Article pubs.acs.org/JPCA
Shock Tube Investigation of CH3 + CH3OCH3 Robert S. Tranter,* Patrick T. Lynch, and Christopher J. Annesley Chemical Sciences and Engineering Division, Argonne National Laboratory, 9700 South Cass Avenue, Argonne, Illinois 60439, United States S Supporting Information *
ABSTRACT: The title reaction has been investigated in a diaphragmless shock tube by laser schlieren densitometry over the temperature range 1163−1629 K and pressures of 60, 120, and 240 Torr. Methyl radicals were produced by dissociation of 2,3butanedione in the presence of an excess of dimethyl ether. Rate coefficients for CH3 + CH3OCH3 were obtained from simulations of the experimental data yielding the following expression which is valid over the range 1100−1700 K: k = (10.19 ± 3.0)T3.78 exp(−4878/T) cm3 mol−1s−1. The experimental results are in good agreement with estimates by Curran and co-workers [Fischer, S. L.; Dryer, F. L.; Curran, H. J. Int. J. Chem. Kinet. 2000, 32 (12), 713−740. Curran, H. J.; Fischer, S. L.; Dryer, F. L. Int. J. Chem. Kinet. 2000, 32 (12), 741−759] but about a factor of 2.6 lower than those of Zhao et al. [Zhao, Z.; Chaos, M.; Kazakov, A.; Dryer, F. L. Int. J. Chem. Kinet. 2008, 40 (1), 1−18].
1. INTRODUCTION Dimethyl ether, DME, is of continued interest in combustion studies because of its potential as a clean replacement for diesel. In recent years there have been a number of experimental and modeling investigations1−15 of the oxidation and pyrolysis of DME. These include shock tube12 and rapid compression machine, RCM,16 studies of ignition delay times, which are important for practical applications as well as tests of proposed mechanisms. Dimethyl ether dissociates by reactions 1a and 1b,13,14 although there is some uncertainty as to the importance of reaction 1b, which is proposed to occur by a roaming mechaniam.13 CH3OCH3 → CH3 + CH3O ΔHr,298K = 84.1 kcal/mol
CH3 + CH3OCH3 → CH4 + CH3OCH 2• ΔHr,298K = −8.7 kcal/mol
From modeling their experimental results they determined a rate coefficient that was about a factor of 3.5 higher at 1000 K than an earlier expression by Curran and co-workers.5,6 However, from simulation of shock tube/laser schlieren densitometry (ST/LS) experiments on dimethyl ether pyrolysis Tranter et al.14 estimated k2 values that were about one-half that of Zhao et al. The literature on reaction 2 is not extensive, particularly at high temperatures. The majority of the previous experiments were conducted at low temperatures (373 < T < 573 K) with photolytic sources of CH3 radicals.3,18−21 Pacey1 studied pyrolysis of dimethyl ether in a flow tube by gas chromatographic analysis of reaction products at higher temperatures, 732−986 K, and pressures in the range 25−395 Torr. Pacey assumed that reaction 2 was the only source of methane detected in the reaction products and obtained k2 from simulations. Later Held et al.22 used a similar apparatus and methodology at 1005 K (10−80 Torr) but augmented with ultraviolet absorption spectroscopic detection of CH3 radicals. Hidaka et al.8 studied DME pyrolysis in shock tubes (900 < T < 1900 K, 0.83−2.9 bar) and by fitting a relatively complex mechanism to their data obtained an Arrhenius expression for k2. The rate coefficients obtained in the low-temperature experiments differ by up to a factor of 5, whereas there is apparently better agreement between the high-temperature measurements of Pacey1 and those Held et al.22 as well as with those of Hidaka et al.8 in the region of overlap. Curran and coworkers5,6 estimated a rate expression for k2 from modeling high-pressure flow reactor experiments on DME pyrolysis.
(1a)
CH3OCH3 → CH4 + H 2CO ΔHr,298K = 0.1 kcal/mol (1b) 13
Sivaramakrishnan et al. determined experimentally that the branching ratio k1b/(k1a + k1b) = 19 ± 7%, but their theoretical study predicted 1000 K increasing to a factor of 2.8 higher than Hidaka et al. at 1500 K. Finally, as mentioned above, Zhao et al.7 found it necessary to increase k2 by about a factor of 3.5 compared to the earlier expression of Curran and co-workers5,6 to account for autoignition in DME. Given the wide range of rate coefficients in the literature for the reaction between methyl radicals and dimethyl ether and its importance in autoignition of DME, additional experimental results are desirable, especially at high temperatures. Consequently, the title reaction has been investigated in a diaphragmless shock tube, DFST, with LS diagnostics by generating methyl radicals from thermal decomposition of a precursor in the presence of a large excess of DME. To the best of our knowledge, this is the first time a radical/molecule reaction has been studied in an LS experiment by decomposing a precursor in the presence of a second reagent. Interpretation of the experimental results requires detailed mechanisms for pyrolysis of the reagents at temperatures and pressures similar to those in the proposed work. Pyrolysis of dimethyl ether was previously investigated in the same apparatus14 used in the current work, and for T < 1400 K dissociation of DME is negligible in the short observation period, 10 μs, of LS experiments. Methyl radicals were generated by pyrolysis of 2,3-butanedione (diacetyl), also previously studied by LS in this apparatus.23 Diacetyl dissociates to acetyl radicals, reaction 3, and cleanly yields CH3 and CO by reaction 4, which is facile and effectively instantaneous. Over the range 1200−1400 K dissociation of diacetyl is quite slow, and to minimize reaction between CH3 and diacetyl the experiments were carried out in a large excess of DME.
3. EXPERIMENTAL RESULTS A total of 51 experiments were performed over the range 1163 < T2 < 1629 K and three pressures, P2 = 60 ± 4, 120 ± 4, and 240 ± 8 Torr. Example raw laser schlieren profiles are shown in Figure 1a−d. The large positive spike and preceding downward
CH3COCOCH3 → 2CH3CO• ΔHr,298K = 72.9 kcal/mol (3) •
CH3CO → CH3 + CO ΔHr,298K = 11.1 kcal/mol
(4)
2. EXPERIMENTAL METHODS The experimental apparatus consists of a diaphragmless shock tube (DFST) equipped for laser schlieren (LS) densitometry behind incident shock waves. The equipment has been fully described elsewhere,24 and only a brief description will be given here. The DFST consists of a circular cross-section, stainless-steeldriven section of 6.35 cm i.d., and a diaphragmless driver section. Optical ports for passage of the laser beam for LS experiments are provided sufficiently far downstream of the driver section to allow the incident shock wave to be fully developed. Five piezoelectric pressure transducers, spaced 120 mm between their centers, are centered on the LS windows and used to obtain the incident shock wave velocity. The shock wave properties are calculated in the normal fashion from the incident shock wave velocity and the initial conditions in the driven section. The estimated error in the shock velocity is about 0.2%, leading to an error in the post shock temperature of around 10−15 at 1500 K in the incident wave. The diaphragmless driver section24 consists of a stainless steel tube approximately 58 cm long by 22 cm i.d., which contains a novel bellows actuated valve that separates the driver
Figure 1. Laser schlieren profiles for dissociation of diacetyl/DME mixtures dilute in krypton.
spike evident in all examples are due to passage of the shock front through the laser beam and have been discussed previously.27 On the immediate right of the large spike, a small dip is observed in some experiments, e.g., Figure 1a and 1b. This is a residual due to diffraction and often seen and always ignored when interpreting LS results. Following the signal due to the shock front, there is a curved portion that is due to chemical reaction. As is normal in ST/LS experiments, the positive spike due to the shock front masks the point, t0, at which reaction starts. Nevertheless, a well-established method locates t0 to within 0.1−0.2 μs.27For analysis the raw laser schlieren profiles are first converted to density gradients, and those derived from the raw signals in Figure 1 are shown in Figure 2. All examples show initial positive gradients that later 7288
dx.doi.org/10.1021/jp302761b | J. Phys. Chem. A 2012, 116, 7287−7292
The Journal of Physical Chemistry A
Article
Table 1. Simplified Mechanism for Simulating Pyrolytic Experiments with Mixtures Containing Diacetyl and Dimethyl Ether Dilute in Kryptona log(A) 1a 1b 2 3
4 5 6 7
Figure 2. Semilog density gradient plots obtained for dissociation of diacetyl/DME mixtures dilute in krypton. Plots are derived from the data in Figure 1; experimental data are represented by symbols, while lines are the results of simulations. Absolute values are plotted. Open symbols represent positive values and closed symbols negative ones. Solid lines represent the final simulations with the mechanism in Table 1: (a) dashed line 0.7k3; dot−dashed line 1.3k3; (b) dashed line k2 from Zhao et al.;7 (c) dashed line 0.7k3; dot−dashed line 1.3k3.
8
9 10 11 a
become negative, indicating a change from net endothermic reaction to net exothermic reaction. For an ideal shock the contribution to the density gradient, dρ/dx, from a given reaction i, is proportional to its rate of reaction, ri, and its heat of reaction, ΔHri, with a small correction due to any change in mole number, ΔNi,26 eq I. Typically, the term CPTΔNi evaluates to 10−15 kcal/mol, and therefore reactions, that are mildly endothermic produce little or no density gradient as do those reactions where the rate of reaction is small. dρ ∝ dx
∑ ri(ΔHr − CpT ΔNi) i
i
DME + M → CH3 + CH3O + M DME + M = CH4 + CH2O + M CH3 + DME → CH4 + C2H5O diacetyl + M → CH3CO + CH3CO + M CH3CO → CH3 + CO CH3O + M → H + CH2O + M H + DME → H2 + C2H5O H + diacetyl → H2 + CH3CO + CH2CO CH3 + diacetyl → CH4 + CH3CO + CH2CO C2H6 → CH3 + CH3 CH3 + CH3 → C2H5 + H H + C2H4 → C2H5
n
Ea/R
ΔHr,298K
ref
65.9
−13.6
45.349
84.1
14
63.9
−13.6
45.349
0.1
14
1.008
3.78
4.847
−8.7
37.268
−7.425
29.317
72.9
present work 23
16.5
−2.09
7.649
11.1
23
13.732
0
7.800
21.1
14
0.595
4.14
0.896
−7.9
14
13.629
0
3.200
12.2
23
11.8
0
4.539
11.3
23
61.084
−13.5
55.193
90.2
23,30
13.732
0
8.079
10.6
23,30
40.85
−8.79
5.817
−36.0
23,30
k = ATnexp(−Ea/RT). Units: kcal, cm3, mole, s, K.
The results of simulations with the final model are shown in Figure 2 by the bold solid line. The overall fit to the experimental data is generally good with small deviation at long times, particularly for the higher temperature experiments. In the simulations all rate coefficients were fixed apart from k3 and k2, dissociation of diacetyl and CH3 + DME, respectively. k3 was varied within the experimental uncertainties23 to obtain the best fit to the early portion of the density gradient. Typically corrections of
