J . Chern. Sac., Faraday Trans. 1, 1985, 81, 259-263 Rate Constants for the Reactions of Hydroxyl Radicals with Propane and Ethane BY DONALD L. BAULCH,* IAN M. CAMPBELL AND SANDRA M. SAUNDERS Department of Physical Chemistry, The University, Leeds LS2 9JT Received 5th June, 1984 The kinetics of the reaction OH + C,H, --+ H,O + C3H7 have been studied in a discharge-flow system under first-order conditions. The OH radicals were generated by the reaction of H atoms with NO, and the concentration of OH, monitored by resonance fluorescence, was followed as a function of reaction time in a large excess of the alkane. A value of k , = (7.2+ 1.1) x 10, dm3 mol-1 s-l at 295 K was obtained. As a check on the technique the rate constant for the reaction OH + C,H, -+ H,O + C,H, (2) was determined in a similar fashion.The value obtained, k , = (1.61 kO.24) x 10, dm3 mol-l s-l, is in excellent agreement with other literature values. The reaction of OH radicals with propane OH + C,H, -+ H,O + C,H, (1) is of importance in combustion processes and the chemistry of the atmosphere. A recent evaluation1 of the rate constant for reaction (1) for use in atmospheric modelling demonstrates a surprising lack of concordance in the available data at 300 K, which scatter over a range of nearly a factor of 3. Equally surprising is the fact that the reaction has not been studied under first-order conditions in a discharge-flow system, a technique well suited to measuring this rate constant. The only previous discharge-flow study2 used comparable concentrations of propane and hydroxyl radicals and required a mass-spectrometric measurement of the reaction stoichiometry to derive a value of k,.Such measurements are difficult to perform accurately and introduce a corresponding uncertainty into the rate-constant value so obtained. In the present work OH radicals generated in a discharge-flow system by means H+NO,-+OH+NO of the rapid reaction were reacted with an excess of propane sufficiently large to ensure that there were no significant routes of OH removal other than reaction (1). The rate constant, k,, was derived from the change down the flow tube of [OH], monitored by resonance fluorescence. To test the satisfactory operation of the apparatus and technique the well established' rate constant for the reaction OH -t- C,H, -+ H,O + C,M, (2) was also measured.259260 REACTIONS OF OH WITH C,H8 AND C,H, EXPERIMENTAL The apparatus used was a discharge-flow system of conventional design, having a 25 mm i.d. Pyrex flow tube 1 m in length with inlet ports and a sliding double injector for introduction of reagents. At a fixed point at the downstream end of the tube a brass fluorescence cell for monitoring the [OH] was mounted. H atoms, produced by a microwave discharge in a stream of He containing a small quantity ( 1 % ) of molecular H,, were introduced into the flow tube through a sidearm and titrated with NO, introduced through the outer inlet of the dual injector. In this way known concentrations of OH were produced by the fast reaction H+NO, -+ OH +NO, (3) with k , = 8 x 10'" dm3 mo1-l s-' at 298 K.Other reactants (C,H, and C,H,) were admitted to the flow tube through the central inlet of the injector. The distance between the point of admission of NO, and the alkane was sufficient to ensure effective completion of reaction (3) under the conditions used. The [OH] was monitored by means of a conventional OH resonance fluorescence system. The OH emission band centred on 309 nm ( A 2C t X "n) was generated by passing a microwave discharge through a stream of argon saturated with water vapour. The fluorescence emission was detected after passage through a collimator and narrow-band filter (Oriel 5703-10 nm band width) by a photomultiplier (EM1 9789Q) placed at angles to both the OH lamp and flow-tube axes. The photomultiplier signal was fed to a photon counter (Brookdeal-Ortec PCS 5C1).Light traps to minimize scattered radiation were placed opposite the resonance lamp and photomultiplier. The initial OH concentration was obtained from the value of [NO,] at the end-point of the titration of H with NO,. The variation of the fluorescence intensity was studied as a function of OH concentration and, as has been observed by others, it was found to be linear at [OH] < cu. 1 OPH mol dmP3. All measurements were carried out with OH concentrations well below this value. To reduce wall removal of the OH radicals the flow tube and injector were coated with Halocarbon wax (series 15-00> and the interior of the fluorescence cell was coated with Teflon. The first-order rate constant for wall removal of the OH was found to be cu.23 s-l and constant throughout the series of measurements. Preliminary experiments demonstrated the necessity of removing trace impurities from the helium carrier gas (B.O.C., CP grade), which was therefore passed through a silica tube at 900 K packed with copper turnings, followed by a liquid-nitrogen trap. Argon (B.O.C., A grade), H, and C,H, (Matheson, research grade) and C,H, (Matheson, CP grade) were used without further purification. The NO, (Matheson, 99.57,) was further purified by mixing with 0,. leaving to stand overnight, freezing and pumping away the 0,. subjecting the NO, to a series of freeze-thaw cycles and finally taking the middle fraction. Concentrations of reagents were calculated from flow rates measured by means of calibrated capillary flow meters and the mean pressure in the flow tube.Operating conditions were: total pressure cu. 0.3 kPa; temperature 295 2 K , linear flow velocities cu. 15 m s-* ; [OH],, z mol dmP3. RESULTS Since the resonance-fluorescence detection system is newly constructed its perform- ance, and the other procedures, were checked by measuring the rate constant for reaction (2). Its value is well established (table 1) and a recent evaluation1 has recommended k , = ( I .63 k0.24) x IOH dm3 mob ssl at 298 K. In a very large excess of C,H, the decay of [OH] down the flow tube will be given In [OH]/[OH], = - (k' + k , ) t where k , is the first-order rate constant for reaction of OH with the walls, k' = k,[C,H,] and t, the reaction time. is related to the reaction distance downD.L. BAULCH, I. M. CAMPBELL AND S. M. SAUNDERS 26 1 I I 1 I 0 0.01 0.02 0.03 0.04 t l s [C,H,] = 7.6 x lo-@ rnol dmP3 and 0, [C,H,] = 2.8 x lo-’ rnol dmP3. Fig. 1. Plot of in[OH] as a function of reaction time for [OH], = mol dm-3: 0, the flow tube, d, and the linear flow rate, f,>, by t = d/fIJ. Good first-order plots were obtained for ln[OH]/[OH], against d, as illustrated in fig. 1 for two different values of [C,H,]. Values of (k’ + k,) obtained from the least-mean-squares value of the slopes of such plots were plotted against [C,H,] (fig. 2). Within the limits of scatter the plot is linear with an intercept k , = 2243 s-l and a slope k, = (1.61 kO.10) x lo8 dm3 mol-l s-l, where the error limits are one standard deviation.This value of k, is in excellent agreement with previous measurements and the value of k , agrees closely with the value of 2 3 k 3 s-l measured in the absence of C,H,. mol dm-3 and although pseudo-first-order kinetic behaviour still appeared to apply, the value of k was based only on measurements at [C,H,] > 1.5 x lop7 rnol dm-3 (i.e. [C,H,]/[OH], > 100) under which conditions only reaction (2) and wall removal of OH should be significant. Reaction (1) was studied in a similar fashion. Values of (k’ + k,) as a function of [C3H,] are shown in fig. 2. A least-squares fit yields k , = (7.2 f 1.1) x lo8 dm3 mol-1 s-l. A detailed analysis of sources of error arising in flow-tube studies33 suggest that an overall error of & 150/;, as quoted here would be more realistic than the 804 standard deviation given by the least-squares fit.Although measurements were made for [C,H,] < 1.5 x262 100 8 0 - I co . h e3 60 5 + 40 20 REACTIONS OF OH WITH C,H8 AND C,H, I I I I 0 1.0 2.0 3.0 L.0 [alkane]/lO-' mol dm-3 Fig. 2. Plot of (k'+k,) as a function of alkane concentration. ., C,H6 and D, C3H,. The lines are least-mean-squares fit to the points. Table 1. Experimentally determined values of k , and k, at temperatures in the region of 300 K ~ ~ ~ _ _ _ ~ ~ _ _ _ alkane k/lOs dm3 mol-' s-' techniquea TIK ref. 'ZH6 I .70 1.59 k0. 10 1.74 k 0.36 1.56k0.24 1.39 k 0.24 1.55 k0.13 1.61 kO.10 7.24 & 0.42 5.00k0.12 6.32 & 0.24 9.09 0.13 7.35 & 0.30 12.16 k 0.60 13.24f3.6 7.2 0.60 f. p .-p . p. f. p .-r .a. d. f.--1.m .r. d.f.-r.f. d.f.-ref. f.p.-r.f. d. f.-r . f. d.f.-e.s.r. f.p.-r.f. smog chamber smog chamber f. p.-r . a. photolysis-product analysis p.f. - r.f. f.p.-k.s. 297 295 k 2 296 298 RTb 298 295 &- 2 297 298 298 300 299 295 k 2 298 295 f 2 5 6 7 8 9 10 this work 5 2 10 11 12 6 13 this work a f.p., flash photolysis; p.p., plate photometry; r.a., resonance absorption; d.f., discharge Room flow; l.m.r., laser magnetic resonance; r.f., resonance fluorescence; k.s., kinetic studies. temperature. DISCUSSION The good agreement between our value for k , and other literature data suggest that the apparatus and procedures are satisfactory. The available values of k , at temperatures close to 300 K are given in table 1. The results from the present study agree closely with the most recent, and probably theD .L. BAULCH, I. M. CAMPBELL AND S. M. SAUNDERS 263 most reliable, flash-photolysis work of Tully et a1.lo and also with the earlier flash- photolysis study of Greiner5 and the relative rate measurements of Darnall et a1.l1 The earlier discharge-flow work gives a low value, probably because of the difficulty of determining the stoichiometry with any accuracy. The other results are all very high, although only one of them, that of Overend et aZ.,6 is clearly well outside the quoted error limits; there is no obvious reason for such a discrepancy since the same group obtained an acceptable value for k , using the same technique. On the basis of the present work and that of G r e i n e ~ , ~ Tully et aLlo and Darnall et al." we suggest a value of k, = (7.0 & 0.6) x lo8 dm3 mol-1 s-l at 298 K.We thank the S.E.R.C. for the award of a studentship to S.M.S. during the tenure of which this work was carried out. * CODATA Task Group on Chemical Kinetics, D. L. Baulch, R. A. Cox, R. F. Hampson, J. A. Kerr, J. Troe and R. T. Watson, Evaluated Kinetic and Photochemical Data for Atmospheric Modelling, Supplement II, J. Phys. Chem. ReJ Data, in press. J. N. Bradley, W. Hack, K. Hoyermann and H. Gg. Wagner, J . Chem. SOC., Faraday Trans. 1 , 1973, 69, 1889. C. J. Howard, J. Phys. Chem., 1979, 83, 3. F. Kaufman, Prog. React. Kinet., 1961, 1, 3. N. R. Greiner, J. Chem. Phys., 1970, 53, 1070. R. P. Overend, G. Paraskevopoulos and R. J. Cvetanovic, Can. J. Chem., 1975, 53, 3374. C. J. Howard and K. M. Evenson, J. Chem. Phys., 1976,64,4303. M. T. Leu, J. Chem. Phys., 1979, 70, 1662. J. H. Lee and I. N. Tang, J. Chem. Phys., 1982,77,4459. lo F. P. Tully, A. R. Ravishankara and K. Carr, Int. J . Chem. Kinet., 1983, 15, I 1 1 1 . I1 K. R. Darnall, R. Atkinson and J. N. Pitts, J . Phys. Chem., 1978, 82, 1581. l2 R. Atkinson, S. M. Aschmann, W. P. L. Carter, A. M. Winer and J. N. Pitts, Int. J . Chem. Kinet., l 3 R. A. Gorse and D. H. Volman, J. Photochem., 1974, 3, 1 1 5. 1982, 14, 781. (PAPER 4/924)