An Experimental and Kinetic Study of

Acetone Oxidation in a Flow Reactor

Marcos Chaos, Zhenwei Zhao, Andrei Kazakov,
Frederick L. Dryer

31st International
Symposium on
Combustion
August 6-11, 2006

Mechanical & Aerospace

Engineering Department

mchaos@princeton.edu
http://www.princeton.edu/~combust

University of Heidelberg,
Germany

Experiments and Kinetic Modeling

Introduction
Acetone (CH3COCH3) is an important species formed during

New acetone oxidation experiments were performed in a flow

reactor at 12.5 atm pressure to measure time histories of stable
species concentrations

Princeton Variable Pressure Flow

Reactor (VPFR)

the free radical oxidation of hydrocarbons (e.g. propane, isobutene, neo-pentane) as well as some proposed oxygenated

The experimental apparatus and flow sampling system are shown

in the figure on the left. Uncertainties in measured species are as
follows: CH3COCH3 - 4%, O2 - 2%; CO - 2%; CO2 - 2%;
CH2O - 6%; CH4 - 2%; C2H2 - 3%; C2H4 - 3%; C2H6 - 4%.

Main N2 Carrier Gas

fuels (aldehydes, alcohols, ethers).

Acetone is widely used as a marker species in two-phase

Oxygen Inlet

Electric Resistance Heater

vaporization/combustion studies (using PLIF), it is an

the atmosphere,

and

its

Experiments were performed at 960 K for equivalence ratios of

0.6, 1.0, and 1.6 at an initial fuel concentration of 1,800 ppm.

Fuel Evaporator

Wall Heaters
Sample Probe

A reactivity experiment was also performed for = 1.0 scanning

a temperature range from 530 to 950 K.

Slide Table

ketene (CH2CO).

Fuel Injector

Understanding of acetone kinetics as well as an expanded

The initial model developed, presented here, consists of 46

species undergoing 248 reversible reactions.

Mixer/Diffuser

validation database are, therefore, important. To date,

The model was assembled in a hierarchal manner and is built

upon a H2/C1-C2 submodel recently developed in our laboratory
based on reaction schemes for H2 [1], H2/CO/CH2O/CH3OH [2],
and C2H5OH [3]. These baseline models have recently been used
to generate a comprehensive dimethyl ether model [4].

however, kinetic modeling and experimental studies of this

molecule under combustion-relevant conditions are scarce.
Here, we present data obtained in a Variable Pressure Flow

Reactor Duct Material: Fused Silica

Pressure:
0.2 - 20 atm
Temperature:
300 - 1200 K

Reactor (VPFR) for the oxidation of acetone under high

VPFR Sampling System

pressure (12.5 atm), and intermediate to high temperatures for

Multiport
sampling
valves

a range of stoichiometries.
It is shown how available acetone kinetic models fail to

Acetone-specific kinetic steps include reaction with O2 as well as

thermal decomposition as initiation processes. Interactions with
the radical pool include abstractions by H, O, OH, HO2, CH3,
CH3O, and C2H3. Rates were obtained from a wide literature
survey (see table). The rate expression of abstraction by OH was
fitted to available experimental data using least-squares analysis.

Back pressure regulator

Exhaust

reproduce the data collected in this study.

Teflon diaphragm pump

As a first attempt, a model based on the hierarchal nature of

The model includes

Needle valve

Sample probe

Only acetonyl (CH3COCH2) decomposition is considered

reacting systems has been assembled based on previous

Flow controller
Calibration gases

Reaction; (cm3/mol/sec/cal); (*) - This study

acetone decomposition and abstraction reactions obtained from

an exhaustive literature review.

Flow meter
FTIR cell

The model is shown to adequately reproduce the present flow

reactor data as well as ignition delay.

Cold trap

Further work is

warranted to optimize the model which is presently the subject

O2

CO 2

Exhaust

CO

of further work.

Results

Ref.

6.310E-16

0.00

8.145E+04

[5]

CH3COCH3+H=CH3COCH2+H2

5.630E+07

2.00

7.700E+03

[6]

CH3COCH3+O=CH3COCH2+OH

1.000E+13

0.00

5.961E+03

CH3COCH3+OH=CH3COCH2+H2O

1.783E+00

3.96

1.369E+03

(*)

CH3COCH3+O2=CH3COCH2+HO2

1.200E+14

0.00

4.600E+04

[6]

CH3COCH3+HO2=CH3COCH2+H2O2

1.700E+13

0.00

2.046E+04

[6]

CH3COCH3+CH3=CH3COCH2+CH4

Mole Fraction

Comparison against species time histories obtained for = 0.6, 1.0, and 1.6 (Figs. 2, 3, and 4 respectively) show
relatively good agreement although the predictive capability of the model deteriorates as the equivalence ratio increases.
Less satisfactory are the comparisons against water (not shown for clarity) which are considerably overpredicted (see Fig.
1).

3.162E+13

0.00

1.648E+04

[5]

1.000E+11

0.00

7.000E+03

[6]

0.00

4.270E+03

[8]

CH3COCH2=CH2CO+CH3

1.000E+14

0.00

3.100E+04

[6]

New high pressure acetone oxidation experiments have been performed in a

flow reactor.
A preliminary kinetic model has been developed based on the hierarchal
nature of reacting systems and shown to reasonably reproduce the present
flow reactor data as well as data obtained in other experimental systems.
additional refinement is needed at the small molecular level. Sensitivity
analyses suggest that ketene kinetics are important for acetone oxidation
which is the subject of further studies currently underway.
J. Li; Z. Zhao; A. Kazakov; F.L. Dryer, Int J Chem Kinet 36 (2004) 566-575.
J. Li; Z. Zhao; A. Kazakov; M. Chaos; F.L. Dryer; J.J. Scire, Jr., Int J Chem Kinet (2006), submitted.
J. Li, Ph.D. Thesis; Dept. Mechanical and Aerospace Engineering, Princeton University (2004).
Z. Zhao, M. Chaos, A. Kazakov, F.L. Dryer, Int J Chem Kinet (2006), submitted.
C. N. Cawood; H. Furue; S.H. Mousavipour; P.D. Pacey, J Phys Chem A 107 (2003) 1896.
Mechanisms found at http://www-cms.llnl.gov/combustion/combustion2.html.
J.T. Herron, J Phys Chem Ref Data 17 (1988) 967-1026.
K. Scherzer; Loser, U.; Stiller, W., Z Chem 27 (1987) 300.
K. Sato; Y. Hidaka, Combust Flame 122 (2000) 291-311.

This work was supported by the Office of Basic Energy

Sciences, U.S. Department of Energy under Grant No.
DE-FG02-86ER-13503.

600

700
800
T (K)

0.4

0.8
1.2
Time (s)

1.6

Figure 2
0.003

tshift = -0.60 s

Mole Fraction

O2 x 0.25
CH3COCH3

0.001

900

0.004
CO x 0.5
CO2
CH4

0.003

tshift = -0.70 s

O2 x 0.5
CH3COCH3

0.002
0.001

0.002

CO x 0.5
CO2
CH4

O2 x 0.5
CH3COCH3

0.001

0
0

0.4

0.8
1.2
Time (s)

1.6

0.4

0.8
1.2
Time (s)

Figure 3

1.6

Figure 4

12000

1000

100

Mixture D - CH3COCH3/O2/Ar - 1/2/97%

10

CO x 0.5
CO2
CH4

0.002

Figure 1

Ignition Delay (s)

Summary

tshift = -0.53 s

500

Figures 5 and 6 show model predictions against data collected in a shock tube [9]. Ignition delays are relatively well
reproduced for both pure acetone as well as hydrogen-doped acetone/O2/Ar mixtures (Fig. 5). Stable species measured in
the single-pulse shock tube are shown in Fig. 6; good agreement is observed, however, the model overpredicts ketene
(CH2CO) and ethylene while underpredicting methane.
Results from sensitivity analyses under both flow reactor and shock tube conditions are shown in Figs. 7 and 8. Figure 7
shows the sensitivity of selected species for the fuel-rich case shown in Fig. 4 at 65% fuel consumption. The system is
sensitive to reactions involving methyl, formaldehyde, and the formyl radical. In addition, and not surprisingly, ketene
plays an important role in acetone oxidation. Figure 8 shows ignition delay sensitivity for the case shown in Fig. 5
(Mixture D, 1500 K). Most of the important reactions are at the small molecular level (C < 2) with acetone thermal
decomposition and abstractions by H and CH3 playing an important role.

O2 x 0.5
CO
CO2
H2O
CH2O x 10

Comparison against a recently published acetone model (developed for high temperature applications) [9] shows its
inability to predict the present results as evidenced by the slow fuel evolution shown by the dashed line in Fig. 2.

[7]

1.229E+11

0.003

0.001

CH3COCH3+C2H3=CH3COCH2+C2H4

0.003
0.002

CH3COCH3+CH3O=CH3COCH2+CH3OH

0.004

Figures 1-4 show data from the present flow reactor experiments compared against model results. As opposed to model
predictions obtained using models available from Lawrence Livermore (see dashed line in Fig. 1), acetone does not
exhibit NTC behavior and the present model captures this trend.

1.
2.
3.
4.
5.
6.
7.
8.
9.

Ea

CH3COCH3=CH3+CH3CO

Mole Fraction

models developed in our laboratory.

Reactor Section Dia.: 10 cm

Mass Flow rate:
10 - 30 g/s
Residence Time:
0.0015 - 5 s

Mole Fraction

species in

oxidation/pyrolysis generates reactive intermediates, such as

Mixture G - CH3COCH3/O2/H2/Ar - 0.1/1/1/97.9%

Mole Fraction (ppm)

important tracer

PrincetonUniversity

CH3COCH3
CO
CH4

C2H2
C2H4

6000

C2H6
CH2CO

8000
4000
4000

2000

Model

0.6

0.65

0.7

0.75

0.8

0
1200 1400 1600

1200 1400 1600

1000/T (1/K)

Temperature (K)

Figure 5

Figure 6

CH3+HO2=CH3O+OH
CH2CO+OH=CH2OH+CO
CH3COCH3+HO2=CH3COCH2+H2O2
CH2O+HO2=HCO+H2O2
CH3COCH3+CH3=CH3COCH2+CH4
H+O2(+M)=HO2(+M)
CH3COCH3=CH3CO+CH3
H+O2=O+OH
CH3+O2=CH2O+OH
CH3COCH3+O2=CH3COCH2+HO2
HCO+O2=CO+HO2
HCO+M=H+CO+M
CH3COCH3+H=CH3COCH2+H2
CH3COCH3+OH=CH3COCH2+H2O
HO2+OH=H2O+O2
CH2O+OH=HCO+H2O
CH2CO+H=CH3+CO
HO2+HO2=H2O2+O2
CH3+CH3(+M)=C2H6(+M)
CH3+HO2=CH4+O2

CH3COCH3
O2
CO
H2O

H+O2<=>O+OH
C2H4+H(+M)<=>C2H5(+M)
CH3COCH3<=>CH3CO+CH3
CH3+OH<=>CH2(S)+H2O
CH2+O2<=>HCO+OH
CH3+HO2<=>CH3O+OH
C2H3+O2<=>O+CH2HCO
C2H2+H(+M)<=>C2H3(+M)
HCCO+O2<=>OH+2CO
C2H6+O<=>C2H5+OH
CH3+HCCO<=>C2H4+CO
2CH3(+M)<=>C2H6(+M)
C2H4+O<=>CH3+HCO
CH3COCH3+CH3<=>CH3COCH2+CH4
C2H6+OH<=>C2H5+H2O
C2H6+H<=>C2H5+H2
CH3COCH3+H<=>CH3COCH2+H2
C2H3+H<=>C2H2+H2
2CH3<=>H+C2H5
CH3+H(+M)<=>CH4(+M)

-0.002
0
0.002 0.004
Sensitivity Coefficient

-1 -0.8 -0.6 -0.4 -0.2 0 0.2

Sensitivity Coefficient

Figure 7

Figure 8