THE REASONS FOR THE D I F F E R E N C E BETWEEN THE GAS AND L I Q U I D PHASE O X I D A T I O N M E C H A N I S M S OF O R G A N I C S U B S T A N C E S G. E. Z a i k o v

and

Z. K. M a i z u s

Institute of Chemical Physics, Academy of Sciences, USSR Translated from Izvestiya Akademii Nauk SSSR, Otdelenie Khimicheskikh Nauk, No. 7, pp. 1175-1184, July, 1962 Original article submitted January 30, 1962

Liquid phase oxidation of organic compounds is as a rule studied in a way distinct from the oxidation of the same compounds in the gaseous phase. However, a comparative study made on reactions of the same compounds in different states of aggregation provides new information on the reactivity of the substance as a function of its properties in the state in which the reaction is being carried out. In our previous papers [1-3], a study was made of the oxidation of butane, ethyl alcohol and methyl ethyl ketone under comparable conditions in the gaseous and liquid phases. It was shown that the composition of the reaction products depends to a large extent on how the reaction is carried out, the greatest difference between gas and liquid phase oxidation being found for ethyl alcohol, and the least for butane. One of the explanations for the observed difference in composition of the reaction products could be that the concentration of the material being oxidized has an effect on the reaction mechanism. According to present-day ideas [4], change in direction of the reaction (composition of the reaction products) during oxidation of hydrocarbons in the gaseous phase is due to competition between the two reactions of the peroxide radical RO2" which leads the chain reaction-namely, isomeriz/ation and decomposition of the RO~ radical: RO2"--~ R'O"-k R"HO,

(1)

where R' and R" are radicals with a smaller number of C atoms than in the material being oxidized, and reaction between the RO~ radical and the original hydrocarbon: RO~' -~- RH -+ RO~H -~- R'.

(2)

From this point of view, when hydrocarbons are oxidized in the liquid phase, where there is a high density of material, the rate of reaction (2) (wz) must be considerably greater than the rate of reaction (1) (wl). Accordingly, in the products of a liquid phase oxidation the compounds formed by transformations of the intermediate hydroperoxide will predominate over decomposition products of the RO~ radical. ActuaUy, in butane oxidation [1] it was found that on going from liquid to gaseous phase oxidation at the same temperature (145 ~ and pressure (50 atm) the composition of'the reaction products changes in accordance with the above ideas as to what effect the butane concentration should have on the competition between the two reactions of the peroxide radical. Thus, if the n-butane concentration is reduced by 84.6% (from 0.89 g / e m s in the liquid phase to 0.061 g / c m s in the gas) the value of wz/w 1 drops by 86.7qo, i.e., by about as much as the n-butane concentration. When a similar set of experiments was made on the polar compounds ethyl alcohol [2] and methylethyl ketone [3], it was found that the change in concentration of the substance being oxidized when its state of aggregation is changed is not sufficient to explain the differences observed in the composition of the oxidation products, The rate ratio w~] w1 in the liquid phase is 320 times greater than the value of wz/w I in the gaseous phase for ethyl alcohol, and 60 times greater for methyl ethyl ketone, while the concentration of material being oxidized does not change more than 6 times in the two cases. In order to find the reasons for the difference in oxidation mechanism of organic compounds in the gaseous and liquid phase, we investigated the oxidation of ethyl alcohol and methyl ethyl ketone when mixed with different amounts of benzene. Benzene, as an ouly slightly polar substance, reduces the dielectric constant of the medium, while remaining inert to oxidation. EXPERIMENT A L The technique used in making the experiments on ethyl alcohol and methyl ethyl ketone oxidation inbenzene mixtures was that described in previous papers [1-8, 5-7]. For convenience in comparing the results obtained at dif-

1102

ferent benzene dilutions, the concentrations of all the compounds present in the system are given in mole % of the ethyl alcohol or methyl ethyl ketone used in the oxidation. The ethyl alcohol was oxidized at 200 ~ and 50 atm with the ethanol: benzene molar ratios 8: 1, 2: 1, 1: 1, 1: 2, and 1: 3. As the amount of benzene in the mixture is increased there is a considerable change in composition of the products of the reaction. This may be clearly seen from the kinetic curves for the build-up of all the oxidation

I//0

C4H~O mole % z,0

~f/~

3,

71/

gg

ft~ /,g

A

1'

4O 4g

2 2/)

E,fg % 40

B

"Y ?.0 /5 1

/0 5

Z 3 Lg-~

0 /5

C

. "~r-g $'

D

d'

/

'0

I0

20 3g Time, hours Fig. 1. Kinetic curves of ethyl alcohol consumption and reaction product build-up during oxidation without solvent (1, 2, 3) and dissolved in benzene in the ethyl alcohol: benzene molar ratio 1 : 3 ( t ' , 2', 3'); 200 ~ 50 atm. A - 1,1') Ethyl alcohol consumption (outside scale); 2,2') acetic acid build-up (inside scale); 3,3') formic acid build-up (right-hand scale). B - 1,1') Acetaldehyde (outside scale); 2,2') formaldehyde (inside scale); 3,3') methyl alcohol (inside scale). C - 1,1') Ethyl acetate (outside scale); 2,2') ethyl formate (inside scale); 3,3') peroxide (inside scale). D - 1,1') COz; 2,2') CO.

0

~

lo

/~

t, hours

Fig. 2. Kinetic curves of methyl ethyl ketone consumption and product build-up during oxidation without solvent (1, 2, 3, 4) and dissolved in benzene with the methyl ethyl ketone: benzene molar ratio 1 : 3 (1', 2', 3', 4'); 145 ~ 50 atm. A - 1,1') Methyl ethyl ketone consumption; 2,2') acetic acid build-up; 3,3') formic acid (inside scale). B - 1,1') Diacetyl; 2,2') acetaldehyde; 3,3') acetone; 4,4') formaldehyde. C - 1,1') Ethyl acetate; 2,2') methyl acetate; 3,3') COz; 4,4') CO. D - 1,1') Peroxide; 2,2') ethyl alcohol; 3,3') methyl alcohol.

products given in Fig. 1 for undiluted ethyl alcohol and for the highest benzene dilution, as well as from the maxim u m product formation rates (Table 1). The relative amounts of products from the bimolecular form of the reaction (acetic acid and ethyl acetate), which in the oxidation of pure ethyl alcohol amount to ,-, 80 mole % of the alcohol reacted, drop off sharply to 20 mole % with 1 : 3 benzene dilution, this being the yield for gaseous phase oxidation of the alcohol [2]. At the same time, dilution with benzene increases the yield of decomposition products

1103

of the RO; radical (formic acid, ethyl formate, methanol, formaldehyde and CO), which for practical purposes do not occur at all in the absence of benzene. The methyl ethyl ketone was oxidized at 14,5~ and 50 aim at the methyl ethyl ketone: benzene molar ratios 1 : 1, 1 : 2, and 1: 3. In this case also a marked change is observed in the composition of the oxidation products when the methyl ethyl ketone is diluted with benzene (Fig. 2, Table 2) and at the methyl ethyl ketone : benzene ratio 1 : 3 the product composition becomes practically the same as when oxidized in the gaseous phase. The quantitative measurement of the rates of formation of the oxidation products of methyl ethyl ketone and ethyl alcohol from the two reactions of the RO2" radical as a function of the benzene dilution was made in the same way as in the previous papers [1-3], from the sum of the maximum build-up rates of all the products of the given reTABLE i. Maximum Build-Up Rate of the Oxidation Products of Ethyl Alcohol from the Bimolecular (w2) and the Monomolecular (wl) Reaction of the RO~ Radical at the G2I-I~OH: C6Hr Molar Ratios 8 : 1, 2: 1, 1 : 1, 1 : 2, and 1 : 3 (Temperature 200 ~ pressure 50 atm) Maximum product build-up rate in mole % of ethyl alcohol ~ per hour' Reaction product., without dilution [r

2:1

8: J

~l

~272

~'1

Acetic acid Formic acid

5 ,O0

o?o 3'2 o~

Ethyl acetate Ethyl formate

4,20

0~9

Acetaldehyde

2~t

Formaldehyde Peroxide " Methanol CO CO2

o~o

~176/ o29 155

Y, te

1,70

o~64 0,25

w2

I:1

I

I

1:2

I

r;:,,

-

Io,5ol-/o,28r

1,~o

o,o,, I - Io,o,,t - Io ;,,l~

1.,40

1 0 , 3 8 1 - /0,t21 o ~ 6 I - - / o , 0 4 / - ,o

t7o

o~7

1:3

~0,t2

ot 7_

0,~0/0,8B/o,IB/O,70JO 38[0,22 0,04 1 - 1o,o5/- Io 04[ -

o:;3

032

0~3

o~3

/~176 1~176 / 037 o,o io?,&Gi o

t3,04

0,40 [ 8,12

0,22

5,24

t ,06 9,07 0,75 3,4O

0,32 12,2310,41tl ,3210,7210,6812,52

TABLE 2. Maximum Build-Up Rate of Oxidation Products of Methyl Ethyl Ketone from the Bimolecular (wz) and the Monomolecular (wl) Reaction of the RO~ Radical at Methyl Ethyl Ketone : Benzene Molar Ratios 1 : 1, 1 : 2, and 1 : 3 (Temperature 145 ~ pressure 50 atm)

Reaction products

Acetic acid Formic acid

Maximum product build-up rate in mole % of methyl ethyl ketone per hour without 1:2 1:3 dilution I:i I/~2

I121

~/22

30,3

0,2

25,t

A cetone

Acetaldehyde Formaldehyde Diacetyl Ethanol MethanolEthyl acetate Methyl acetate CO 2 CO Peroxide Y.w

1104

1,4 0,2

I/)~

wt

2t,6

2,8 0,4

W2

Wl

t4,0

3,5 0,9 0,4 2,2 0,2

0,3 1,0 0,1

---

0,t 0,6

1o~2 0,1

9,8 0,t

---

8,2 0,t

9,5 0,4

7,6

8,4

--

6,2

5,t 0,2 t,5

0,1

~3 o-3 ~t o,~ ~,o o,1 ~,8 L0'9 49,5

0,4

45,3 [ 2,6

1,1

0,t 0,2

0,8 38,0

419

&

3t ,9

0,t 0,3 O;4 8,1

action (see Tables 1 and 2). The fraction of acetic acid and CO 2 formed when methyl ethyl ketone is oxidized a c cording to the two reactions of the peroxide radical was measured in the following way. the maximum build-up rates of the products (with the exception of CH3COOH and CO2) formed by reactions(I), (2) (w' 1 and w'2) were added up, and then it was assumed that the ratios w'~/(w'2 + w'1) and w' t / ( w ' 2 + w'l) determine the fraction of acetic acid and COs formed in each of the reactions. A similar calculation gave the fraction of acetaldehyde and CO2 formed in the oxidation of ethyl alcohol by reactions (1) and (2). In a previous paper [3] this method of attack was found to be correct in special experiments made by adding acetaldehyde to methyl ethyl ketone being oxidized, which made it possible to find the acetaldehyde consumption rate constant in the gaseous phase (k). Since under these experimental conditions all the acetaldehyde is converted into acetic acid, the value of k makes it possible to find the rate of acetic acid formation from acetaldehyde. For practical purposes the resulting value is in complete agreement with the one calculated by the method we have used. If we consider the results obtained by benzene dilution of substances undergoing oxidation, we have to consider the possibility of a sharp change in over-all reaction rate coming from the critical phenomena observed re~2 u % mole % 1'

I .....

--

30g Io gO -g s 1o : t _ .

g~

/o /o o~

, 2

P

g

0,5

o,2

~4

l,

,

o,s

Fig. 3.

t

o,8

.,

t-

I .....

r

@

0,s

t"

1,0o~

Fig. 4.

Fig. 3. Over-all oxidation rate of methyl ethyl ketone (1, outside scale) and ethyl alcohol (2, inside scale) as a function of the mole fraction of the substance undergoing oxidation (a) in a mixture of benzene and methyl ethyl ketone or ethyl alcohol. The corresponding oxidation rates of ethyl alcohol and methyl ethyl ketone in the gaseous phase are given as dotted lines (1' and 2')~ Fig. 4. Ratio w2/w20 of the reaction rates of methyl ethyl ketone (1) and ethyl alcohol (2) with the corresponding peroxide radicals as a function of mole fraction of the substance undergoing oxidation when mixed with benzene (wz0 is the rate of the ROz + RH reaction in the experiments without dilution). The dotted lines are the values of wz/w20 for gaseous phase oxidation of ethyl alcohol (1')and methyl ethyl ketone (2'). cently in the oxidation of butane [8] and methyl ethyl ketone. In our experiments on methyl ethyl ketone oxidation we also observe a reduction in the over-all reaction rate practically down to zero value for an insignificant increase in benzene concentration in the original mixture from 77 to 80 mole % (Fig. 3). However the methyl ethyl ketone: benzene ratio 1 : 3 (the highest dilution) corresponds with a benzene concentration less than critical and the overall oxidation rate of methyl ethyl ketone does not change significantly in this range of benzene concentrations. In the oxidation of ethyl alcohol, no critical phenomena are observed under our conditions even up to 80 m o l e % benzene in the initial mixture. The consumption rate of the ethanol when diluted in 1 : 1 ratio with benzene becomes equal to the gaseous phase oxidation rate, and does not change with further increase in the amount of benzene used. Thus, in neither case (oxidation of methyl ethyl ketone or ethyl aleohoi) can the inhibiting action of benzene on the over-all reaction rate have any effect on the results of our experiments, Comparing the ratio w 2 / w t for different concentrations of methyl ethyl ketone and alcohol in benzene with the values of w z / w 1 for gaseous phase oxidation of these substances (Table 3), we come to the conclusion that the greater the reduction in polarity of the medium from dilution with benzene, the less difference there is between the

1105

reaction mechanisms in the gaseous and liquid phases. Actually, although in ethyl alcohol oxidation the ratio of the values w ~ / w 1 for liquid and gaseous phases was equal to 320, with benzene dilution in the ratio 1 : 3 it becomes equal to 2.45. Allowing for the concentrations of ethyl alcohol in the benzene mixtures (c) and in the gaseous phase (Cg) this gives a change in the quantity B = (w2/wl) Cg from 67 to 2.8. With methyl ethyl ketone, the value of 8 (w2/wD g c

changes from 9.9 for the experiments without dilution to 1.14 for a methyl ethyl ketone : benzene molar ratio of 1:3. Thus, the conclusion may be drawn that if the initial substance is diluted sufficiently with benzene the composition of the oxidation products of ethyl alcohol and methyl ethyl ketone comes to agree with the ideas that have been advanced as to the effect which concentration of the substance undergoing oxidation has on the reaction m e chanism, in the same way as has been observed in the oxidation of the nonpolar compound n-butane [1]. The relative change in the quantity w~ when the dielectric constant of the medium is reduced is given in Fig. 4,which shows w2/w~0 [where w20 is the rate of the reaction (2) in the experiments made without dilution] as a function of the mole fraction of the substance undergoing oxidation in the benzene mixture. It is clearly seen that the effect of dilution on ethyl alcohol oxidation is considerably greater than in methyl ethyl ketone oxidation, and that in both cases if the dilution is great enough the value of w~ approaches the value of W2 for gaseous phase oxidation. D I S C U S S I O N OF E X P E R I M E N T A L R E S U L T S The bimolecular reaction of the RO2" radical with methyl ethyl ketone or ethyl alcohol may be thought of as an interaction between two polar compounds. It is well known that when two dipoles with the moments Pt and p~ interact the rate constant k 2 of the reaction must be a function of the dielectric constant of the medium [9]; thus

lg k2 = lg k o

~

~

~ -- 1

2,3kT

2~+ i " ~ 7 g '

(3)

where k 0 is the true rate constant of the reaction in the gaseous phase, and k is Boltzmaun's constant.

For the

reactions considered, the quantity ~ -~- may be written as

~2 ~-fi---

~trI r--5 --'-'~ R.

I~O~

I~

(4)

4

a

where PRH' PRO'-' rp,H' and rRO- are the dipole moments and the effective radii of the reacting substances and RO~ radicals, whhe p # and r ~"-'2are the same thing for the activated complex. The values for the dielectric constant of methyl ethyl ketone and ethyl alcohol in the temperature range that we are interested in were found by extrapolating the tabular data from the formula ~ = A +(B/T)[10]. The dielectric constants of the various alcohol and methyl ethyl ketone-benzene mixtures were calculated from the formula [11] lg e =

x l l g q + x ~ l g e 2,

where x is the mole fraction of each component of the mixture.

TABLE 3. Effect of Solvent on Direction of Oxidation of Ethyl Alcohol and Methyl Ethyl Ketone Ethyl alcohol Methyl ethyl ketone t methyl ethyl ]~ I ketone in ben -[ ~ o / zene T~ N [

i :,1, :21i': alg,~l

solution of ethyl alcohol in be 1 z e n e 8:t.

2:t

I:I

1:2

1:3

I

w"c, g/cm s wt /

k wl /

7,8 3,85i2,1 32,7 37,0 16,4 5,4 1,830,25( 0,t02 0,2 3,t430,11~ 0,t20 0,6 [0,32 O,201'0',t610, lq 0,58 0,4! ;63 9 16~ '3: 53,5 t8,0 2,45 1,0 i9,2 t 8,3 3,7 t,831/,0 20

t

9,9

1106

2,6 t,85 l fl4/l,O

67,ol 89,0

58,7 28,2 15,0 12,8

1,0

The true values of the quantity k2 for the oxidation of methyl ethyl ketone and ethyl alcohol are unknown. However for an orientational calculation of how the reaction rate varies with polarity of the medium we can use the corresponding values of w z instead of k z. The experimental data found for the oxidation of methyl ethyl ketone and ethyl atcohol in benzene mixtures give a good fit to straight lines when plotted in the form log wz vs. ( ~ - 1)/(2e + 1) (Fig. 5). This supports the idea that the bimolecular reaction of the peroxide radical is an interaction between two dipoles, and that the rate of the reaction drops off with reduction in polarity of the medium. Thus, the sharp increase in the ratio wz/w 1 observed on going from gaseous phase oxidation of polar substances to liquid phase oxidation actually comes from an increase in the rate of reaction (2) brought about by the electrostatic forces characteristic of the liquid state of the material. If the two reactions, the oxidation of methyl ethyl ketone and ethyl alcohol, are considered together, we do not observe any correspondence between changes in the dielectric constant of the medium and the value ofw 2. Thus the slope of the straight line of logl0w z vs. (g - 1)/(2e + 1) (see Fig. 5) is considerably greater for ethyl a l c o hol than for methyl ethyl ketone, while the dielectric constant of the alcohol is less than that of the ketone [3], In order to explain why this lack of correspondence occurs we have tried to use our experimental data to calculate the dipole moments of the activated complexes for the two reactions and compare them with the dipole moments of the activated complexes found by calculation for all the concentrations of the complexes that are in principle possiblg. If the experimental values agree with any of the calculated dipole moments for an activated complex, this would give us some information on the structure of the activated complex and whether or not the change in dielectric constant of the medium is the loglow2 only reason for the change in w 2, i.e., whether it is legitimate to apply Eq.(3) to the system we are talking about. 2,I

From the equation for the straight lines of Fig. 5 for ethyl alcohol

~,e 4,e

~

z lgw2 = - - 2,64 -k 9,2 2~ T t

0,e -g and for methyl ethyl ketone

q0-

-qe 2

,

4

I

o,3

I

,~

oAe-t

Fig. 5. Loglowz as a function of ( e - 1)/(28 + 1) in the oxidation of ethyl alcohol (1) and methyl ethyl ketone (2) in benzene mixtures,

lgw~-- 1,t4 + 1,37ff-Tl we find the value of the quantity ~ 7 tz ' equal to 18.7 .10-1s for the alcohol and 1.8 . 10-t3 for methyl ethyl ketone. To find the dipole moment of the activated complex from gq. (4), in addition to the quantity ~ 1~2 r3 , we need

to know the dipole moments of the reacting substances (Pm4 and /~RO-) and their effective radii (rRH and fRO0 . Values of PRH' equal to 1.7 9 t0 ~ CGS traits for ethyl alcohol (at 200;~ and 2 . 7 5 ~ 1 0 - u c G s units for methyl e~nyI ketone (at 145"), were taken from the handbook literature [12]. To find the values of rRH, we used the values for the molecular volumes obtained from the molecular weights and the experimentally determined densities (for ethanol d ~~176 = 0.48 g / c m a, and for methyl ethyl ketone d ~s = 0.6 g/em~). In finding pr, O. and r ~ , we started from the assumption that the constants of the peroxide radical would r~ 2 t~'v2 not be very different from the constants of the corresponding hydroperoxides. After having looked over the density values for a large number of compounds of the same elementary composition as the hydroperoxides of methyl ethyl ketone and ethyl alcohol, we came to the conclusion that density of the material is not strongly dependent on its structure, and calculated r.,-~,, from the mean values of these quantities extrapolated to the temperature of the experiment: for ethyl alcoho~U~r~o~ = 4 . 7 " 10-z3 cm s, and for methyl ethyl ketone r~o. = 5 . 1 0 " 2 w s. The values for the effective radii of the activated complexes were found from the formula r ~ = Zr~lH + r~o ~ [13]. For, ethyl alcohol, r~= 8.5 "10 -m and for methyl ethyl ketone, r3#=9.8 . 1 0 - ~ c m 3. Since the dipole moments of alcohol and ketones are independent of the structure of the radical R, and are determined solely by the presence of the functional groups [12], knowing the dipole moments of these groups and the dipote moment of HzO~ [12] as well as the angles between the C atoms and the bond lengths, we were able to calculate the dipole moment of the RO~ radical. The dipole moment of the peroxide radical of ethyl alcohot is PRO" = 2.3 9 10"18 CGS units. We can imagine two structures for the peroxide radical of methyl ethyl ketone with dipole moments, PROz;equal to 4.4 9 10"18 CGS units (a) and 0.8 9 10-Is CGS units (b):

1107

H I

C H ~,\

CH.~\ ~

>L%--Z,s

o0/0 O~'l

O/ a b Substituting all the values found in gq. (4) we find as the dipole moments of the activated complexes ( g r for ethyl alcohol 11.5 9 10"18 CGS units, and for methyl ethyl ketone 8,4 9 10"xs CGS units (a) and 5 . 8 6 9 1 0 " u c ~ s units (b). When these values obtained from the e x p e r i m e n t a l data were compared with the c a l c u l a t e d values of the dipole moments, consideration was given to all possible structures of the activated complexes. The dipole moments of the activated complexes were c a l c u l a t e d from the parallelogram rule allowing for the fact that the bonds between the C" "" H" "" O atoms which help to form the complex can blow up to one and o n e - h a l f times their value. For the reaction between the peroxide radical and methyl ethyl ketone, values of p very close to the experimental ones were found for the two complexes having the structure H .I

Ctt.~\

H [

0

(11

I

0

O~@ ~

CHs

I H

(II)

I

o0

c'3.. @L -- ~ - - C - - C H ~

')~/,0

I

H

The good agreement between the calculated v a l u e p r = 8.1 9 10"18 and the value p r = 8.4 9 10 "18 CGS units found from the experimental data on the assumption that the peroxide radical has the structure "a" speaks in favor of the a c t i v a t e d complex formed having the structure (II). In addition, the agreement between the experimental and c a l c u Iated values of p r shows that the straight line 2 of Fig. 5 gives the true variation of the rate at which RO~ reacts with methyl ethyl ketone as a function of polarity of the m e d i u m . In the case of ethyl alcohol, not a single one of the possible structures for the a c t i v a t e d complex has a dipole m o m e n t anywhere near the e x p e r i m e n t a l value (/a = 11.5 9 10-is), the calculation giving the m a x i m u m value p ~ = 4.5 9 10 "zs CGS units. Thus, the rate at whichRO~ reacts with ethyl alcohol depends not only on the polarity of the medium but on other factors, for e x a m p l e association of t h e alcohol molecules to form hydrogen bonds. The linear form of curve 1 of Fig. 5 is apparently caused by the fact that the electrostatic forces and the association of the alcohol molecules both act in the same direction of increasing w 2. This is also shown by the fact that the line 1 has a greater slope than line 2, in spite of the fact that /~RH and PROs. are smaller for the alcohol than for the ketone. The role played by hydrogen bonds in the oxidation of ethyl alcohol apparently consists in the fact that it is not just one RH m o l e c u l e reacting with one ROs radical, but two aggregates in each of which the molecules are bound together by hydrogen bonds. Here the RO~ radical associates itself with alcohol molecules, i.e., if we assume that each aggregate consists of n particles, there will be some of them that contain n alcohol molecules and others that contain n - 1 alcohol molecules and one RO~ r a d i c a l . In order to convince ourselves that this assumption is ~2 correct and find the approximate value of n, l e t us write the expression for ~ - ~ as given in Eq. (4) applied to r e actions between two aggregates. Then the effective radii of the reacting aggregates and of the activated complex m a y be conceived of as r~ ---- n. r~tH;

r~ = ( n - - 1) r~H -4- r 3R %."'

r~ -= 2 n . r~H

-

-

r~H -~- r aRo~

Hence after some transformations we obtain an expression for the effective radii of the reacting aggregates and of the activated complex: t

l--z --

t

,),IH

4

where

t

2

(.-' r~u

z

1108

l--z

(n - - ~) ~ a

x+z

)

r~o~ ( . - - t)

r~

' RII

(n

--

I) r~C.H

The dipole moments of the reacting substances and of the activated complex will be x ~RII'

9 ~

x

2

'

+ Ro; 4 2(,,, -

RtI

Here the quantity n x < n takes account of the fact that the dipole moments of the molecules forming the aggregate are vector quantities, and so when they are summed they can partially cancel one another out. To find the dipole moment of the activated complex to a zeroth approximation, we assume that x = z = O, and, substituting all the values found in Eq. (4),we obtain ~-~ rS

' ( n - - t) r~H

[nx (nx

1) ~I~R H -F nxliRii~RO2]

9

Assuming further that PRH ~ P R O " and using the experimentally determined value ~, - ~ = 1.37 9 10- 2 , we 32 n2x I 37.t0 .2 rRH find for n > 1 ' -g X ( ..-7~--~ = 5, which gives the value n = 5 for n x ~ n. A similar calculation n

\ ~tRH ]

for methyl ethyl ketone, as might be expected, gives a value o f n close to unity, i.e., no association of the reacting dipoles occurs in this case. Thus, the assumption that there are aggregates consisting of several particles reacting in the oxidation of ethyl alcohol enables us to understand why the dielectric constant of the medium has a different effect on the rate of the bimolecular reaction between the peroxide radical and the substance undergoing oxidation (wz), depending upon whether we are oxidizing ethyl alcohol or methyl ethyl ketone. For the same dielectric constants in both cases, the value of w 2 for ethyl alcohol decreases more rapidly when it is diluted with benzene than is the case with methyl ethyl ketone,as a result of destruction of the hydrogen bonds. The authors express their thanks to N. M. l~manugl', N. D. Sokolov, S. G. ~ntelis and M. I. Binnik for their valuable advice in discussing the results of this work. SUMMARY 1. The differences in oxidation mechanism of polar organic compounds observed when the reaction is carried out in the gaseous and liquid phases are caused by the fact that the rate of the reaction between the peroxide radical and the substance undergoing oxidation is dependent on the dielectric constant of the medium (methyl ethyl ketone, ethyl alcohol) and on the formation of intermolecular hydrogen bonds (ethyl alcohol). 2. The reaction between the peroxide radical and methyl ethyl ketone or ethyl alcohol is an interaction between two dipoles, 3. A proposal is made as to the structure of the activated complex formed in the reaction between methyl ethyl ketone and the peroxide radical. 4. In the oxidation of ethyl alcohol what is reacting is not individual molecules but aggregates consisting of five or more particles bound together by hydrogen bonds. 1. 2.

3. 4. 5. 6~ 7. 8. 9.

LITERATURE CITED E . A . glyumberg, G. E. Zaikov, and N. M. Emanuel', Dokl. AN SSSR 139, 99 (1961); Neftekhimya 1 , 2 3 5 (1961). E . A . Blyumberg0 G. E. Zaikov, Z. K. Maizus, and N. M. ~manu$1'. Dokl. AN SSSR 13___33,144 (1960); Kinetika iKataliz 1, 510 (1960)o G. E, Zaikov and Z. K. Maizus, Kinetika i Kataliz (1962), N, N. Semenov, Some Problems of Chemical Kinetics and Reactivity [in Russian], Academy of Sciences Press. USSR, po 131 (Moscow, !958). ~o A. Blyumberg, Z. K. Maizus, and N. M. Emanuel', Collection "Liquid Phase Oxidation of Hydrocarbons" [in Russian] Academy of Sciences Press USSR, p. 125 (Moscow, 1959). G. E. Zaikov, Zh, analit, khimii 1_~5,104 (1960); 1._~5,639 (1960), G. E. Zaikov, Zh. analit, khimii i 7 , 1 1 7 (1962). t~. A. Blyumberg, A, D. Malievskii-:-and N. M. E m a n u ~ ' , Dokl. AN SSSR 13_~6,1190 (1961). S~ Glasstone, K. Leider, and G, gyring, Theory of Absolute Reaction Rates [Russian translation] IL, p.404(1948).

1109

10. 11. 12. 13.

G.C. Akerl6f and H. Y. Oshry, J. Amer. Chem. Soc. 72, 2844 (1950). S . G . Entelis, G. P. Kondrat'eva, and N. M. Chirkov, V---ysokomolek.soed. 3, No. 7, 1044 (1961). Lendolt-B~Srnstein, Zahlenwerte und funktionen aus physik, chcmie,astronomie, geophysik, technik, vol. 1, pt. 3 (molekuln II), pp. 389,415 (Berlin, Gottingen, Heidelberg, 1951). S. Glasstone, K. Leider, and G. Eyring, Theory of Absolute Reaction Rates [Russian translation] IL, p. 408 (1948).

All abbreviations of periodicals in the above bibliography are letter-by-letter transliterations of the abbreviations as given in the original Russian journal. S o m e or all of this per/o d i c s l l i t e r a t u r e may well be a v a i l a b l e in E n g l i s h translation. A complete l i s t of the cover-tocover English translations appears at the back of this issue.

1110