2913
J. Phys. Chem. 1981, 85,2913-2916
Kinetics of the Reactions of the Hydroxyl Radical with Xylenes J. M. Nlcovlch, R. L. Thompson, and A. R. Ravishankara' Molecular Sclences Group, Engineering Experiment Station, Georgia Institute of Technology,Atlanta, Georgia 30332 (Received: AprU 21, 198 1; In Final Form: June 10, 1981)
Absolute rate constants for the reactions of the hydroxyl radical with o-, m- and p-xylene are reported for the temperature interval 250 IT I 970 K. Three major reaction pathways are observed. At T I 320 K, electrophilic addition of OH to the ring is dominant,while above 500 K abstraction of hydrogen atoms from the side-chain CH3groups is dominant. In the temperature range 320-400 K, the system is complicated because of the reverse decomposition of the addition product.
Introduction The reactions of OH and other free radicals with aromatic hydrocarbons have been studied, until now, with only their tropospheric importance in mind. Therefore, all investigations have been carried out at temperatures near 298 K. However, these reactions are quite important in combustion systems, and their behaivor at higher combustion temperatures could be different from that seen at 298 K. For this reason we have recently started a series of studies of radical-aromatic hydrocarbon reactions using the direct kinetic technique of flash photolysis-resonance fluorescence which extend measurements to combustion temperatures (Le., T = 1000 K). This paper is the second in this series and describes OH radical reactions with xylenes
-
OH
+ o-xylene
OH
+ m-xylene
OH
+ p-xylene
kl
products
(1)
products
(2)
products
(3)
ka
ka
over the temperature range 250-970 K. It has been previously shown that OH reactions with aromatic hydrocarbons at low temperatures (i.e., T 5 298 K) proceed mainly via addition to the ring;1-9 ring hydrogen abstraction is a negligible channel, and the sidechain hydrogen abstraction channel is small but significant. Furthermore, the variation of the overall reactivity of various aromatic hydrocarbons toward reaction with OH has been correlated with the nature and position of the substituent groups on the aromatic ring,' indicating, again, that electrophilic ring addition is the dominant reaction (1)D. D. Davis, W. Bollinger, and S. Fischer, J. Phys. Chem., 79,293 (1975). (2)A. C. Lloyd, K. R. Darnall, A. M. Winer, and J. N. Pitts, Jr., J. Phys. Chem., 80,789 (1976). (3)D. A. Hansen, R. Atkinson. and J. N. Pitts, Jr., J. Phvs. Chem., 79,1763 (1975). (4) R. A. Perrv. _ .R. Atkinson. and J. N. Pitts.. Jr.., J.Phvs. Chem.., 81.. 29s'ii977). (5)G.Doyle, A. C. Lloyd, K. R. Darnell, A. M. Winer, and J. N. Pitts, Jr., Enuiron. Sci. Technol., 9,237 (1975). (6)R. A. Perry, R. Atkinson, and J. N. Pitta, Jr., J.Phys. Chem., 81, 1607 (1977). (7)A. R. Ravishankara, S. Wagner, S. Fischer, G. Smith, R. Schiff, R. T. Watson, G. Tesi, and D. D. Davis, Znt.J. Chem. Kinet., 10,783(1978). (8)R. A. Kenley, J. E. Davenport, and D. G. Hendry, J.Phys. Chem., 82, 1095 (1978). (9)F.P.Tully, A. R. Ravishankara, R. L. Thompson, J. M. Nicovich, R. C. Shah, N. M. Kreutter, and P. H. Wine, J.Phys. Chem., 85,2262 (1981). 0022-3654/81/2085-2913$01.25/0
pathway. Since meta substitution of the ring enhances electrophilic addition, it is easy to understand why mxylene reacts faster than either p-xylene or o-xylene with OH at 298 K; Le., kz > kl and kz > k3, while kl = ka. Before this work, the only kinetic measurements of the temperature dependence of OH-xylene reactions which have been made are those of Perry et alS4Using the flash photolysis-resonance fluorescence technique over the temperature range 296-473 K, these authors found sharp structures in In k vs. 1/T plots of the overall rate constant (for the disappearance of OH). Perry et al. interpreted their results in terms of variations with temperature of the branching ratios for the competitive reaction channels. The results of our study confirm their interpretation and provide branching ratios as a function of temperature.
Experimental Section Our utilization of the flash photolysis-resonance fluorescence technique in the study of OH(X211)radical reaction kinetics is amply described in the 1iterature.lO Recently, we have extended the temperature range of this method to T > 1000 K. A detailed description of the modifications required to allow measurements at high temperatures is given el~ewhere.~JlThe experimental procedure adopted during the current study was identical with that employed in the study of OH + benzene, toluene reactionsg and hence will not be repeated here. OH radicals were produced by flash photolysis of HzO at wavelengths between 165 nm and the onset of continuum absorption at 185 nm. The initial hydroxyl radical concentrations, [OHIO,were in the range 2 X lolo-1 X loll em9, while the concentration of the aromatic hydrocarbon, RH, was greater than 4 X 10l2~ m - the ~ ; ratio [RH]/[OHIo was always greater than 100. Therefore, pseudo-first-order conditions prevailed during all experiments, and secondary reactions did not contribute to the measured OH decay rates. The temporal profile of OH under the pseudofirst-order conditions is given by the equation [OH], = [OH]oe-(k[RHl+ kd)t = [OHIoe-k't where k 'is the measured pseudo-first-order rate constant (Le., slope of In [OH] vs. t plot) and k is the bimolecular rate constant for the reaction OH
+ RH
k
products
(10)P.H. Wine, N. M. Kreutter, and A. R. Ravishankara, J. Phys. Chem., 83,3191 (1979),and references therein. (11)A. R. Ravishankara, J. M. Nicovich, R. L. Thompson, and F. P. Tully, J.Phys. Chem., 85, 2498 (1981). (12)E. D. Morris, Jr., and H. Niki, J. Phys. Chem., 75, 3640 (1971).
0 1981 American Chemical Society
2914
The Journal of Physical Chemistty, Voi. 85, No. 20, 1981
[RH] is the constant aromatic hydrocarbon concentration and kd is the first-order rate constant for loss of OH in the absence of RH. k’was measured as a function of [RH] and k derived from the slope of k’vs. [RH] plots. In order to avoid the accumulation of photolysis or reaction products and to minimize any uncertainties in [RH] arising from aromatic hydrocarbon adsorption on the reactor walls, all experiments were carried out under “slow flow” conditions. The flow rate through the cell was such that each photolysis flash encountered a fresh reaction mixture (photolysis repetition rate, -0.3 Hz). The aromatic hydrocarbon RH was taken from a 12-L bulb containing an RH/diluent gas mixture, and the water mixture was generated by bubbling diluent gas at 800 torr through distilled water at room temperature. The RH/diluent gas mixture, the HzO/diluent gas mixture, and additional diluent gas were mixed before entering the reaction cell. The concentration of each component in the reaction mixture was determined from measurements of the appropriate mass flow rates (measured by using calibrated mass flowmeters) and the total pressure. The fraction of aromatic hydrocarbon in the RH/diluent gas mixture was checked frequently by simultaneous measurements of the aromatic hydrocarbon absorption at 253.7 nm and the total pressure of the mixture. These determinations were carried out by using a Hg pen-ray lamp as the light source, an 80-cm long absorption cell, and a photomultiplier tube fitted with a band-pass filter. The absorption cross sections at 253.7 nm used to calculate the RH concentrations in the source mixtures were measured during the course of the experiments; the obtained values are as follows: o-xylene, 5.98 X cm2;rn-xylene, 5.22 X cm2;and p-xylene, 6.16 X cm2, The absolute [RH] values quoted for each experiment are accurate to better than 8%. The diluent gases used in this study had the following stated purities: Ar > 99.9995% and He > 99.9999%. o-Xylene and rn-xylene were obtained from Eastman Kodak Co. and had stated purity levels of greater than 99.9%. p-Xylene was from Baker Chemicals Co. (Ultrex) and had a stated purity of >99.99%. All three xylenes were degassed before use.
Results and Discussion Three different types of kinetic behavior were identified with specific temperature ranges for all three reactions. At T < 320 K (335 K for p-xylene), the measured [OH] decays were exponential and the measured values of kl, kz, and k3 were nearly independent of temperature. In the temperature range 320 < T < 400 K, nonexponential [OH] decay curves were observed. The leading portion of the nonexponential decay was faster than the tail. At temperatures above 400 K, [OH] decays were exponential and the measured values of kl,kz, and k3 increased monotonically with temperature. This general behavior is identical with that seen in the reactions of OH with benzene and tol~ene.~?~ Table I lists as a function of temperature the measured bimolecular rate constants. The quoted errors are (2a) overall error estimates which include errors in concentration measurements. For those temperatures characterized by nonexponential [OH] decays, the listed k values are identified by a superscript. These values are included merely to reflect the decreasing trend in reactivity through this temperature region. The values compiled in Table I are also shown in Figure 1(for the sake of pictorial clarity, the error bars are not included). As pointed out earlier, all experiments were carried out under pseudo-first-order conditions. The initial concentration of OH, the photolysis energy, and the diluent gas
Nicovich et al.
TABLE I: Rate-Constant Data for the Reactions of OH with Xylenes 10’2k,acm3 molecule-’ s-l temp, K o-xylene m-xylene p-xylene 250 26.5 * 2.5 269 25.6 i 4.3 298 14.2 f: 1.7 25.4 * 3.5 13.5 * 1.4 320 15.8 * 1.8 13.8 rt 1.1 330 5.2b 335 12.5 i 1.3 357 5.1 4.3b 400 2.39 f 0.20 2.47 0.41 1.71 * 0.28 484 3.70 i 0.64 508 4.19 * 0.48 3.44 f 0.34 526 3.40 f 0.48 576 5.42 f 0.45 4.60 i 0.54 5.03 * 0.88 647 6.87 i 0.91 6.01 * 0.59 684 6.2 .I 1.1 757 10.20 f 0.91 9.3 f 1.1 9.66 i 0.85
*
30T
o
m-XJene
p
20
“E
IO
0
o o-Xylene 0
0
A p-Xylene
..
‘0
A
t ‘ I
5 A
2
3
4
1000 T (K) Flgure 1. Arrhenius plots of In kvs. 1000/ T(K) for reaction of OH with o-xylene (0),m-xylene (a),and p-xylene (A). The filled polnts refer to data obtained under conditions where nonexponentlal [OH] decays were observed.
pressure were varied at selected temperatures to investigate the possible dependence of kl, kz, and k3 on these parameters; no dependence was found. This confirmed our expectations that the measured values of kl,k2, and k3 are devoid of contributions from secondary reactions involving photolysis fragments and reaction products. At the highest temperatures, the residence time of the xylenes in the reactor were also varied to check for the possible thermal decomposition of these compounds; again, no evidence was found for such a process. Based on the results of previous investigations and the present study, a consistent picture for the mechanism of OH reactions with aromatic hydrocarbons has emerged.41~ At low temperatures, i.e., T 5 320 K, OH predominantly adds to the aromatic ring. The adduct formed in this reaction is thermally stable toward decomposition into the
Kinetics of the Reactions of
The Journal of Physical Chemistty, Vol. 85, No.
OH with Xylenes
20, 198 1 29 15
TABLE 11: Summary of Rate-Constant Data for Reactions of OH with Xylenes at 298 K 101lk, cm3 molecule-ls'' compd o-xylene m-xylene p-xylene techniqueb
ref
12
2
18.7a
2 1 . 5 * 4.3
DFMS
EC
3
4
5
7
this work
15.3 f 1.5 23.6 f 2.4 12.2 ?: 1.2 FPRF
14.3 f 1.5 24.0 t 2.5 15.3 * 1.7 FPRF
1 2 . 8 t 3.8 23.2 t 1.7 12.3 2.5 EC
12.4 0.10 20.6 * 1 . 3 1 0 . 5 0.60 FPRF
* *
14.2 t 1.7 25.4 t 3.5 13.5 jl 1.4 FPRF
*
a A mixture of all isomers was studied. EC = environmental chamber; k measured relative to the rate coefficient for the reaction of OH with n-butane; h o ~ + ~= 3.0 . bX ~ ~ cm3 ~ molecule-' s-'. FPRF = flash photolysis-resonance fluorescence. DFMS = discharge flow-mass spectrometry.
reactants. As the temperature is raised, however, the adducts become thermally unstable and their rates of decomposition into OH and aromatic hydrocarbon increases rapidly with temperature in the range 320-400 K. Eventually, by -400 K, the decomposition is sufficiently rapid that OH is not observed to react with the aromatic hydrocarbon via addition. It is the varying rates of adduct decomposition that account for the observed non-firstorder kinetics in the 320-400 K temperature range. At temperatures above 500 K, reaction proceeds exclusively via hydrogen atom abstraction. Of course, both the ring hydrogen and the side-chain hydrogen atoms can be abstracted. As first explained by Perry et ala4and confirmed in the present study, reactions 1-3 follow the above scheme. At this point we can compare our 298 K values of kl,k2, and k3 with previous results. As noted above, the 298 K reactions 1-3 proceed mainly via addition. However, no pressure effect is observed since the reactions have already reached their second-order high-pressure l i m i t ~ . 4Table ~~~~ I1 lists the values of k3 measured in all reported studies. It is seen that, in general, the agreement between the present study and all previous investigations is excellent. Only in the case of p-xylene, there is a disagreement of -30% between the results of Ravishankara et al.7and the present study. Ravishanha et al.' utilized a static reactor, and the loss of p-xylene to the walls was probably significant, thereby yielding an apparent lower rate constant. In fact, during the present study both o-xylene and p xylene were found to be quite "sticky" at T I298 K, such that these compounds could not be studied at temperatures below 298 K. The values of kl,k2,and k3 in the temperature range 320-298 K (320-250 for reaction 2) are nearly independent of temperature. This behavior compares extremely well with the results of Perry et al.,4 who saw essentially no change in the rate coefficient over the temperature range 298-325 K. In addition, this behavior can be compared with that of the reactions of benzene and toluene with OH in the same temperature range? where again very little temperature dependence is found. At higher temperatures (Le., T > 500 K) the values of lzl, kz, and k3 increase monotonically with temperature and seem to follow Arrhenius behavior. (The quality of our data is such that any subtle non-Arrhenius behavior that might be present is not discernable.) We have fitted our high-temperature data to the following Arrhenius expressions: ki(7') (6.5 f 1.1) X 10-l' exp[-(1.42 f 0.12) x 1o3/q cm3 molecule-'s-'
k2(7') = (6.8 f 2.3) X 10-l' exp[-(1.54 f 0.24) X 103/T] cm3 molecule-' s-l k3(7')
(6.4 f 2.4) X exp[-(1.44 f 0.25)
X
103/r] cm3 molecule-' s-l
for 500 < T < lo00 K. All quoted errors are 2a values and Q A = Aam. From the above three expressions (and Figure l), it is very clear that the high-temperature rate constants for the reactions of OH with xylenes is independent of the position of CH3 groups on the aromatic ring. These values are in reasonable agreement with Arrhenius parameters kl = (5.0 f 6.3) X 10-l' exp[-(152 f 758)/q cm3 molecule-' s-l ilzz = (5.0 f 0.6) X
lo-" exp[-(ll60 f 7 5 8 ) / T ] cm3 molecule-'^-^
k3 = (6.3 f 0.6) X exp[-(1212 f 7 5 8 ) / T ] cm3 molecule-'^-^ derived by Perry et aL4 which covered a much smaller temperature range and had very large error bounds. Comparison of the measured values of kl, Itz, and k3 with those for the reaction of OH with benzene OH
k4
C6H6 -* C&5' -I-H2O
(4)
and toluene OH
+ C6H5CH3--k C6H4CH3.+ HzO
kw
C&&H2.
+ H2O
(5a) (5b)
-
shows that kl,k2, and k3 are -2 times as large as k5 = (k5a + k5b) and 10-20 times larger than k4 in the temperature range 500-1000 K. We9 have shown how kSaand k5bcan be approximately separated by making the assumption that k5* = (5/6)k4; i.e., the ring hydrogen abstraction in reaction 5 is neither hindered nor helped compared to reaction 4 by the presence of the methyl group in toluene. We calculated kSb (/asHin our previous notation) to be kbb = ksH = (2.4 f 0.6) x exp[-(1.3 f 0.2) x i03/T] cm3 molecule-l s-l (I) This approach can be easily extended to xylenes (by subtracting (2/3)k4from kl,k2, and k3) and it yields for oxylene kslH = (4.9 f 0.9) x IO-'' exp[-(1.32 f 0.12) x 1 0 3 / q cm3 molecule-' s-'
(114 for m-xylene ksZH = (4.6 f 2.1) X 10-l' exp[-(1.39 f 0.31) X 1 0 3 / q cm3 molecule-'^-^ (IIb) and for p-xylene k ~ =3(4.6 ~ f 1.9) X lo-'' exp[-(1.33 f 0.27) x 1 0 3 / q om3 molecule-l s-l (IId
2916
J. Phys. Chem. 1981, 85, 2916-2923
Again, the quoted errors are 21s and uA = A a M . Three points become clear when expressions IIa-c are compared with each other and with expression I. First, the rate constant for the side-chain hydrogen abstraction in xylenes by OH is independent of the CH3 group position (this is not very surprising and was obvious from Figure 1). Secondly, the activation energy for side-chain hydrogen abstraction is independent of the number (and position) of the CH3 groups on the ring. Lastly, the side-chain hydrogen abstraction rate constant in xylenes is twice that in toluene (kslH = kSZH= kS3H=.2kBH).This result again is reasonable since there are twice as many abstractable side-chain hydrogen atoms in xylene as there are in toluene. Finally, extrapolation of expressions IIa-c to 298 K yields the branching ratios for side-chain hydrogen ab-
stractions at that temperature. We get -4% side-chain hydrogen abstraction for reactions 1 and 3 and -2% for reaction 2 at 298 K. This compares reasonably with the results of Perry et al. except for reaction 1,for which they which would yield a lower limit of 10%. obtain 0.203;:, This number is too high compared to our result and, in fact, there is no obvious reason why the branching ratios for reactions 1 and 3 should be very different.
Acknowledgment. We thank Mr. N. M. Kreutter and Mr. D. Semmes for carrying out some of the experiments reported here and Dr. P. H. Wine for his help during the course of this investigation. This work was supported by the Division of Chemical Sciences, Office of Basic Energy Sciences, U S . Department of Energy, under Contract No. ER-78-S-05-6030.
Electron Spin Resonance Study of Oriented Allyl-Type Free Radlcals of Butene and Butadienecarboxylic Acids and Derivatives L. Muszkat" Department of Structural Chemistty, The Weizmann Institute of Science, Rehovot, Israel (Recelved: June 24, 1980; In Final Form: April 16, 198 I )
Single-crystalelectron spin resonance studies of free radicals formed by y irradiation were performed on a series of unsaturated carboxylic acids and their derivatives. As a rule, allyl-type extended ?r-electronfree radicals are the most persistent species observed in the butene series molecules (2- and 3-hexenedioicacids) and in the
butadiene series molecules (sorbic acid, sorbamide, and dimethyl muconate).
Introduction In a previous study of persistent free radicals formed by the y irradiation of unsaturated dinitriles, two different structure types could be 0bserved.l In muconodinitrile, NCCH=CHCH=CHCN, the localized imino radical NCCH=CHCH=CHCR=N is formed by an addition reaction, while in 3-hexenedinitrile, NCCH,CH=CHCHzCN, the extended allyl radical NCCH,CH=CH;-;CH=CN. is obtained by a hydrogen atom abstraction pr0cess.l As a continuation of this research the present work is concerned with the crystal-state radiation chemistry of unsaturated molecules containing the same four carbon chain skeleton of the butene and of the butadiene series, XCH,CH=CHCH,Y and XCH=CHCH=CHY. In these two series (Table I) were included molecules in which the two end groups are -COOH, -COOCH3, -CONHz, and CH3. (For preliminary results obtained with other unsaturated systems, e.g., muconic acid, dipotassium muconate, and monoalkyl and dialkyl fumarates, see ref 24.) Quite unexpectedly, the present work indicates that solid-state y irradiation of unsaturated molecules of these two series gives rise in a very uniform way to the corresponding allyl radical system XCH,CH;-;CH;-;CHY, irrespective of the substrate molecule. Thus the butene series (I and 11)react by loss of a hydrogen atom while the butadiene series (111-V) react by the addition of a hydrogen atom (or of a carbon-centered free radical Re). Some of *Address correspondence to this author at the Pesticide Chemistry and Residue Research Laboratory, Agricultural Research Organization, The Volcani Center, Bet Dagan, Israel.
the reasons for such reactivity patterns have been considered at length elsewhereS2 The present results obtained with both butene and butadiene series widen considerably the range of known extended a free radical systems for which information from single-crystal studies is available. Some of the systems included in the present study provide exceptions to the common tenets on free radical guest-crystal host structural relationships. (a) In two cases y-ray-damaged molecules were observed to assume positions differing from those of the undamaged molecules. This finding is a t variance with the common conclusion7Jsof single-crystal studies that the atoms of the free radical occupy the same positions as in the undamaged molecules. (b) While free radicals trapped in crystalline matrixes are present usually in one singular conformation, dimethyl muconate is in this respect an exception yielding one and the same free radical system existing in two distinct molecular conformations. Among the previously described extended systems for which single-crystal results are available, one should note the allyl systems HOOCCH;-;CH;-;CHCOOH in glutaconic acid,3 CH,CH;-;CH;-;CCOOH in furoic acid,4 and the 0extended molecular system in potassium hydrogen maleate.6
-
(1) L. Muszkat, Int. J. Radiat. Phys. Chem., 7, 597-602 (1975). (2) S. Sharafi-Ozeri, L. Muszkat, and K. A. Muszkat, 2.Naturforsch. A , 31, 781-5 (1976). (3) H. C. Heller and T. Cole, J. Chem. Phys., 37, 243 (1962). (4) R. J. Cook, J. R. Rowlands, and D. H. Whiffen, Mol. Phys., 7, 57 (1963).
0022-3654/81/2085-2916$01.25/00 1981 American Chemlcal Society