Sequence Studies in Liquid Phase Hydrocarbon OxidationPart 4.-Hydroperoxide-Alcohol and Hydroperoxide-Ketone Transitions in theOxidation of EthylbenzeneBY &A DAN~CZY, ISTVAN NEMES AND DEZS~ GAL*Central Research Institute for Chemistry of the Hungarian Academy of Sciences,1025 Budapest, Pusztaszeri ut 59-67, HungaryReceived 10th March, 1976The rates of consumption of cc-phenylethylhydroperoxide molecules during the oxidation ofethylbenzene at 120°C: have been measured using radioactive tracer procedures ; balance equationsfor the total activities and concentrations were used for the calculations. In order to obtain a coherentset of rates a hydrogen transfer reaction between hydroperoxide molecules and peroxy radicals had tobe taken into account.It was established that ketone and alcohol molecules were formed mainly as a result of the induceddecomposition of the hydroperoxide and the contribution to their formation by termination pro-cesses was <lo % of their total formation rates.The formation rates of alcohol and ketone viahydroperoxide molecules were very similar.The relative reactivity of hydroperoxide and ethylbenzene molecules toward peroxy radicals wasfound to be two. Further kinetic data, such as ease of oxidation and kinetic chain length both forthe oxidation of the parent hydrocarbon and its main intermediate, have been determined.In previous papers we have reported on the kinetics and mechanism of theconsumption of 1-phenylethanol (ROH) formed during the oxidation of ethylbenzene.In the present work we report on the transformation of the main intermediate, a-phenylethylhydroperoxide (ROOH) in the course of the oxidation.Pritzkow and Holm and Emanuel et aL6 performed earlier studies on the trans-formation of hydroperoxides in oxidative systems.EXPERIMENTALMATERIALSEthylbenzene (RH) was used after careful purification as described previously.1-214GEthylbenzene was prepared by the reduction of 4C-acetophenone (RCOR”) syn-thesized from Ba14C03 by a Grignard reaction followed by a Friedel Crafts process. Theradiochemical purity after distillation was 100 % and the molar activity of the product was3.4 Ci mol-I.14C-a-Phenylethylhydroperoxide was obtained by oxidation of 14C-RH using azobis-isobutyronitrile (AIBN) as initiator or sodium stearate as catalyst.After removal of thewater-soluble 2-cyanoprop-2-ylhydroperoxide, formed from AIBN, the ROOH was pre-cipitated as its sodium salt, washed thoroughly and neutralised with hydrochloric acid. Theproduct so obtained was then dried and its purity checked by gas chromatography. Molaractivity of the 14C-ROOH was 0.8 Ci mol-1 and its radiochemical purity was 92 % ; RCOR”(2.5 %) and ROH (1.2 %) were found as chemical impurities.ANALYSISGas chromatography and radio gas chromatography were used. Working conditionsexcept that in order to obtain a sharp peak for ROOH, the were as previously reported13I36 LIQUID PHASE HYDROCARBON OXIDATIONcolumns were filled with 10 ”/, Ge XE 60 on Gas Chrom S (120-140 mesh) and operated at75°C with argon as carrier gas.REACTION CONDITIONSThe experimental technique has been described previously.‘RESULTSFrom the experiments carried out at 120°C in the presence of 14C-ROOH theamounts of ROOH formed and consumed during the oxidation were calculated bythe kinetic isotope method.’.However, the amounts of ROOH consumed exceededthe equivalent amounts of the products formed from the ROOH and observedanalytically. Possible explanations for this discrepancy may be : (i) During theconsumption of the hydroperoxide further products are formed which have not beendetected by our analytical method, i.e. the expression assumed in Part 3where ARH is the amount of ethylbenzene consumed, is not valid. (ii) The sequencenetwork * assumed in Part 1 does not represent the overall process correctly becauseof unaccounted reactions.[ARH] = [ROQH] + [ROH] + [R’COR] +[phenol] (1)CONTROL OF THE VALIDITY OF THE BALANCE EQN ( I )Labelled ethylbenzene with a total activity of 22.7 mCi dm-3 was oxidized for50 11.Samples were analysed for the total activities of RH, ROOH, ROH and R’CORand for the total amount of the latter three compounds. The amounts of RHconsumed were calculated from its total activity losses and from the activity increases1.6I .4m 1.2Ec) 1.0E .s 0 . 820 . 68I-3& . c -0 -40.20.time/hFIG. I .--Kinetics of product formation in the oxidation of ethylbenzene. [ROOH] + [RC’OR] +[ROHI f [phenol] against time, + sum of product concentrations from total activity increase, x sumof product concentrations from activity decrease of ethylbenzeneE. DANOCZY, I .NEMES AND D. GAL 137of the products, and compared with the total amounts of products determined analy-tically. The results are given in fig. 1 and show that the amounts of ethylbenzeneconsumed as calculated by the balance equation (1) and based on activity measure-ments agree to within + 5 %. The results also indicate that eqn (1) can be used upto 15-20 % conversion.DETERMINATION OF THE FORMATION A N D CONSUMPTION RATES OF THEROOH I N THE OXIDATION OF ETHYLBENZENEThe amounts of labelled ROOH introduced into the system varied between 3.5and 33 x It was established previously that the introduction ofROOH causes a shift to higher conversions, but the kinetic curves can be matchedby a simple translation along the tiEe axis.in01 dm-3.DISCUSSIONPrior to the quantitative calculations we had to determine those processes whichare responsible for the activity decrease of the ROOH without its simultaneous con-centration change.The most likely such process is the hydrogen transfer reaction(2) investigated in detail by Howard and co-workers ’* l owhere RO: are a-phenylethylperoxy radicals. Although this reaction is ari identityprocess in the present system, since ROOH is introduced in labelled form, reactioii(2) might nevertheless result in activity losses if its rate is significant.Ingold and co-workers have fouiid that the rate constant of reaction (2) variedbetween 102-103 dm3 mol-1 s-l.A similar rate constant was obtained by Niki etal.’ 9 l2 at 60°C. Values observed by Thomas and Tolman and by Hiatt, Gouldand Mayo l4 are also -lo2. Consequently we included the hydrogen transfer pro-cess in the sequence network :RO’, + ROOH + ROOH + RO’, (2)w13I RONi -3)K’COR”‘“1 4where ivXy are the corresponding rates and * refer to the transfer processes.Since, the total activities, molar activities and concentrations of the peroxy radicalsare not known, the equations of the kinetic isotope method applied earlier 1 * cannotbe used. Consequently we introduced the balance equations for the total activitiesand concentrations :dA,,,H = - a 2 w 2 3 - a 2 ~ 2 4 - a 2 ~ ~ 1 + a l ~ ~ 2 + a 1 ~ v 1 2 = b,dt (4)r 5 ) d A,,, - = a 2 ~ ~ 2 3 + a l w , , - a 3 w 3 , = b6d138 LIQUID PHASE HYDROCARBON OXIDATION= a 2 ~ ~ 1 - a , w ~ 2 - a l w ~ 3 - a l w , 4 = b, dA,o;dt (7)whers a l , a2 and a3 refer to the molar activities of RO:, ROOH and ROH, respec-tively.Values of al, [RO’,], w ; ~ and wY2 are not known.We expressed certain rates by using reliable rate constants from literature dataand thus obtained the values of al. According to Tsepalov and Shlyapintokh l5k13 varies between 1.9 and 2.2 x lo7 dnl3 mol-1 s-l while Gadelle and Clement l6and Howard et aL9 give k I 3 = 2 x lo7 dm3 mol-’ s-l.Tsepalov and Shlyapintokh foundk12 = 9.6 x lo5 exp(-8500/RT)giving k12 = 18 dm3 mol-1 s-l at 120°C. According to Gadelle and Clement l6k12 = 20.6 dm3 mol-l s-l.In our calculations we have used the average values at120°Ckigo = 2 x lo7 dm3 mol-l s-land= 19.2 dm3 mol-l s-I.In addition, values of w33 given in Part 1Using the above literature data and assuming that a fast equilibrium takes placewere used.in the hydrogen transfer process, i.e. wT2 N wil, eqn (4)-(11) yieldFor long kinetic chains and at not too high conversionsw12 = [al(b2 + b3 + b4)-(a2b2 + b6 + b7)l/(al -a2)* (1 2)w1 = kl 2[RH](~i/2kl 3)9 (1 3)where wi is the total rate of initiation. Since Wdeg, the rate of the degenerate branching,substantially exceeds w ~ , ~ , the rate of primary initiation, the latter can be neglectedOn the other hand, the rate of initiation equals the rate of termination, wi =~ 1 3 + w , ~ . Recent results indicate l7 that ketone molecules can be formed fromperoxy radicals by the isomerization and subsequent decomposition of the latterinto ketone and 6 H radicals.Therefore, we express the rate of termination as 2w13assuming a Russel-type termination mechanism. Thus, wi = 2 ~ ~ 3 .From eqn (4)-(11) values of w13 are given bySO that M’i = Wdeg.w13 = [a2(b3 +W34)-b6-a3w3411(a2-al). (148. DANOCZY, I. NEMES AND D. GAL 139From eqn (12)-(14) we obtain a quadratic equation for al+b3 + b4)2 - al[2(b2 + b3 + b4)(a2b2 + b6 + b7) + c] + (a2b2 + b6 + b7)2 + a2c = 0(15)(16)whereC = (G2CRH12/k13)[bci +a3W34-&(b3 + W34)I.Calculations were carried out for the initial 25 h.After solving eqn (15) for al we can calculate the corresponding rates from eqnSome of the values obtained are given in table 1.(4141 1).TABLE 1 .-MOLAR ACTIVITIES, CHAIN LENGTH (v) AND HYDROGEN TRANSFER RATES CALCULATED[14C-ROOH] = 33 x mol dm-35295132175220270315360410460515570625680735865.51006.3937.11081.41092.31352.81479.51543.51235.61025.5930.0860.8825.3682.8580.2470.2254.7181 .o134.8104.883.369.5659.050.243.237.132.1528.224.9721.9521 5.0184.0157.0146.0137.0132.5128.3123 .O112.0102.094.589.586.583.078.72.73.23.64.24.95.66.37.99.9511.512.613.013.6514.315.56.967.427.267.447.37.37.08.088.89.259.559.79.9Fig. 2 shows the total rate of consumption (giving alcohol and ketone molecules)of ROOH and the formation rates of alcohol and ketone in the pathway, omittingthe intermediate formation of ROOH.The main source of the alcohol and ketoneformation is the transformation of hydroperoxide molecules. The data also showthat ROOH molecules are transformed into alcohol and ketone molecules with similarprobabilities wZ3 - ~ 2 4 .We can calculate the relative reactivities of the hydroperoxide and ethylbenzenemolecules toward the chain carrier radicals, if the following considerations are takeninto account. The total consumption rate of ROOH, ( ~ 2 3 + ~ 2 4 ) consists of threedifferent rates ; thermal decomposition into radicals, induced decomposition andthermal decomposition into molecules. The last one, according to the data of Emanueland co-workers,18 can be neglected.As mentioned already, the rate of the thermaldecomposition into radicals (degenerate branching) is equal to 2w1 3. Consequently,the differences ~ 2 3 + ~ 2 4 - 2wl correspond to the rates of induced decomposition ofthe ROOH molecules, which are plotted in fig. 3 against [ROOH].Comparing fig. 2 and 3 reveals that the consumption of ROOH proceeds mainlyvia an induced decomposition. Since the " correction " values 2wI3 are very small(and consequently bear the highest error) it was necessary to verify the calculationsindirectly1 40 LIQUID PHASE HYDROCARBON OXIDATIONE 1.0100 2 0 0 3 0 0 400 5 0 0 600 7 0 0[ROOH]/103 mol dm-3FIG. 2.-(n) Total consumption rate of ROOH (wz3 + w24) and (b) termination rates ( ~ ~ 3 , w14) againstthe actual [ROOHJ concentrations.According to literature data l8 the decomposition rate constant of ROOH intoradicals at 120°C is 0.56 x s-l.According to our earlier data it lies betweens-l. Present results, calculated from 21v13, yield a value of 4 x lo-'s-l, in good agreement with the above values though somewhat lower than thosemeasured directly using inhibitors,Assuming that the prevailing process which determines the induced decompositionof the ROOH is its reaction with a-phenylethylperoxy radicalsandkROOHRO', + ROOH '-+ product +radicaland knowing the rate of the induced decomposition of ROOH (shown in fig. 3) wecan calculate the ratio kFooH/k~H at 120°C where kFH refers to the processkpRHRO', + RH -+ Ro+ ROOH.The value of this ratio is -2, less than the kFoH/kEH = 4.3 calculated previouslyand should be considered as an average, since the calculations were carried out upto [ROOH] concentrations at which the contribution of dimers in process (17) isvery likely.This assumption is supported by attempts to calculate the absoluterate constant of process (2), where it was established that the k* values, though theirorder of magnitude is in good agreement with literature data, lo3 dm3 mol-l s-I, arenot constant but decrease reproducibly with increasing [ROOH] concentrationsi. DANOCZY, I. NEMES A N D D. GAL 141A similar phenomenon was observed by Niki et aZ.ll* l2 Howard et aZ. attributedWe computed some further values ; the oxidation rate of ethylbenzene, 4.3 xthis to dimer formation.inoI-% dms s-g, and the oxidative decomposition of ROOH, 8.4 xs-?.are given in table 1.mol-3 dm3The changes in the kinetic chain length during the oxidation of ethylbenzeneX13: 233 3 3 3 400 5QE 530 7rJ9[ROOH]/103 mol dm-3FIG.-;.-Rates of the induced decomposition of the hydroperoxide (wZ3 + 1u24-2~v13) against theactual [ROOHJ concentrations.Since the rates of consumption of ROOH are known and the rate of the initiationis constant throughout the whole system, it was possible to calculate the kinetic chainlength of the oxidative consumption of ROOH. These are also given in table 1 andare seen to increase with increasing conversion.We thank Drs. F. Dutka, A. Mirton and T.Komives for preparing the 14C-elhylbenzene.E. Danbczy, G. Vasviri, J. Phys. Chem., 1972, 76,2785.T. Vid6czy. 8. Dan6czy and D. Gh1, J. Phys. Chem., 1974, 78,828.8. Danbczy, I. Nemes, T. Vidbczy and D. Ghl, J.C.S. Faraday I., 1975, 71, 841.' D. GAl, 8. Danbczy, I. Nemes, T. Vidbczy and P. Hajdu, Ann. N. Y. Acad. Sci., 1973,213,51142 LIQUID PHASE HYDROCARBON OXIDATIONW. Pritzkow and I. Holm, J. pvakt. Chem., 1962, 16,287.N. M. Emanuel, Z. K. Maizus and L. G. Privalova, Int. J. Appl. Rad. Isotopes, 1959, 7, 111. ’ M. B. Neiman and D. 681, The Kinetic Isotope Method and its Applications (Elsevier, Amster.dam, 1971).* I. Nemes, L. BotBr, T. Vidbczy and D. Ghl, in press.J. A. Howard, W. J. Schwalm and K. U. Ingold, Adv. Chem. Ser., 1968,75,6.J. A. Howard and J. H. B. Chenier, Canad. J. Chem., 1975,53,624.E. Niki and Y . Kamiya, J. Fac. En.u. Uiiiv. Tokyo, 1972, 31, 4.l2 E. Niki, K. Qkayasu and Y. Kamiya, Int. J. Chem. Kinetics, 1974, 6, 279.l3 J. R. Thomas and C. A. Tolman, J. Amer. Chem. Soc., 1962, 84, 2079.l4 R. Hiatt, C. W. Gould and F. R, Mayo, J. Org. Chem., 1964,29, 3461.V. F. Tsepalov and V. Ya. Shlyapintokh, Kinetika i Kataliz, 1962, 3, 870; V. F. Tsepalov,V. Ya. Shlyapintokh and P. M. Shou, Zhur. Fiz. Khim., 1964, 38, 351.l6 C. Gadelle and G. Clement, Bull. SOC. chim. France, 1967, 1175 ; 1968, 44.l7 G. E. Zaikov, Z . K. Maizus and N. M. Emanuel, Kinefika i Kataliz, 1966, 7, 401.l 8 I. P. Skibida, Z . K. Maizus and N. M. Emanuel, Neftekhimiya, 1964, 4, 82.l9 T. Vidbczy, personal communication.(PAPER 6/482