and Cohen (2) that with certain halogenated benzoyl peroxides polymerization is initiated principally by addition of the halogenated benzoate radicals.

+

B. EFFECT O F BENZOIC ANHYDRIDE ON THE DECOZlPOSITIO% I S BENZEKL

In the preceding section it was shown that the presence in considerable amount of a substance whose structure is somewhat skin to that of the solvents used, results in no sensible increase in the observed rate of decomposition. However,

950

l l E S J A b l I S BARNETT .\XD WILLIAM E. V.IUGH.AN

when benzoic anhydride is added, a considerable decrease in rate is observed even with relatively small additions. In the determination of the rate of decomposition in benzene, increasing tjhe peroside concentration may lead not only to higher-order reaction, but, also to a decided change in the nature of the solvent. The latter effect alone may cause a shift in the observed first-order constant. Thc only means of separating these two effects appears to be an indirect, one, bascd on a comparison of the change in the decomposition rate of the peroside on adding t i comparatively inert substance. When benzoic anhydride is added in increasing proportions at 75°C. to solutions of t,he peroside in benzene containing 0.11 mole of pcroside per kilogram of solution after addition of the anhydride, the ohserved first,-order rate falls off steadily, as shown in figure 3. The rate at, which thc velocity demaser; becaiisc

1.0 c u

0.4 0.0 Moles Benzoic Anhydriddkg. SdutiCn

FIG.3. Decomposition of benzoyl peroside. Effect of benzoic anhydride on decomposition rate of 0.14 mole/!ig. solution of peroxide in benzene. 2' = 75°C. of the presence of the anhydride is about one-quarter the rate of increase of k t (obs.) with the concentration of peroxide and in the absence of anhydride. I n view of the structural similarity between benzoic anhydride and benzoyl peroxide, it is possible that the latter also has a certain inhibiting solvent effect, but if t>hisoccurs it is more than offset by the progress of t>hehigher-order reaction.

v. R A T E S

DECONPOSITIOX OP REXZOl.1, PE:ROXIDE I N ~ A R I O C SSOLVEKTS PRESENCE OF OXYGEK:PEROXIDE-SEXSITIZED FORMATIOX OF PEROXIDES OF THE SOLVESTS I n certain solvents, such as benzene, aldehydes, and ketones, the rate of decrease of oxidizing pov-er is not sensibly different in an air atmosphere from the rate in one of nitrogen. I n other solvents, such as the alcohols, this rate in air is markedly less a t the start than later in the reaction. In still another group of solvents, which include the alkyl-substituted benzenes, an atmosphere of oxygen will bring about a considerable perturbation in thP init,ial rrtt,e nf 1nw of oxidizing power. OF

IS THE

DECOMPOSITION O F BhKZOYL PEROXIDE.

I1

95 1

I n the cases of alcohols and acids the effect is to simulate an induction period, as is illustrated by figure 1. I n substituted benzenes, as in ethylbenzene and cumene, and in cycloparaffins, the oxidizing power actually rises in the presence of oxygen, often quite rapidly, although the benzoyl peroxide is decomposing during this interval The appearance of more oxidizing power can only be due to the formation of new peroxide, apparently a hydroperoxide of the solvent, by absorption of oxygen by the reacting mixture. A.

FORM.ATIOS

O F PEROXIDE I S ETHYLHEKZESE

The partial pressure of oxygen in air has only a slight effect on the rate of loss of oxidizing power in solutions of benzoyl peroxide in ethylbenzene a t 80°C. In the presence of pure oxygen, hon-ever, a11 initial quite rapid rise in Oxidizing power is observed until the oxygen above the reacting mixture is exhausted. Thereafter, the oxidizing poiver decreases (see figure 5 ) . The left section of the upper curve in this figure indicates slight autocatalysis, an effect which ivould probably be more pronounced if it were not for the rapid decrease in the partial pressure of oxygen. Autocatalysis could result by the following path, of a chain character, to hydroperoxide: CCHbCO

I/ ij

+

CsH,CHCH,

CcH5CHCJHa 0 2

CfiHjC'HzCH,

-+

CGH5COH

II

-t

C'RHSCHCHI

(9)

0

+

0 2

--

-+

+ CsHbCHzCHx

CsHjCHCH, 0 2 --+

C6HbCHCHs

+

CsHbCHCH3 (11)

02H

Phenyl radicals, as well as benzoate, could initiate this chain. While the t,otal oxidizing pon-er is rising to a maximurn, the benzoyl peroxide presumably is decomposing a t the normal rate in the absence of oxygen and, a t the maximurn the partial pressure of oxygen above the reaction mixture has dropped so loiv that the difference betn-een the rates of hydroperoxide formation and decomposition is just balanced by the rate of decomposition of benzoyl peroxide. For a short distance beyond the maximum the latter rate predominates, and thereafter the rate of decomposition of benzoyl peroxide is the only measui,ed rate. This is shown by comparing the calculated curve for the decomposition of the peroxide in a nitrogen atmosphere with the right-hand portion of the experimental curve representing total oxidizing power. The two curves become parallel a short distance beyond the maximum, supporting the presumption that during the first 2 hr. the benzoyl peroxide decomposes at the same rate in an atmosphere of oxygen as in a nitrogen atmosphere. On this basis the oxidizing power found a t the maximum \vas corrected for the benzoyl peroxide decomposed up to this point. This correction is indicated in figure 5 . The hydroperoxide of ethylbenzene is evidently stable under these conditions. While the experimental curve mas rising to its maximum in this experiment

c 0 c

.-

$-=go

I

2

3

4

Time, hr.

FIG.4. Decomposition of benzoyl peroxide. Decomposition rate in n-butyl alcohol in nitrogen and air atmospheres. T = 75°C.

h

8

-a

0.13

0-

c

2

w

g

0.11

0)

0 3

P

a .E .- 0.09

60

.-x

-0

50

0

a

.-8 c

d 405

0.07

30

B 0

4

p

20

0.05

5

B

IO

-

FIG.5. Decomposition d benzoyl peroxide. Peroxide-semitixed formation of Iiydroperoxide in ethylbensene in atmosphere of oxygen at the start. T S0"C. 952

DECOMFOSITION OF BENZOYL PEROXIDE.

I1

953

0.0155 mole of benzoyl peroxide decomposed and 0.0530 mole of hydroperoxide formed. The ratio: 10.0530 - l , i 2 0.0155 -

is the average number of hydroperoxide molecules formed per benzoate radical produced, This is probably a minimum value, for it depends on the rate of diffusion of oxygen into the reaction mixture during the first part of the experiment, or on the mean equilibrium concentration of oxygen in this period. If the former is relatively slow, or t’he latter low, many of the ethylbenzene radicals formed by reactions I) and 1 1 may dimerize or form est,er nith benzoate radicals ld’ore they can encounter an oxygen molecule. By passing air or oxygen continuously through the reaction mixture at atmospheric pressure the above ratio has been rsised to 3.6. Thr oxygen above the reaction mixture is removed fairly rapidly, perhaps R S rapidly as the rate of diffusion of the gas into the reaction mixture will permit. I n a closed system there occurs the rapid drop in pressure represented by the lower curve, a minimum pressure being observed a t the time the maximum is olitainetl in the curve representing changes in the oxidizing power of the mixture. Sozaki and I3artlett (12) state that the effect of oxygen and air is t o inhibit per s e the radical-induced decomposition uf benzoyl peroxide. The use of decomposing peroxides to sensitize peroxide formation is not new. For csample, George (7) hns used benzoyl peroxide in the air-oxidation of tetralin, and the principle is implicit in such processes as thaf of catalyzing Diesel fuel ouidation by inoculating with a portion of n previously oxidized bat,rh of fuel (15). 13. FORhl.ATIOS

OF PKROXIDE I S .II,COHOLS

‘1 typical rate curve for the decomposition of benzoyl peroxide in ti-butyl alcohol a t 75°C. in an atmosphere of air is shown in figure 4. The effect of peroxide formation is to give a pseudo induction period. Actually this is due t o the approximate balance between the rate of formation of new peroxide by absorption of oxygen and the normal rate of decomposition of benzoyl peroxide. Since the absorption of oxygen is probably rapid in both n-butyl alcohol and ethylbenzene, the approximate balance observed in the case of the alcohol may he due to the knonn greater decomposition ratr of benzoyl peroxide in this solvent than in ethylbenzene. When the oxygen above the reacting mixture has been largely removed, the rate of loss of oxidizing power in n-butyl alcohol approximat,es that found when 9, nitrogen atmosphere is present from the start; this is shown by the slopes drawn in figure 4. A possible course t o peroxide in alcohols would appear t o be by way of the radical RCHOH, which would eventually lead to hydroperoxide by a mechanism similar to that suggested for ethylbenzene. The presence of aldehyde has been found to reduce markedly the initisl

954

BEKJAMIS H.%KStT'I' A S D \VII.l.IAM

1,;.

V.IGOH.IS

period of slow change in oxidizing power in solutions of benzoyl peroxide in alcohols decomposing in an atmosphere of air. For example, the effect a t 80°C. of adding increasing proportions of propionaldehyde to solutions of benzoyl peroxide in n-propyl alcohol is shown in figure fi. The presence of 2.5 per cent by volume of the aldehyde reduces the initial period of sloiv change to about one-quarter its value in pure n-propyl alcohol,and in the presence of 10 per cent by volume of propionaldehyde this interval is completely eliminated. The rate corresponding t o the straight-line plot for the mixture containing 10 per cent by volume of aldehydp is sensibly less than the rate in pure n-propyl alcohol in

FIQ. 6. Decomposition of benzoyl peroxide. Decomposition rate in n-propyl alcohol and in n-propyl alcohol containing propionaldehyde. Air atmosphere. T = 80°C.

the absence of air, but this is in accord with the lower decomposition rates of benzoyl peroxide in aldehydes than in alcohols. C. FORMATION O F P E R O X I D E J l i BCETIC ACID .9ND IN CYCLOPARAFFlXS

At 75"C., in an atmosphere of air, acetic acid initially displays a well-defined period of slow change in oxidizing power, but the effect is only about one-quarter that found in n-propyl alcohol. Beyond this interval the decomposition of the peroxide is closely first order up to over 70 per cent decomposition. Here also, a rapid contraction occurs in a closed system owing to the absorption of the oxygen above the reaction mixture. Similar results are obtained when benzoyl peroxide decomposes in a n oxygen

DECOXIPOSITIOK OF BEKZOYL PEROXIDE.

I1

atmosphere at 80°C. in cyclohexane and in methylcyclohexanr. carbons gave R value of cn. 2 for the ratio

955

These hytlro-

moles hydroperoxide formed 2(rnoles benzoyl peroxide decompose>? \vhen a constant stream of air or oxygen \vas passed through the reacbt,ing mixtures at atmospheric piwsiirc. 1'1. S K N > I . \ R Y 1 . I t is found t,hat at lo\v peroxide concentrations, t)he decomposition of benzoyl peroxide in tn-enty-three different solvents is hasirally first order. alt,hough varying some t\\-entyfoldin magnitude. 2 . The presence of 20 per rent by volume of polymerizing styrene has littlr effect on the ratrl of decomposition in benzene at, temperatures up to S0"C. This result, bogether \\-ith other evidenre presented, appears to Tveaken t,he asslimption of the existenre of an equilihrium hetireen benzoyl peroxide and hrnzoate radicals. 3. Decomposition rates of benzoyl peroxide in many other solvents indicate that the occurrence of higher-order reactions accompanying a basic first-order change is common to most, if not all, solvents. 4, Many solirents, particularly hydrocarbons, alcohols, and acids, form hydroperoxides \\.hen solutions of benzoyl peroxide in t#hesesolvents are allowed to decompose in ai1 atmosphere of air or oxygen. HF:FEREKCES ( 1 ) BARSETTA S U VAUGHAS: J . Phys. Colloid Chem. 61, 026 (1947).

(2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) i14)

(15)

BARTLETT A N D COHEN:J. Am. Chem. Soc. 66, 543 (1943). BEREZOVSKAYA A N D \rARI"OLOMEEVA: J. Phys. Chem. (LT. s. s. R.) 14, (2) 936 (1940). COHEK:J . Am. Chem. SOC.6 7 , l i (1945). GELISSEN ASD HERMASS: Ber. 68, 765 (1925). GELISSEKA N D HERIMASS: Der. 69, 63 (1926). GEORGE:Trans. Faraday Soc. 42, 210 (1946). GLASSTONE, LAIDLER, ASD EYRING: Theory of Rak Pmcesses, pp. 410-17. McGrawHill Book Company, Inc., New York (1941). HILDEBRAND: Solubility, 2nd edition, Chap. V. The Chemical Catalog Company, Inc., Sew York (1936). K A a l E s s K A Y A AKD MEDWEDEV: ."eta Ph,ysicochim. c. R . s. s. 13, 565 (1940). hkCLURE, ROBERTSON, AND CI:TABERTSON: Can. J. Research POB, 103 (1942). KOZAKI A N D BARTLETT: J. Am. Chem. SOC.68,1686 (1946). PRICEA N D TATE:J. Am. Chem. SOC.66, 517 (1943). RICEAND RICE:Aliphatic Free Radicals, p. 165. The Johns Hopkins University Press, Baltimore (1935). U. S . patent 2,365,220 (to Standard Oil Co. of California), December 19, 1944.