7838
J. Phys. Chem. 1989, 93, 7838-7840
Kinetic Measurements of the Gas-Phase Reactions of OH Radicals with Hydroxy Ethers, Hydroxy Ketones, and Keto Ethers Philippe Dagaut,+ Renzhang Liu, Timothy J . Wallington,t and Michael J. Kurylo* Chemical Kinetics Division, Center f o r Chemical Technology, National Institute of Standards and Technology,$ Gaithersburg, Maryland 20899 (Receiued: March 17, 1989; In Final Form: M a y 23, 1989)
Absolute rate constants were determined for the gas-phase reactions of hydroxyl radicals with a series of hydroxy ethers as well as the simplest hydroxy ketone and keto ether with use of the flash photolysis resonance fluorescence technique. At 298 K, the measured rate constants were as follows (in units of cm3 molecule-I s-I): 2-methoxyethanol, 12.5 f 0.7; 2-ethoxyethanol, 18.7 f 2.0; 2-butoxyethanol, 23.1 f 0.9; 3-ethoxy-l-propanol, 22.0 f 1.3; 3-methoxy-1-butanol, 23.6 f 1.6; acetol, 3.0 f 0.3; and methoxyacetone, 6.8 h 0.6. The kinetic data for 2-methoxyethanol obtained between 240 and 440 K were used to derive the following Arrhenius expression: k , = (4.5 f 1.4) X IO-'* exp[(325 f 100)/T](cm3 molecule-' s-'). The results for all seven reactants are discussed i n terms of the prediction of OH rate constants for oxygenated organic compounds
Introduction We have recently reported the results from several studies of the gas-phase reactivity of hydroxyl radicals with a series of ketones,'-2 alcohol^,^-^ and ethers4-' in which we have tried to delineate structure-reactivity relationships for these oxygenated organics. In the present work, we have extended our measurements to difunctional oxygenates for which very few data are available in the literature.* These studies have practical importance as well since such compounds are released to the atmosphere during diesel fuel combustion9 as well as during their industrial production and use as solvents.IO In this report, we present the results of our flash photolysis resonance fluorescence investigation of the kinetics of the reactions of OH radicals with hydroxy alcohols (2-methoxyethanol, 2-ethoxyethanol, 2-butoxyethanol, 3-ethoxy- 1-propanol, and 3-methoxy- I-butanol), a hydroxy ketone (acetol), and a keto ether (methoxyacetone).
of reactant (attributed to diffusion out of the viewing zone and reaction with possible impurities in the argon diluent gas); and k , is the bimolecular rate constant for the reaction of OH with the reactant, R. Values of kist were determined for various reactant concentrations by nonlinear least-squares (eq I ) analysis of the OH fluorescence decay curves and ranged from 40 to 1000 s-'. k , was also measured experimentally (for the various conditions of total pressure, flow rate, and temperature) in the absence of reactant gas and ranged from 40 to 70 s-] at T I 298 K, increasing to 60-90 s-I at higher temperature. These ko values were constant to within 10 s-' for a given set of experimental conditions. Values for k , were then determined from linear least-squares analyses of plots of kist - ko versus the reactant concentration. Exponential decays of the resonance fluorescence signal were observed over at least three half-lives in all experiments, and the calculated decay rates were linearly dependent on reactant conExperimental Section centration. The 298 K rate data for 2-methoxyethanol, reaction I (see Figure l), demonstrate the data statistics characteristic of The apparatus and procedures used in the present work were all of the studies over the 25-50-Torr total pressure range. A linear typical of earlier flash photolysis resonance fluorescence experleast-squares analysis of these data (indicated by the solid line iments performed in our laboratory1q2and do not require further in the figure) yields a value of k,(298 K) = (1.25 f 0.07) X elaboration. All of the oxygenated reactants had manufacturers' cm3 molecule-I s-I where the quoted error is 2 standard deviations. stated purities of at least 98% and were further purified by reAlthough it is difficult to quantify systematic errors in such studies, peated freeze/pump/thaw cycles followed by fractional distillation. we estimate a maximum additional uncertainty of 5510% for these The argon diluent gas used in preparing the reaction mixtures had contributions in the present work. The measured rate constants a manufacturer's stated purity of 299.998% and was used directly were insensitive to variation of the flash energy (and hence of from the cylinder. (OH],) by a factor of 3 and of the reaction mixture residence time The initial hydroxyl radical concentration could be estimated by a factor of 4. These observations suggest a lack of complications from comparisons with our previous experiments and with similar due to secondary reactions involving either reaction products or For the apparatus and systems used by other photofragments. The complete set of 298 K rate constant data flash-lamp geometries and operating conditions used in this investigation, we conclude that 1Olo I[OH], I10" molecules ~ m - ~ . measured in the present work is given in Table 1. The results of the k , temperature-dependence study over the temperature Thus, pseudo-first-order kinetic conditions with respect to the OH decay were maintained throughout the reagent concentration range of (0.9-18.1) X IO" molecules cm-j. Results At low concentrations, the intensity of the OH fluorescence is directly proportional to the OH concentration, and the integrated first-order rate expression can be written as Fl = F,, exp{-kIst(t - r,)} = F,, exp{-(ko
+ k , [ R ] ) ( r- t o ) {
(I)
where Flo and F, are the OH radical fluorescence intensities at times to and t , respectively; kist is the total first-order decay rate; k , is the first-order rate constant for OH removal in the absence 'Present address: CNRS, CRCCHT, IC Ave. de la Recherche Scientifique, 45071 Orlzans Cedex 2, France. f
Present address: Ford Motor Company, Scientific Research Laboratory,
P.O. Box 2053, Dearborn, MI 48121. Formerly the National Bureau of Standards.
( I ) Wallington, T. J.; Kurylo, M. J. J . Phys. Chem. 1987, 91, 5050. (2) Dagaut, P.; Wallington, T. J.; Liu, R.; Kurylo, M. J. J . Phys. Chem. 1988, 92, 4375. ( 3 ) Wallington, T. J.; Kurylo, M. J. Int. J . Chem. Kinet. 1987, 19, 1015.
(4) Wallinnton. T. J.; Danaut. P.: Liu, R.: Kurvlo. M. J. Enuiron. Sci. Technol. 1988; 22, 842. ( 5 ) Wallington, T. J.; Dagaut, P.; Liu, R.: Kurylo, M. J. I n t . J . Chem. Kinet. 1988, 20, 541. ( 6 ) Wallington, T. J.; Liu, R.; Dagaut, P.; Kurylo, M. J. Int. J . Chem. Kinet. 1988, 20, 41. (7) Liu, R.; Wallington, T. J.; Dagaut, P.; Kurylo, M. J. Acta Phys.-Chim. Sin. 1989, 5, 210. (8) Atkinson, R. Int. J . Chem. Kinet. 1987, 19, 799. (9) Aaronson, A. E.; Matula, R. A. International Symposium on Combustion, 13th; The Combustion Institute: Pittsburgh, 1971; pp 471-481. ( I O ) Graedel, T. E. Chemical Compounds in the Atmosphere; Academic Press: New York, 1978. ( I I ) Wallington, T. J. Inr. J . Chem. Kinet. 1986, 18. 487. (12) Witte, F.; Urbanik, E.; Zetzsch, C. J . Phys. Chem. 1986, 90,3251.
This article not subject to U S . Copyright. Published 1989 by the American Chemical Society
Gas-Phase Reactions of OH Radicals with Hydroxy Ethers
The Journal of Physical Chemistry, Vol. 93, No. 23, 1989 7839 7
IVI r
I r
'VI h
0
Y
I
c VI r
Y v
I
0 r 7
Y
Figure 1 . Plot of klS' - ko vs 2-methoxyethanol concentration at 298 K: 0,25 Torr; A, 30 Torr; 0 , 40 Torr; 0, 50 Torr. The line represents a linear least-squares analysis. TABLE I: Measured and Predicted Rate Constants from the Present Work k(298 K), cm3 molecule-' s-I predicted GR reactant measured' (i)' (ii)d SARb Hydroxy Ethers 13.1 10.3 11.2 2-methoxyethanol 12.5 f 0.7 17.6 14.8 15.6 2-ethoxyethanol 18.7 f 2.0 19.6 2-butoxyethanol 23.1 f 0.9 26.6 23.8 19.3 17.0 3-ethoxy-I -propanol 22.0 & 1.3 23.5 e e 20.6 3-methoxy-I-butanol 23.6 f 1.6 Hydroxy Ketones acetol (hydroxyacetone) 3.0 f 0.3 methoxyacetone
Keto Ethers 6.8 f 0.6
0.5 1
2.0
P( 2-methoxyet hanol) , mTorr
3.1
2.4
2.3
6.8
5.7
4.9
'Quoted errors represent 2a from a least-squares analysis. Details may be found in ref 8. 'Predicted values in this column assume combined activation effects in the hydroxyethers and combined effects operating over the carbonyl group of the hydroxy ketone and keto ether as discussed in the text. dPredicted values in this column assume only the activation effect of the ether group in the hydroxy ethers and combined activation effects not operative over the carbonyl group in the hydroxy ketone and keto ether as discussed in the text. 'A G R prediction could not be made due to the lack of a value for the C H group reactivity in ethers.
2.5
3.0
3.5
4.0
4.5
1000/T (K-')
Figure 2. Arrhenius plot for the reaction of hydroxyl radicals with 2-methoxyethanol. The line represents a linear least-squares fit to the Arrhenius equation given in the text. TABLE 111: Group Reactivities (lo-'* cm3 molecule-' s-') in the Indicated Oxygenated Organics at 298 K Used To Predict the Rate Constants Listed in Table I ROH ROR' RC(0)R' RH CH3 0.8" 1.25' 0.1 I' 0.14' 0.4* 0.78 0.37k CH2 2.5' 4.50h 0.91' ==l.lm 1.2d 3.46k CH 4.2c i i Calculated for methanol through l - p e n t a n ~ l . ~bCalculated in a tert-butyl group.4 'Valid for the a through p positions with respect to the O H group.5 dValid for positions beyond 6 with respect to the OH group.5 e Calculated from the rate constant for 2-propanol.' 'Calculated from the rate constant for dimethyl ether? Assumed to be valid for the a through 6 positions by analogy with CH2.5.7 gCalculated in a tert-butyl group.' *Valid for the a through 6 posit i o n ~ . ~ ,'No ' value available. jValid for the a position. Calculated in ref 2 (revised from ref I ) . kValid for the p position. Calculated in ref 2 (revised from ref I ) . 'Reactivity in C2H6 calculated from ref 8. Average reactivity calculated from ref 8.
where the error limits are again 2u.
10-l2 cm3 molecule-' SKI)were measured by Hartmann et aI.l3 at room temperature for 2-ethoxyethanol and 2-butoxyethanol, respectively. There is no obvious explanation for the fact that these values are ~ 4 0 %lower than reported here. It should be noted (as will be discussed below) that predictions derived from Atkinson's structure activity relationship (SAR)* or from group reactivity values derived in our previous work are in better agreement with the present results than with those of ref 13. The primary aim of the present study was to evaluate the utility of the room-temperature hydrocarbon group reactivity coefficients determined in our earlier alcohol, ether, and ketone studies in predicting the rate constants for the reactions of OH radicals with difunctional oxygenates. The comparison of the measured and predicted values for these compounds should give some indication of how the reactivity enhancements due to one functional group are affected by the presence of a second. The hydrocarbon group reactivity (GR) values that we have derived for alcohols, ethers, and ketones are given in Table III. Also listed are the alkane values as determined by the SAR approach. Thus, the actual reactivity enhancement (or reduction) due to the presence of an oxygen-containing functional group may be considered to be the difference between the GR (oxygenate) and SAR (alkane) values. The predicted r a t e constant values given in T a b l e I were determined in several ways as we will now describe. For the hydroxy ethers, the first (larger) GR-predicted value was computed under the assumptions that the individual CH, or CH, reactivities associated with each of the two functional groups are additive but that the enhancement effect of the alcohol
Discussion Published OH rate constants exist for only two of the reactants listed in Table I. Rate constant values of 12 and 14 (in units of
( 1 3 ) Hartmann, D.; Greda, A,; Rhasa, D.; Zellner, R. Proceedings of the 4th European Symposium on Physico-Chemical Behavior of Atmospheric Pollutants; Riedel: Dordrecht, The Netherlands, 1987.
TABLE 11: Rate Constants for the Reaction of 2-Methoxyethanol with Hydroxyl Radicals Measured in the Present Work between 240 and 440 K temp, K ki(T)" Ab EIR, K-I 240 1.88 f 0.13' 298 1.25 f 0.07 3 50 1.10 + 0.06 400 1.04 f 0.08 440 1.01 f 0.06 -325 f 100 240-440 4.5 f 1.4 IO-" cm3 molecule-' s-l. cm3 molecule-' errors are 20 from a least-squares analysis.
s-l.
'All quoted
range 24C-440 K are reported in Table I1 and plotted in Arrhenius form in Figure 2. A linear least-squares analysis of these latter data corresponds to the line drawn in the figure and is represented by the expression k, = (4.5 f 1.4) X exp[(325 1 0 0 ) / T ] ,cm3 molecule-] SKI
*
7840 The Journal of Physical Chemistry, Vol. 93, No. 23, I989
functional group is not transmitted across the oxygen atom of the ether function. Thus, the reactivity of alkyl groups positioned between the two functions is computed as the sum of the group reactivity given in the GR ether column plus the reactivity enhancement calculated from the GR alcohol and SAR alkane columns. The second (lower) GR prediction was calculated assuming that all reactivity enhancement is due to the ether function. (Activation due to the hydroxyl group, being significantly lower, was neglected.) No G R prediction could be made for 3-methoxy-I-butanol due to the lack of a previous determination of a C H group reactivity in ethers. For the hydroxy ketone (acetol) and the keto ether (methoxyacetone), the GR predictions were also computed under two assumptions: that the activation effects due to the hydroxyl or ether functions were additive and (i) were and (ii) were not operative over the carbonyl group. It should be noted that, for the GR predictions, the small reactivity of the H atom in the alcohol group was not included. As it can be seen from the Table I comparison, the two sets of predicted rate constants from the GR approach for the hydroxy ethers are in quite reasonable agreement with the measured values. While both sets may not be statistically distinguishable from the measurements within the combined uncertainties of the experiments and the calculations, it appears that the calculation that ignores any reactivity enhancement contribution due to the alcohol group generally leads to an underestimation of the rate constant. Thus, in the hydroxy ethers, we believe that the alcohol hydroxyl function adds to the reactivity enhancement provided by the ether function, particularly for those hydrocarbon groups positioned between the two functions. Another way of assessing the level of any added reactivity enhancement is to compute the reactivity for C H 2 groups situated between the two functions in hydroxy ethers from the 298 K rate constants for 2-methoxyethanol, 2ethoxymethanol, 2-butoxyethanol, and 3-ethoxy- 1-propanol. This cm3 molecule-’ analysis yields a reactivity of (5.4f 0.8) X s-I per C H 2 (a value that is approximately 20% larger than we determined in simple ethers and is slightly less than the sum of the group reactivity of C H 2 in ethers plus the reactivity enhancement of CH2 in alcohols). In all cases, the group reactivity predictions for the hydroxy ethers (combining the functional group enhancements) are in better agreement with the experimental values than the SAR predictions; however, the differences are not very significant. For acetol (hydroxyacetone) and methoxyacetone,
Dagaut et al. the comparison between the various predicted and experimental values suggests that the enhancement of reactivity due to the hydroxyl (alcohol) and ether groups is operative over the carbonyl group and cannot be neglected. For both molecules, our group reactivity predictions with combined group effects give slightly better agreement with the measurements than the SAR. The suggestion that reactivity activation effects in these difunctional oxygenates is, to some extent, additive can be contrasted with our single observation for the diether, diethoxymethane,’ which indicated little or no effect of the second ether function on the group reactivity of the central CH2. It is instructive, at this point, to consider the temperature dependence of k , , the rate constant for the reaction of OH with 2-methoxyethanol. The present determination of E , / R = -(325 f 100) is statistically indistinguishable from the value of -(270 f 100) measured earlier for OH diethyl ether.6 As discussed at that time, this “negative” temperature dependence can be rationalized in terms of zero or near-zero activation energy associated with hydrogen atom abstraction from weak C-H bonds combined with the inverse temperature dependence of the preexponential factor in the Arrhenius expression. As can be seen in Figure 2, there is a suggestion of curvature in the Arrhenius plot for reaction 1. For this reason, use of the derived Arrhenius parameters to predict rate constants at temperatures well outside the temperature region of these measurements may not be valid. Such behavior has been addressed previously in the literature.8 From the atmospheric chemistry point of view, the rapidity of the reactions measured in the present study makes reaction with O H the most important sink for these difunctional oxygenated organics in the atmosphere, photolysis, reaction with NO,, and reaction with O3being negligibly slow by comparison. Assuming an average tropospheric OH concentration of IO6 radicals ~ m - ~ , we calculate atmospheric lifetimes ranging from 12 to 92 h for the organics considered in the present study.
+
Acknowledgment. The research described herein was conducted with the partial support of the National Aeronautics and Space Administration (Agreement W-15,8 16). Registry No. OH, 3352-57-6; 2-methoxyethanol, 109-86-4; 2-ethoxyethanol, 110-80-5; 2-butoxyethanol, 1 11-76-2; 3-ethoxy-l-propanol, 1 1 1-35-3; 3-methoxy-l-butanol, 2517-43-3; acetol, 116-09-6; methoxyacetone, 5878- 19-3.
