J. Phys. Chem. A 2010, 114, 11529–11537
11529
Shock Tube/Laser Absorption Measurements of the Reaction Rates of OH with Ethylene and Propene Subith S. Vasu,* Zekai Hong, David F. Davidson, Ronald K. Hanson, and David M. Golden Mechanical Engineering Department, Stanford UniVersity, Stanford, California 94305, United States ReceiVed: June 30, 2010; ReVised Manuscript ReceiVed: September 8, 2010
Reaction rates of hydroxyl (OH) radicals with ethylene (C2H4) and propene (C3H6) were studied behind reflected shock waves. OH + ethylene f products (rxn 1) rate measurements were conducted in the temperature range 973-1438 K, for pressures from 2 to 10 atm, and for initial concentrations of ethylene of 500, 751, and 1000 ppm. OH + propene f products (rxn 2) rate measurements spanned temperatures of 890-1366 K, pressures near 2.3 atm, and initial propene concentrations near 300 ppm. OH radicals were produced by shock-heating tert-butyl hydroperoxide, (CH3)3-CO-OH, and monitored by laser absorption near 306.7 nm. Rate constants for the reactions of OH with ethylene and propene were extracted by matching modeled and measured OH concentration time-histories in the reflected shock region. Current data are in excellent agreement with previous studies and extend the temperature range of OH + propene data. Transition state theory calculations using recent ab initio results give excellent agreement with our measurements and other data outside our temperature range. Fits (in units of cm3/mol/s) to the abstraction channels of OH + ethylene and OH + propene are k1 ) 2.23 × 104 (T)2.745 exp(-1115 K/T) for 600-2000 K and k2 ) 1.94 × 106 (T)2.229 exp(-540 K/T) for 700-1500 K, respectively. A rate constant determination for the reaction TBHP f products (rxn 3) was also obtained in the range 745-1014 K using OH data from behind both incident and reflected shock waves. These high-temperature measurements were fit with previous low-temperature data, and the following rate expression (0.6-2.6 atm), applicable over the temperature range 400-1050 K, was obtained: k3 (1/s) ) 8.13 × 10-12 (T)7.83 exp(-14598 K/T). C3H6 + OH f products
1. Introduction Small alkenes including ethylene (C2H4) and propene (C3H6) are important intermediates in the oxidation of hydrocarbon fuels1 and are formed in large quantities in practical engines. Ethylene and propene are also present as components of practical fuels and are emitted into the atmosphere through anthropogenic and natural sources. The oxidation kinetics of ethylene and propene are thus important to the hierarchical development of the kinetic mechanisms of real fuels.1-4 Furthermore, alkenes contribute to soot production (and other pollutant formation), therefore strategies for mitigating pollutant formation in advanced combustion systems depend in part on alkene oxidation chemistry. Lastly, we mention that ignition delay times and species time-histories for hydrocarbon fuels can be strongly sensitive to alkene concentration levels and oxidation rates.1 Hydroxyl (OH) is an important transient radical during combustion and in atmospheric chemistry, and it is widely accepted that OH radical reactions with ethylene and propene form the major oxidation route for these molecules under atmospheric and combustion conditions.1,5,6 Ethylene and propene react with OH to form various products, including water (the final product):6,7
C2H4 + OH f products
(1)
* To whom correspondence should be addressed. E-mail: subith@ stanford.edu,
[email protected]. Phone: +1 650-725-6771. Fax: +1 650723-1748.
(2)
Because ethylene is an important stable intermediate (and the simplest alkene) during combustion of higher hydrocarbons, extensive experimental and theoretical studies have been conducted on reaction 1. Table 1 lists all studies above 500 K along with some of the low-temperature studies of reaction 1.5,8-25 Despite the multitude of measurements, relatively large variations in the rate constants (magnitude and activation energy) for C2H4 + OH still exist in recent combustion kinetic mechanisms for practical fuels (see Figure 1). Shown are rates used in JetSurF 1.0,2 Westbrook et al.,3 Ranzi,4 the Galway natural gas III,26 and Chaos et al.27 mechanisms. Relatively very few studies5,28-32 have been conducted on the reaction of OH with propene (see Table 1). To the best of our knowledge, there has been no experimental study of reaction 2 above 1210 K. Hence, investigations at higher temperatures appear to be warranted for both reactions 1 and 2. In the present study, direct OH laser absorption measurements of reactions 1 and 2 are presented. As well, at lower temperatures a rate constant for the decomposition of tert-butyl hydroperoxide (TBHP) has been determined. The GRI 3.033 mechanism (with slight modifications) was used to model the OH time-histories. Measurements of the OH + alkene reaction rate constants are then compared to canonical transition state theory (TST) calculations. 2. Experimental Setup Experiments were performed behind reflected and incident shock waves in a high-purity, stainless steel, helium-driven shock tube with inner diameter of 14.13 cm. Test gas mixtures
10.1021/jp106049s 2010 American Chemical Society Published on Web 10/05/2010
11530
J. Phys. Chem. A, Vol. 114, No. 43, 2010
Vasu et al.
TABLE 1: OH + Alkenes (Ethylene, Propene) Experimental Studies ref
temperature range
Bott and Cohen5
1197 K
Smith8 Bradley et al.9
1220 K ∼1300 K
Baldwin et al.10
813 K
Bhargava and Westmoreland11 Westenberg and Fristrom12 Tully13 Tully14 Liu et al.15,16 Westbrook et al.17 Greiner18 Fulle et al.19 Srinivasan et al.20 Baldwin et al.21
1455-1740 K 1250-1400 K 291-591 K 650-901 K 343-1173 K 1003-1253 K 299-497 K 300-814 K 1463-1931 K 773 K
Diau and Lee22 Zellner and Lorenz23 Hoare and Patel24 Avramenko and Lorentso25
544-673 K 296-524 K 734-798 K 350-451 K
Bott and Cohen5
1204 K
Tully and Goldsmith28 Smith et al.29 Baldwin et al.30
293-896 K 960-1210 K 773 K
Atkinson and Pitts Jr.31 Yetter and Dryer32
297-425 K 1020 K
experimental method and comments OH + Ethylene Reflected shock tube (S.T.), 1 atm, OH resonance microwave absorption at 309 nm Laser pyrolysis/laser induced fluorescence (LIF) Incident S.T., OH UV lamp, rate determined relative to methane + OH Reaction vessel, pressure and gas sample analysis using method similar to gas chromatograph Laminar flames, molecular-beam mass spectrometry (MBMS) Flames, electron spin resonance (ESR) spectroscopy Flash photolysis, LIF Laser photolysis, LIF Pulse radiolysis/OH resonance microwave absorption, 1 atm Jet-stirred reactor (JSR) Flash photolysis/kinetic spectrograph Laser flash photolysis/saturated LIF Reflected S.T., OH UV Lamp Reaction vessel, pressure and gas sample analysis using method similar to gas chromatograph (GC) Laser photolysis/LIF Laser photolysis/resonance fluorescence Reaction vessel/GC Discharge flow system, OH measured by UV absorption. OH + Propene Reflected S.T., 1 atm, OH resonance microwave absorption at 309 nm Laser photolysis, LIF Laser pyrolysis, LIF Reaction vessel, pressure and gas sample analysis using method similar to gas chromatograph. Relative rate method using OH+tetramethylbutane (TMB) reaction Flash photolysis, LIF Flow reactor, 1 atm, GC
used research grade argon (99.999%), ethylene (99.999%), and propene (99.8%) supplied by Praxair Inc., and 70% TBHP in water supplied by Sigma-Aldrich. Mixtures were prepared manometrically in a stainless steel mixing chamber equipped with a magnetic stirrer assembly and mixed for about 6.5 h to ensure homogeneity and consistency. Details of the shock tube and mixing procedure can be found elsewhere,34-38 and only a brief description is given here. In all cases the mixing assembly and shock tube driven section were held at room temperature. Preshock GC analysis for such TBHP mixtures showed that less than 0.3 ppm of TBHP decomposes in the mixing tank to form acetone.35,39,40 Modeling the reaction system with this decom-
Figure 1. Comparison of rate constants for C2H4 + OH (k1) used in several current mechanisms. Variations of a factor of ∼5.5 are evident at high temperatures near 1000 K. Shown are rates used in (1) JetSurF 1.0,2 (2) Westbrook et al.,3 (3) Galway natural gas III,26 (4) Ranzi,4 and (5) Chaos et al.27 mechanisms.
position taken into account showed no discernible effect on the rate constant measurements. Also, kinetic simulations showed that water vapor in the initial reactant mixture had no significant effect on the rate constant measurements. The shock tube test section and mixing manifold were routinely evacuated below 1 µTorr using two turbo-molecular pumps (Varian V-250, and V-80, respectively) to ensure purity of the test mixture. The leak-plus-outgassing rates for the shock tube were less than 5 µTorr/min. Incident shock velocity was measured using five axially spaced PCB Piezotronics 113A26 piezoelectric transducers (PZT’s) and a six-channel amplifier (PCB 483B08) connected to four Philips PM6666 interval timers with a resolution of 0.1 µs, and linearly extrapolated to the endwall. Average incident shock wave attenuation rates were between 0.6 and 2% per meter. Ideal shock relations and the thermodynamic database from ref 41 were used to calculate reflected shock temperature and pressure (T5 and P5). Uncertainty in the initial reflected-shock temperature and pressure (i.e., test conditions) is typically less than 0.7 and 1%, respectively, arising out of 0.3% uncertainty in the endwall shock velocity measurements (see34,36,42). OH radicals were monitored using a narrow-line width ring dye laser tuned to the center of the R1(5) absorption line in the (0,0) band of the A2Σ+ r X2Π transition at 306.7 nm. This particular absorption feature was chosen since it had the highest peak absorption and is well isolated from adjacent lines at the current experimental conditions. Absorption spectroscopy of the OH radical is well-established at combustion temperatures and near-atmospheric pressures, and the current laser absorption method is fully described in refs 1, 34-36, and 39. The minimum detectable absorption was less than 0.1%, leading to
Reaction Rates of OH with Ethylene and Propene
J. Phys. Chem. A, Vol. 114, No. 43, 2010 11531
ppm-level detectivities for OH at microsecond time scales. OH mole fractions (XOH) were determined from the measured absorption using the Beer-Lambert relation: I/I0 ) exp{-kνP5XOHL}. P5 (atm) is the initial reflected shock pressure, and L is the path length (14.13 cm). The well-characterized absorption coefficient, kν, was calculated incorporating the collision-broadening and collision-shift parameters previously measured.34,36 Several uncertainties contribute to the determination of XOH with uncertainties in temperature (T5) having the largest effect. All data were recorded at 2 MHz using a highresolution data acquisition system. The overall estimated uncertainty in measured XOH is ∼3%. Measurements were also performed with the laser tuned off the OH absorption line and with the laser turned off to verify that there was no significant interfering absorption or emission. 3. Kinetics Measurements A total of 28 shock wave experiments were carried out to determine the rate constants of OH with ethylene and propene. Measurements were conducted over temperatures from 973 to 1438 K and pressures from 2.0 to 10.2 atm for three initial concentrations of ethylene (500, 751, and 1000 ppm). OH + propene rate constant measurements spanned temperatures of 890-1366 K and pressures of 2.0-2.6 atm at an initial propene concentration near 300 ppm. Nominal mixtures with 96-150 ppm TBHP and water (based on partial pressures) diluted in argon were prepared. To model the OH mole fraction time-history, the GRI 3.033 natural gas oxidation mechanism was chosen as the base mechanism for ethylene. At T > 1000 K, TBHP decomposes almost instantaneously (half-life ∼ 0.1 µs) to form an OH radical and a tert-butoxy radical, (CH3)3CO, which immediately (halflife < 0.1 µs) falls apart to form acetone and a methyl radical (acetone and methyl radicals can also react with OH):
(CH3)3sCOsOH f OH + (CH3)3CO
(3)
(CH3)3CO f CH3 + CH3COCH3
(4)
CH3COCH3 + OH f CH3COCH2 + H2O
(5)
Rate constants measured by Vasudevan et al.35,39 were used for k3 and k5, respectively, whereas the evaluation of Benson and O’Neal43 was used for k4. The acetone decomposition reaction, CH3COCH3 f CH3CO + CH3 was also added with rates taken from Vasudevan et al.39 For all species, the GRI 3.033 thermodynamic data were used except for the following species: CH3COCH3, TBHP, and (CH3)3CO data were taken from ref 41, OH from ref 44, and CH3COCH2 from ref 3 (which was calculated using the THERM program45). To the above modified GRI 3.033 mechanism, a propene submechanism containing 53 reactions (for propene and products formed from propene) and 14 species (and associated thermo data) from JetSurF 1.02 was added to model propene experiments. All simulations were conducted assuming homogeneous adiabatic conditions behind reflected shock waves, with the common constant-internal-energy, constant-volume constraint (constant U, V) using CHEMKIN 4.1.1.46 Using only a OH diagnostic, it is not possible to distinguish the various product channels of reactions 1 and 2. However, because of the pseudofirst-order behavior of the removal of OH, the total rate is independent of the branching ratios, that
Figure 2. Example OH + ethylene rate constant measurement at 1201 K, 2.00 atm, 500 ppm ethylene, Ar bath gas, and ∼23 ppm TBHP (fit). Model predictions using the best-fit predictions for the rate and 50% variation from the inferred rate are also shown.
TABLE 2: OH + Ethylene Rate Constant Data, k1a T (K) 973 1014 1078 1321 1438 1301 1232
P (atm)
k1 best-fit to simulation
k1 pseudofirst-order determination
1000 ppm Ethylene, 20-40 ppm TBHP in Ar 2.52 1.16 1.03 2.46 1.4 1.51 2.45 1.7 1.82 2.19 3.5 3.30 2.09 4.69 2.19 3.1 3.39 2.31 2.9 2.98
1007 1085 1201
500 ppm Ethylene, ∼21 ppm TBHP in Ar 2.56 1.55 2.28 1.8 1.99 2.6
1055 1102
751 ppm Ethylene, ∼24 ppm TBHP in Ar 10.2 1.5 7.00 2.0
a
All rate constants are in units, 1012 cm3 mol-1 s-1.
is, the choice of products has no discernible effect on our overall rate determination. The product of C2H4 + OH for modeling purpose was taken as C•HCH2 + H2O and the products of C3H6 + OH were taken to be water and the three abstraction products: CH3C•CH2, CH3CHC•H, and C•H2CHCH2 (allyl). 3.1. OH + Ethylene, k1. A sample OH absorption trace at the middle of the ethylene temperature range is shown in Figure 2. For the conditions of the experiment shown (1201 K, 2.00 atm, 500 ppm ethylene/argon), the measured peak OH yield is ∼23 ppm. Because wall adsorption and condensation of TBHP is possible, the initial TBHP mole fraction was inferred from the peak OH concentration measured after the decomposition of TBHP behind the reflected shock wave. For the experiment shown in Figure 2, an overall rate constant of 2.6 × 1012 cm3 mol-1 s-1 was obtained from the best-fit of the simulation with the experimental data. Model predictions for a variation of 50% of the inferred rate constant are also shown. Other experiments were performed with varying amounts of TBHP and ethylene. Table 2 summarizes the current measurements (best-fit to simulation). The current experiments are insensitive to the type of mechanism used to model OH data; k1 determined using the Galway mechanism26 for the conditions of Figure 2 is also 2.6 × 1012 cm3 mol-1 s-1. An OH radical sensitivity analysis (for the conditions in Figure 2) is provided in Figure 3. Sensitivity is defined as S ) (dXOH/dki)/(ki/XOH), where XOH is the local OH mole fraction
11532
J. Phys. Chem. A, Vol. 114, No. 43, 2010
Vasu et al.
Figure 3. OH sensitivity plot for rate measurement of OH + ethylene at conditions of Figure 2 (1201 K, 1.992 atm).
and ki is the rate constant of reaction i. This analysis shows that the reaction between OH and ethylene is the dominant reaction over the entire time frame of the experiment with only slight interference from the following reactions:
CH3COCH3 + OH f CH3COCH2 + H2O
(5)
CH3 + OH f 1CH2 + H2O
(6)
CH3 + OH(+M) f CH3OH(+M)
(7)
CH3 + CH3(+M) f C2H6(+M)
(8)
Because the initial concentration ratio of the reactants ([ethylene]0/[TBHP]0) ratio is larger than 20, the chemistry is pseudofirst-order and hence rate constants can also be determined without using a detailed mechanism fit (see Cook et al.47). By measuring the initial slope of the ln [OH] versus time plot in some of the experiments (where the plot is linear), pseudofirstorder rate constants can be obtained for the overall OH + (ethylene, CH3, and other OH-removing species) reaction. By conducting experiments in similar TBHP and argon (ethylene replaced by argon) concentrations, a pseudofirst-order rate constant for OH + (CH3, and other OH-removing species) was obtained. The difference of these two pseudofirst-order OH
removal rates gives k1 (OH + ethylene) and is tabulated in Table 2. The pseudofirst-order determination was only possible in experiments where both these graphs are linear and for identical TBHP concentrations. k1 determined using the pseudofirst-order method is typically within 10% of the more accurate value of k1 determined using the fit to detailed modeling, as expected from the sensitivity analysis. A detailed error analysis was carried out to set uncertainty limits on the measured rate constants. The individual contributions of each error source were determined by perturbing each error source to the estimated positive and negative bounds of its 2σ uncertainty and refitting the experimental traces by adjusting the rate constant of interest.48-50 The resulting uncertainties in the rate constant for each individual error source were combined using the root-sum-squares (RSS) method. The main uncertainty categories and their effect on the target reaction rate of OH + ethylene, k1, for the experiment at 1201K, 2.00 atm are shown in Table 3. An overall uncertainty estimate of (22.8% at 1201 K and 1.992 atm was found. 3.2. OH + Propene, k2. The OH + propene data were reduced in the same manner as the ethylene + OH data. An example OH absorption trace at the middle of the propene temperature range is shown in Figure 4. For the conditions of the experiment shown (1136 K and 2.02 atm) for 300 ppm of initial propene, the measured peak OH yield is ∼20 ppm. For this experiment, an overall rate constant of 7.84 × 1012 cm3 mol-1 s-1 was obtained using the best-fit approach. Model
TABLE 3: Uncertainty Analysis for OH + Alkenes Reactionsa
error source temperature fitting (signal/noise) OH absorption coefficient (kν) mixture concentration wavemeter reading
CH3COCH3 + OH f CH3COCH2 + H2O CH3 + OH f 1CH2 + H2O CH3 + OH(+M) f CH3OH(+M) CH3 + CH3(+M) f C2H6(+M) total 2σ RSS uncertainty in (k1, k2):
uncertainty in (k1, k2) from positive perturbation
2σ uncertainty in error source Experimental (0.7%,48 (5% (3%,48 (0.43 ppm,48 (0.0.01 cm-1,39
Uncertainties (
