Plug-flow Reactor

CATALYSIS, KINETICS AND REACTION ENGNEERING Chinese Journal of Chemical Engineering, 21(11) 1269—1283 (2013) DOI: 10.1016/S1004-9541(13)60624-2

Chemical Effects of CO2 Concentration on Soot Formation in Jet-stirred/Plug-flow Reactor* ZHANG Yindi (张引弟)1,2,**, LOU Chun (娄春)2, LIU Dehua (刘德华)1, LI Yong (李勇)3 and RUAN Longfei (阮龙飞)1 1

Yangtze University Research for China National Petroleum Corporation Key Laboratory of Oil Gas Production, Yangtze University, Wuhan 430100, China 2 State Key Laboratory of Coal Combustion, Huazhong University of Science and Technology, Wuhan 430074, China 3 Gas Lift Technology Center of Tuha Oilfield of China National Petroleum Corporation, Hami 838202, China Abstract Soot formation was investigated numerically with CO2 addition in a jet-stirred/plug-flow reactor (JSR/PFR) C2H4/O2/N2 reactor (C/O ratio of 2.2) at atmospheric pressure. An updated Kazakov mechanism emphasizes the effect of the O2/CO2 atmosphere instead of an O2/N2 one in the premixed flame. The soot formation was taken into account in the JSR/PFR for C2H4/O2/N2. The effects of CO2 addition on soot formation in different C2H4/O2/CO2/N2 atmospheres were studied, with special emphasis on the chemical effect. The simulation shows ZZX CO + OH is responsible of the reduction of hydrocarbon intermediates that the endothermic reaction CO2 + H YZZ in the CO2 added combustion through the supplementary formation of hydroxyl radicals. The competition of CO2 for H ZZX O + OH radical through the above forward reaction with the single most important chain branching reaction H + O2 YZZ reduces significantly the fuel burning rate. The chemical effects of CO2 cause a significant increase in residence time and mole fractions of CO and OH, significant decreases in some intermediates (H, C2H2), polycyclic aromatic hydrocarbons (PAHs, C6H6 and C16H10, etc.) and soot volume fraction. The CO2 addition will leads to a decrease by only about 5% to 20% of the maximum mole fractions of some C3 to C10 hydrocarbon intermediates. The sensitivity analysis and reaction-path analysis results show that C2H4 reaction path and products are altered due to the CO2 addition. Keywords fuel enrichment, carbon dioxide, kinetics modeling, soot formation, jet-stirred/plug-flow reactor

1

INTRODUCTION

Exhaust gas recirculation systems and oxy-fuel burners with CO2 recycling have been used successfully in a number of industrial processes and shown to improve combustion efficiency and reduce pollutant emissions. However, these industrial processes are mostly accompanied by NOx and soot particles emission which have a detrimental effect on human health and contribute significantly to global warming [1]. So, the radiative properties of oxy-fuel flames and the role of soot is a persistent problem in many processes that involve fuel-rich combustion of hydrocarbon fuels [2]. Much research has been devoted to soot formation in flames of methane, acetylene, ethylene and benzene burning at reduced or atmospheric pressures [3-9]. Most of those kinetic simulations were performed up to a specified aromatic size: two to four rings. Several groups extended the detailed description to include soot particle dynamics [4-7]. In their efforts, the transition from the gas phase to soot particles is described assuming nucleation takes place with collisions of pyrene and larger aromatics or benzene molecules. Surface growth is treated either empirically or based on chemical analogy to aromatic chemistry. Coagulation of soot particles is modeled either using a

discrete sectional method or the method of moments. Although the exact chemical pathways and processes that lead to soot are not completely understood, it is generally agreed that soot formation from gas phase hydrocarbons mainly involves such steps: formation of the first ring, formation of polycyclic aromatic hydrocarbons (PAHs), soot inception, and subsequently soot growth [10-12]. A number of investigations have been focused on reducing soot emission by introducing several techniques, including the use of different fuels and addition of diluents on the oxidizer or fuel side in diffusion flames. Especially, study of the effect of CO2 addition on soot emission is important for suppressing soot production tendency, such as those by Du and co-workers [13], Ni and co-workers [14], and Liu et al [15]. These investigations found that the addition of CO2 in either fuel or in oxidizer has chemical effects on soot formation reduction, which might be to promote the concentrations of oxygen atom and hydroxyl that in return increase the oxidation of soot precursors in soot formation regions, in addition to the dilution and thermal effects. More recently, two experimental investigations were conducted by Renard et al. [16] and Vandooren et al. [17] on CO2 addition to rich C2H4/O2/Ar and CH4/O2/Ar premixed flames. The results showed that

Received 2012-10-28, accepted 2013-05-08. * Supported by the Foundation of State Key Laboratory of Coal Combustion, the National Natural Science Foundation of China (51306022, 51176059) and the Natural Science Foundation of Hubei Province (2013CFB398). ** To whom correspondence should be addressed. E-mail: [email protected]

1270

Chin. J. Chem. Eng., Vol. 21, No. 11, November 2013

CO2 and H2O addition are responsible for reduction of hydrocarbon intermediates in flames and the addition of CO2 suppresses soot emission chemically in both diffusion and premixed flames. However, the chemical effects of CO2 addition on flame structure in premixed flames and suppression of sooting tendency remains unclear in a jet-stirred/plug-flow reactor (JSR/PFR) reactor, especially the computational investigation is not sufficient. The main objective of this study is to numerically investigate the mechanism of chemical effect of CO2 addition on soot formation in a JSR/PFR reactor at atmospheric pressure. An updated mechanism of C2H4/air combustion based on Kazakov et al. [6] with emphasis on CO2 addition was modified from previous studies in the literature. The main combustion concentration profiles and soot volume fraction have been predicted using the updated mechanism for combination of soot particles by Frenklach [12] in a JSR/PFR reactor. In addition, the effects of O2 and CO2 concentrations variation on flame characteristics and soot formation are studied in a JSR/PFR reactor. The kinetic models were scrutinized with sensitivity, rate of production analysis and path of reaction analysis. 2 COMBUSTION MECHANISMS AND SOOT MODEL 2.1

Reaction mechanism without CO2 additive

A detailed kinetic model of soot formation, from the perspective of computer implementation, can be considered to consist of two principal components: gas-phase chemistry, which determines the flame structure, and soot particle dynamics, which describes the evolution of the particle ensemble. The correctness of the particle dynamics sub-model relies, first of all, on the accuracy of the species profiles supplied by the gas-phase sub-model, that define the soot particle nucleation and surface growth rates. In this work, the gas-phase original comprehensive reaction mechanism was developed by Kazakov et al. [6] for the oxidation of C2H4 in air combustion. The detailed chemical kinetic model consisted of 680 reactions and 156 species. Calculations were performed using the Premixed Laminar Model and JSR/PFR Model of the CHEMKIN-PRO package [18]. Thermodynamic properties were obtained from CHEMIKIN database [19]. Transport properties were obtained from the Sandia CHEMIKIN transport data base [20] and Frenklach and Wang [21]. The soot volume fraction was predicted based on the work of Frenklach and Wang [21] and has been verified by Marr [5] against the experimental data in a JSR/PFR reactor without CO2 addition to the atmosphere. The main species involving soot precursors and soot formation are discussed in detail, such as reactants, products, acetylene, benzene, pyrene and so on. The role of resonantly stabilized radicals were investigated in aromatic, branched aromatic and polycyclic aromatic

hydrocarbon formation in a premixed, rich, laminar, C2H4/O2/Ar flame and in a rich C2H4/O2/N2 JSR/PFR reactor. For soot volume fraction predictions, a modified version of the original soot model of Frenklach and Wang [21] has been used, where the main feature is implementation of detailed PAH chemistry into the coagulation and surface growth modules allowing for the selection of a set of PAHs involved in soot inception by PAH coagulation. 2.2

Reaction mechanism with CO2 additive

The computed results on PAH formation for C2H4 oxidation in different O2/CO2 atmospheres are analyzed in terms of a detailed gas-phase chemical kinetic model, which was updated from the mechanism of Kazakov et al [6]. This mechanism provides in general reasonably good agreement between modeling and experimental results, even though some modifications are needed to take into account the effect of the presence of high CO2 concentration. An updated mechanism was developed by the present authors to identify the chemical effects of an additive (CO2) introduced to the fuel mixture. Therefore, in the present work, the mechanism is updated, with particular emphasis on the effect of the presence of an O2/CO2 atmosphere instead of air or O2/Ar. A number of reactions involving CO2 species which do not appear in the initial model but may be important during the C2H4 oxidation process under oxy-fuel conditions are included in this model. Moreover, revisions of rate constants of the most sensitive reactions in the mechanism under the specific conditions studied were modified based on previous studies in the literature. The modifications are listed in Table 1. These modifications and the main reactions of interest are described below. Calculations are performed using the PREMIX model. The reverse rate constants were taken from the same source [6] as the mechanism. In establishing the reaction mechanism, special emphasis was put on reactions that converted CO2 to CO (Table 1), thereby activating chemically the comparatively inert CO2 molecule. Thermal dissociation of CO2, ZZX CO2(+M) CO + O(+M) YZZ

(R1)

is strongly endothermic and occurs only at high temperature. Reactions of CO2 with free radicals may proceed at lower temperature. At medium temperature, the reaction with atomic hydrogen, ZZX CO + OH CO2 + H YZZ

(R2)

is important. This reaction would be considered in establishing a CO/CO2 partial equilibrium under combustion conditions. The rate constant for the reaction is known accurately over a wide range of temperatures and pressures [19-29]. Two reactions of other species with CO2, i.e.: ZZX CO + HO2 CO2 + OH YZZ (R4)

1271

Chin. J. Chem. Eng., Vol. 21, No. 11, November 2013 Table 1 Reaction

R1

ZZX CO2(+M ) CO + O(+M) YZZ ②

low-pressure limit Troe parameters: 1.0, 10−30, and 10−30 ZZX CO2 + H CO + OH YZZ

③

β

Ea/kJ·mol−1

Ref.

1.8×10

10

0

9.981808

[19, 22]

1.4×10

24

−2.79

17.54772

[22, 23]

8.0×1010

0

0

[24]

3.994398

8.8×105

1.77

R3

low-pressure limit Troe parameters: 1.0, 10−30, and 10−30 ZZX HOCO CO + OH YZZ

2.0×1016

−5.6

12.06275

[24]

R4

ZZX CO2 + OH CO + HO2 YZZ

1.0×105

2.18

751.1478

[25]

R5

ZZX CO2 + O CO + O2 YZZ

4.7×1012

R6

ZZX CO2 + H HOCO YZZ low-pressure limit Troe parameters: 1.0, 10−30, and 10−30 ZZX CO2 + H2O HOCO + OH YZZ

0

253.3135

[26]

3.5×10

56

−15

194.6955

[24]

2.5×10

69

−18

251.22

4.6×1012

0

−0.37264

6

2

−0.37264

R8

low-pressure limit Troe parameters: 1.0, 10−30, and 10−30 ZZX CO2 + H2O HOCO + OH YZZ

9.9×1011

0

0

[28]

R9

ZZX CO2 + H HCO + O YZZ

3.0×1013

0

0

[29]

R10

ZZX CO2 + OH + H HCO + HO2 YZZ

R11

ZZX CH3 + CO2 CH3O + CO YZZ

9.5×1025

R12

3

R13

1

R7

②

R14 ① ② ③

A/cm3·mol−1·s−1

No.

R2

Reactions subset for CO2①

9.5×10

3.0×10

13

[27]

0

0

[30]

−4.93

38.01796

[31]

ZZX CH2O + CO CH2 + CO2 YZZ

1.0×10

11

0

4.187

[32]

ZZX CH2O + CO CH2 + CO2 YZZ

1.1×1013

0

0

[32]

ZZX HCO + CO CH + CO2 YZZ

8.8×106

1.75

−4.35448

[11]

Parameters for use in the modified Arrhenius expression k = ATβexp(−E/RT). Units are moles, centimeters, seconds, and joule. Enhanced third-body efficiencies: H2 = 2.5, H2O = 12, CO = 1.9, CO2 = 3.8. Expressed as the sum of the constants.

ZZX CO + O2 CO2 + O YZZ (R5) are important for slow reactions. The reverse steps, CO + HO2 (R4) and CO + O2 (R5), were significant in CO oxidation under conditions with low OH concentrations, but recent work [30, 31] has proven them to be much slower than indicated by the early estimates. The most abundant radical in oxidation process of C2H4 is CH3. The methyl radical may react with CO2 to form CH3O and CO: ZZX CH3O + CO (R11) CH3 + CO2 YZZ The reaction has only been measured in the reverse direction and shown little consistency [30]. Following Rasmussen et al. [24], this value was adopted in [32], assuming that CH3 + CO2 is the sole product channel from CH3O + CO. It is expected that this rate constant leads to an over-prediction of the CH3 + CO2 reaction rate, but the reaction is quite slow and only of minor importance at the conditions of the present study. Reactions of small hydrocarbon fragments with CO2 include 3

ZZX CH2O + CO CH2 + CO2 YZZ

(R12)

1

ZZX CH2O + CO CH2 + CO2 YZZ

(R13)

ZZX HCO + CO CH + CO2 YZZ

(R14)

Radical CH2 has two states that are both important in combustion processes. Compared to the singlet methylene (1CH2), triplet state (3CH2) is more reactive toward radicals but reacts slower with stable molecules. Rate constants for 3CH2 + CO2 reaction in room temperature have been measure by Laufer and Bass [31], which are used in this work. The rate constant for the reaction between CH and CO2 (R14) was measured over a temperature range (298 to 3500 K) [33]. The available data for Arrhenius expression was shown in Table 1. 2.3

Soot model

A computational model for the prediction of soot formation in flames consists of three components. Firstly, initial PAH formation in the gas phase chemistry, which determines the flame structure, formation of the first aromatic ring, and its subsequent growth up to a prescribed size. Secondly, planar PAH growth, homogeneous nucleation of PAH clusters beyond the prescribed size. Thirdly, particle coagulation, particle surface reactions (growth and oxidation), and particle agglomeration results in spherical particle are formed and grown/destroyed.

1272

Chin. J. Chem. Eng., Vol. 21, No. 11, November 2013

2.3.1 Gas-phase chemistry The C2H4/O2/Ar combustion mechanism [6] is used in the simulation. This reaction mechanism consists of 156 species and 680 reactions which include PAH growth reactions up to pyrene, A4. The code names for some species are shown in Table 2. According to Frenklach et al. [34], soot particles are created by the dimerization of pyrene molecules. 2.3.2 Growth of PAHs beyond pyrene While still within the realm of gas-phase processes, the growth of PAHs beyond pyrene is treated as a separate part of the model—to differentiate between reactions assigned on an individual basis, though still categorized by classes on physical grounds, and those whose numerical treatment is founded entirely on the reaction-class concept. The latter is described as an infinite polymerization process with the technique of linear lumping [34]. The equations, as specifically applied to the PAH problem, are reported in [34]. The reaction classes adopted for this part of the model of Kazakov et al. [6] with the rate coefficients modified as described above for the initial part of the gas-phase PAH reactions. It is pertinent to mention that the adopted model for PAH growth beyond pyrene is of minimal size, comprised of only the dominant mass-addition route—that of acetylene. 2.3.3 Particle dynamics Particle inception, coagulation, PAH condensation, and surface reactions are treated by solving the Smoluchowski master equations [35] with the method of moments [36]. 2.3.4 Soot mass growth and oxidation reactions One of the advantages of the Particle Tracking Feature is that soot mass growth and oxidation reactions Table 3 No.

①

Table 2 No.

Aromatic species

Name

Structure

1

phenyl, A1

C6H5

2

benzene, A1

C6H6

3

toluene

C7H8

4

benzyl, C7H7

C6H5CH2

5

phenyl acetylene, A1C2H

C6H5C2H

6

ethynylphenyl radical, A1C2H

7

styrene, A1C2H3

8

phenylvinyl radical, A1C2H3

9

n-styryl, n-C8H7

*

C6H4CCH C6H5C2H3

*

C6H4CH

CH2

C6H5CH

CH

10

indene

C9H8

11

naphthalene, A2

C10H8

12

biphenyl, P2

C12H10

13

methylnaphthalene, A2CH3

C11H10

14

ethynylnaphthalene, A2C2H

C12H8

15

acenaphthalene, A2R5

C12H8

16

phenanthrene, A3

C14H10

17

methylphenanthrene

C14H12

18

phenanthrylacetylene, A3C2H

C16H10

19

pyrene, A4

C16H10

20

pyreneacetylene, A4C2H

C18H10

can be provided as regular surface reactions. For example, the H-Abstraction-C2H2-Addition (HACA) soot growth sequence proposed by Frenklach et al. [34] can be given in Table 3.

Soot growth and oxidation mechanism①

Reaction

k = ATnexp(−E/RT) −1

3

A/cm ·mol ·s

−1

n

E/kJ·mol−1

Ref.

S1

2A4 ⎯⎯ → 32C(B) + 20H(S) + 28.72(S)

2.00×108

0.5

0

37

S2

H + H(S) ⎯⎯ → (S) + H2

4.20×1013

0.0

54.431

37

S3

→ H(S) + H H2 + (S) ⎯⎯

3.90×10

12

0.0

39.02284

37

S4

H + (S) ⎯⎯ → H(S)

2.00×1013

0.0

0

37

S5

H(S) + OH ⎯⎯ → H2O + (S)

10

0.734

5.98741

37

S6

→ OH + H(S) H2O + (S) ⎯⎯

3.68×10

8

1.139

71.5977

37

S7

C2H2 + (S) ⎯⎯ → H(S) + 2C(B) + H

8.00×107

1.56

15.9106

37

S8

A1 + 6H(S) ⎯⎯ → 6C(B) + 6(S) + 6H2

0.2

0.0

0

37

S9

A1C2H + 6H(S) ⎯⎯ → 8C(B) + 6(S) + 6H2

0.21

0.0

0

37

S10

→ 10C(B) + 16(S) + 12H2 A2 + 16H(S) ⎯⎯

0.1

0.0

0

37

S11

A2R5 + 16H(S) ⎯⎯ → 10C(B) + 16(S) + 11H2 + C2H2

0.1

0.0

0

37

S12

A3 + 20H(S) ⎯⎯ → 14C(B) + 20(S) + 15H2

0.1

0.0

0

37

S13

→ 16C(B) + 20(S) + 15H2 A4 + 20H(S) ⎯⎯

0.1

0.0

0

37

S14

OH + (S) + C(B) ⎯⎯ → CO + H(S)

0.20

0.0

33.496

37

1.00×10

(S): site, active data; (B): bulk, solid-phase component; n is the reaction order.

1273

Chin. J. Chem. Eng., Vol. 21, No. 11, November 2013

The model assumes that nucleation of soot particles is due to the coalescence of two large size PAHs, pyrene (A4), into a dimer. Then the particle size increases or decreases due to the particle coagulation, surface growth and oxidation. In this paper, simulation models for the original gas phase chemistry was used by Kazakov et al. [6] without CO2 addition; for planar PAH growth, homogeneous nucleation of PAH clusters was reported in [34]; for particle coagulation, particle surface reactions (growth and oxidation), and particle agglomeration was developed by Frenklach et al [34].

Here ρ is the (mass) density and u the axial velocity of the gas, which consists of Kg species; Wk is the molecular mass of species k, and sk is the molar production rate of this species by all surface reactions. The quantities A and ai,m are the cross-sectional (flow) area and the effective internal surface area per unit length of the reactor, respectively. Both A and ai,m can change as arbitrary functions of x. Eq. (1) states simply that the mass flow rate of the gas can change as a result of generation or consumption by surface reactions on all materials in the reactor. A similar equation can be written for each species individually. gas-species conservation equation:

3 NUMERICAL METHODS AND EXPERIMENTAL DATA 3.1

Model and governing equations

The experimental JSR/PFR system developed at MIT [5] provides a good platform for kinetic studies of soot formation and growth because, under the JSR and PFR conditions, the influence of mass diffusion on gas phase species profiles is minimized. The JSR serves as the pre-heat and flame zones of a premixed flame and the PFR is used to simulate the postflame region. From the prospect of model simulation, the JSR/PFR implementation reduces greatly the complexity of the numerical process as well as the run time. The JSR/PFR experiment can be modeled by one PSR (perfectly stirred reactor) and two PFRs shown in Fig. 1 in series. The first PSR is for the upstream (or flame zone) JSR, the following PFR is to model the transition zone between JSR and PFR in the experimental setup, and the last PFR is for the postflame PFR where measurement was performed. The main purpose of the transition PFR is to allow the JSR exhaust to cool down from 1630 K to 1620 K before entering the test section. The diagram view of this three reactor network is given in Fig. 1. Since Particle Tracking Feature is activated by special keywords in surface reaction mechanism, all soot simulations will need both gas phase and surface chemistry input files. In this paper, the PSR and PFR models are employed to simulate one flame of premixed C2H4/O2/N2. The simulation includes mole fractions of major gas phase species and mass concentrations of PAH and soot at various locations inside the PFR. The equations governing the behavior of a plug-flow reactor (PFR) incorporated in CHEMKIN-PRO are simplified versions of the general relations for conservation of mass, energy, and momentum. They can be described as follows. mass continuity equation: Kg dA du dρ M ρu + ρ A + uA = ∑ ai,m ∑ sk ,mWk (1) dx dx dx m =1 k =1

Figure 1

ρ uA

K

M g dYk + Yk ∑ ai,m ∑ sk ,mWk dx m =1 k =1

⎛M ⎞ (2) = Wk ⎜ ∑ sk , m ai, m + ω k A ⎟ ⎝ m =1 ⎠ Here Yk is the mass fraction of species k and ω k is its molar rate of production by homogeneous gas reactions. Such reactions cannot change the total mass of the gas, but they can alter its composition. energy equation:

⎛ Kg

dYk dT du ⎞ + Cp + u ⎟⎟ + dx dx ⎠ ⎝ k =1 dx Kg ⎛ Kg 1 ⎞M ⎜⎜ ∑ hk Yk + u 2 ⎟⎟ ∑ ai, m ∑ sk , mW 2 ⎠ m =1 k =1 ⎝ k =1

ρ uA ⎜⎜ ∑ hk

M

= ae Qe − ∑ ai,m m =1

Kb

∑

k = K bf

sk ,mWk h

(3)

where hk is the specific enthalpy of specie k, C p is the mean heat capacity per unit mass of the gas, T is the (absolute) gas temperature. In the right-hand summation, sk ,m is the molar production rate of bulk solid species k by surface reactions on material m. ae is the surface area per unit length, Qe is the heat. K bf is one of bulk solid species between the K bf and K b . Momentum equation: for the gas it expresses the balance between pressure forces, inertia, viscous drag, and momentum added to the flow by surface reactions. Thus, K

A

M g dP du dF + ρ uA + + u ∑ ai,m ∑ sk ,mWk = 0 (4) dx dx dx m =1 k =1

where P is the absolute pressure and F is the drag force exerted on the gas by the tube wall, to be discussed below. The pressure is related to the density via the ideal gas equation of state. A perfectly stirred reactor (PSR) consists of a

Diagram of the system used to simulate the JSR/PFR experiment

1274

Chin. J. Chem. Eng., Vol. 21, No. 11, November 2013

chamber, which may or may not allow heat loss, having inlet and outlet ducts. There may be more than one inlet defined for each reactor. Conservation of mass, energy, and species for a well mixed reactor (PSR) or homogeneous system include net generation of chemical species within the reactor volume, and net loss of species and mass to surfaces in the reactor. These equations are stated as follows. global mass balance equation: N

d ( ρV )( j ) = dt N PSR

∑ m

(r )

r =1

Rrj − m

( j)

inlet ( j )

∑ i =1 M

+∑

m =1

m i∗( j ) +

Am( j )

Kg

∑

k =1

gas energy equation:

sk( ,jm) Wk

j = 1," , N PSR

dU s(jsj ) (5)

Here j is the reactor number, ρ is the mass density, V is the reactor volume, m ∗ is the inlet mass flow rate, and m is the outlet mass flow rate. Ninlet ( j ) is the number of inlets for each reactor j, while N PSR is the total number of reactor modules in the reactor network. Rrj is the fraction of the outflow of reactor r that is recycled into reactor j. The outlet mass flow differs from the sum of the inlet and recycled mass flow when deposition or etching of materials within the reactor occurs, as represented by the last term on the right-hand side. In this term, Am is the surface area of the mth material defined within the reactor, and Sk ,m is the molar surface production rate of the kth species on the mth material per unit surface area. There are K g gas-phase species and M materials. species conservation equation:

( ρ kV )

( j)

dYk( j ) = dt

N

inlet ( j )

∑ i =1

(

)

m i∗( j ) Yk∗,i − Yk +

N PSR

M

Kg

m =1

k =1

∑ m k(r ) Rrj ⎡⎣Yk(r ) − Yk( j ) ⎤⎦ − Yk( j ) ∑ Am( j ) ∑ sk( ,jm) Wk + r =1

(ω kV )

( j)

M

Wk + ∑ Am( j ) sk( ,jm) Wk m =1

In Eq. (6), Yk is the mass fraction of the kth species, Wk is the molecular mass of species k, and ω k is the molar rate of production of species k by gas-phase chemical reaction per unit volume. The superscript * indicates inlet stream quantities. For steady-state conditions, the nominal residence time τ in the reactor is ρV τ= N (7) N PSB ⎡ inlet( j ) ⎤ ⎢ ∑ m i∗( j ) + ∑ m i( r ) Rrj ⎥ r =1 ⎣⎢ i =1 ⎦⎥

(6)

dt

N

=

inlet ( j )

∑ i =1

Kg

(

mi∗( j ) ∑ Yk∗,i hk∗,i k =1

N PSR

Kg

r =1

k =1

∑ m k(r ) Rrj ∑ (Yk hk )

(r )

)

( j)

+

−

dV ( j ) j = 1," , N PSR (8) dt The total internal energy Usjs consists of the internal energy of the gas, surface phases, deposited or etched solid phases, and walls. Qloss is the net heat flux directed out of the reactor. Qsource is the net heat flux generated by the heat source. ( j) ( j) Qloss + Qsource − P( j )

3.2 Reactor geometry, boundary conditions and numerical method

In this work, the geometric model and boundary conditions are shown in Fig. 2. The combined reactor model includes one PSR and two PFRs. The reactor geometric parameters, inlet initial conditions and boundary conditions are as shown in Fig. 2. The initial conditions of fuel and oxidant components respectively is X(C2H4) = 0.1005, X(O2) = 0.1370 and X(N2) = 0.7625. PFR reactor assumed that there is no backmixing in the axial (flow) direction but perfect mixing in the transverse direction. The lack of transverse gradients

Figure 2 Calculation geometry and boundary conditions

Chin. J. Chem. Eng., Vol. 21, No. 11, November 2013

implies that all entering fuel elements have the same flow and reaction behavior. The plug flow reactor model is also computationally efficient since its model is first-order ordinary differential equations (ODE’s). The equations governing the behavior of a plug-flow reactor are simplified versions of the general relations for conservation of mass, energy, and momentum et al. [Eqs. (1)-(8)]. 4

RESULTS AND DISCUSSION

The experimental data used in this work is from the post-flame of fuel rich combustion in a closely coupled jet-stirred/plug-flow reactor system (JSR/PFR) investigated by Marr [5]: Equivalence ratio Φ = 2.2, reactant components and reaction conditions were X(C2H4) = 0.1005, X(O2) = 0.1370, X(N2) = 0.7625, X(CO2) = 0, P = 0.1 MPa, v0 = 9.85 g·s−1. Well characterized combustion environments are fed into the PFR from combustion of C2H4/O2/N2 fuel mixtures within JSR at Φ = 2.2, at a pressure of 0.1 MPa, and a temperature of 1620 K. The PFR emulates the post-flame zone of a flame and provides 14 ms of near-isothermal plug flow. Fixed gases (CO, CO2, H2), light hydrocarbons (C1 through C4 species), PAHs (2 to 9 fused rings, including cyclopenta compounds, cyclophanes, and aliphatic-substituted PAH), tars (all PAH soluble in methylene chloride), and soot (methylene chloride insoluble material) were sampled from the PFR using a water-cooled stainless steel probe, which is inserted at a fixed distance from the top of the flow straightener. Since the gas velocity through the PFR is known for each of the experimental conditions (≌25 m·s−1), the distance from the top of the flow straightener to the probe tip converts directly into residence time, and the analyzed species concentration profiles are plotted as mole fraction vs. residence time for each PFR condition studied. The JSR/PFR experiment can be modeled by one PSR and two PFR’s in series. In this work, five cases in JSR/PFR reactor were simulated, and the basic conditions were as follows: P = 0.1 MPa, Φ = 2.2 in

1275

C2H4/O2/N2 flame without CO2 addition in JSR/PFR reactor. Several other calculations were executed at P = 0.1 MPa, Φ = 2.2 in C2H4/O2/CO2/N2 flame with CO2 replacing part of the N2 in JSR/PFR reactor. 4.1

Base case in JSR/PFR reactor (no CO2 addition)

One simulation with both H-abstraction-C2H2addition (HACA) and PAH condensation growth reactions is performed for the soot mass growth mechanism. Results of the TJSR = 1630 K and Φ = 2.2 case are presented in Figs. 3 to 5. PFR2 was used for analysis. Soot formation is generally considered in the literature [5] being prone to occur within a certain average temperature range, such as from 1300 K to 1800 K. Therefore, soot precursors and some intermediates which influences soot formation are chosen as monitoring species, such as H2, OH, CO, CO2, C2H2, C2H4, C6H6, C10H8, C14H10 and C18H10, etc. Figure 3 shows the comparison of predicted mole profiles some intermediates and C1, C2 species with the experimental profiles in the C2H4/O2/N2 flame at Φ = 2.2. As can be seen from Fig. 3, the good agreement is obtained for CO2, CO and C2H4 between the prediction and experimental data [Fig. 3 (b)]. The CO2 and CO show the upward trend with the PFR residence time increase, but C2H4 shows the opposite trend. This is because C2H4 is reactant which decreases with enhances of reaction time, but CO2 and CO is combustion products which increase with enhance of reaction time. It also can be seen that some intermediates, such as H2, OH [see Fig. 3 (a)] and C2H2 [see Fig. 3 (b)], which indicate some different. The key factor for the difference is the limitation of experimental conditions, such as the intermediates are not easily measured. However, H2, OH and C2H2 show the consistent trend with enhance of reaction time. The results are acceptable for consideration the experimental conditions. Figure 4 shows the comparison of predicted mole profiles some intermediates species with the experimental profiles in the C2H4/O2/N2 flame at Φ = 2.2.

(a) Comparison for H2 and OH (b) Comparison for CO2, CO, C2H4 and C2H2 Figure 3 Comparison of predicted mole profiles some intermediates and C1, C2 species with the experimental profiles in the C2H4/O2/N2 flame at Φ = 2.2, T = 1620 K line: modeling; scatter: Marr-data

1276

Chin. J. Chem. Eng., Vol. 21, No. 11, November 2013

(b) Comparison for C14H10 (A3) and C18H10 (A4) (a) Comparison for C6H6 (A1) and C10H8 (A2) modeling; ■ Marr-data [5]; ● Harris [38]; ▲ Fwar [37] modeling ■ Marr-data [5]; ▲ Grow [10]; Figure 4 Comparison of predicted mole profiles some intermediates species with the experimental profiles in the C2H4/O2/N2 flame at Φ = 2.2, T = 1620 K

surely affect the simulation results in the PFR section behind the JSR. Comparison of the predicted and measured soot mass concentration profiles in the PFR is presented in Fig. 5. The soot mass growth rate predicted by the HACA + PAH mechanism shows a much better agreement with the experimental data. According to Marr [5], the soot nucleation is still occurring as the gas mixture entering the PFR. 4.2 Figure 5 Comparison of predicted soot volume fraction with experimental data in the C2H4/O2/N2 flame at Φ = 2.2 HACA + PAH; ■ Marr data [5]

As can be seen from Fig. 4, the good agreement is obtained for C6H6 (A1) and C10H8 (A2) between the prediction and experimental data [Fig. 4 (a)]. The C6H6 (A1) and C10H8 (A2) show the downward trend with the PFR residence time increase. But some intermediates, such as C14H10 (A3) and C18H10 (A4) [see Fig. 4 (b)], which indicate some different. The specific reasons will be described as follow. Fig. 5 shows the comparison of predicted soot volume fraction with experimental data in the C2H4/O2/N2 flame at Φ = 2.2. As can be seen from these figures (Figs. 3 to 5), the predictions obtained by the Particle Tracking Feature are in good agreement with the experimental data, but in general the gas-phase species are slightly under-predicted in comparison to Marr’s data [5]. For C6H6 (A1) and C10H8 (A2), predictions of mole fraction over-predict Harris [38] [Fig. 4 (a)]. There are many factors that can contribute to the discrepancies shown in Figs. 4 and 5. For example, Marr [5] did not provide details of the composition and temperature of the inlet gas mixture to the JSR. Since the temperature of the JSR is maintained by adjusting the N2 fraction in the inlet gas stream, uncertainties in inlet condition, reactor heat loss, and reactor residence time will

With CO2 addition in JSR/PFR reactor

An updated mechanism was developed to identify the chemical effects of an additive (CO2) introduced to the fuel mixture in Table 1. Calculations are performed using the PREMIX model. The reverse rate constants were taken from the same source as the mechanism. Four simulations were conducted, unlike basic case, with CO2 replacing part of N2 in C2H4/O2/N2 flame at Φ = 2.2 in this section. Figure 6 shows mole fractions of H and OH with CO2 replacing part of N2 in C2H4/O2/N2 flame at Φ = 2.2. The mole fraction of OH increases with increasing CO2 additive, just the opposite of the change in H in Fig. 6. This is because the endothermic process ZZX CO + OH (R2), which is the most imCO2 + H YZZ portant reaction associated with the direct chemical participation of CO2, is comparatively fast even at medium temperatures and responsible for the promoted concentration of hydroxyl. R2 has been identified as the primary pathway for the chemical effect of CO2 in several previous studies [39]. This reaction would be expected to be the dominating reaction in establishing a CO/CO2 partial equilibrium under combustion conditions. The competition of CO2 for the H radical through the forward reaction of R2 with the single ZZX most important chain branching reaction H + O2 YZZ O + OH [39] significantly reduces the concentrations of important radicals, i.e. O, H, and OH, leading to significant reduction of the fuel burning rate. The reduced fuel burning rate extends the reaction time. So, the residence time increases with increasing CO2 additive (see Fig. 7).

Chin. J. Chem. Eng., Vol. 21, No. 11, November 2013

Figure 6 Comparison of H and OH mole fraction with different CO2 addition in C2H4/O2/CO2/N2 flames at Φ = 2.2, T = 1620 K O2/N2 (20/80); O2/CO2/N2 (20/20/60); O2/CO2/N2 (20/40/40); O2/CO2/N2 (20/60/20); O2/CO2 (20/80)

1277

Figure 7 Comparison of residence time with different CO2 addition in the C2H4/O2/CO2/N2 flames at Φ = 2.2, T = 1620 K O2/N2 (20/80); O2/CO2/N2 (20/20/60); O2/CO2/N2 (20/40/40); O2/CO2/N2 (20/60/20); O2/CO2 (20/80)

(a) CO (b) CO2 Figure 8 Comparison of CO and CO2 mole fraction with different CO2 additive in the C2H4/O2/CO2/N2 flames at Φ = 2.2, T = 1620 K O2/N2 (20/80); O2/CO2/N2 (20/20/60); O2/CO2/N2 (20/40/40); O2/CO2/N2 (20/60/20); O2/CO2 (20/80)

Figure 7 shows comparison of residence time with different CO2 addition in the C2H4/O2/CO2/N2 flames at Φ = 2.2. It can be found that the residence time increases with increasing CO2 additive. In addition to the reasons consistent with Fig. 6, another factor is that with increasing CO2 fraction, the numerator increases in Eq. (7) slower than the denominator, due to the fuel burning rate is reduced. Soot oxidation by OH is the most important oxidation for soot. The increased mole fraction of OH around the flame as shown in Fig. 6 also helps reduce soot emission through elevated oxidation attack of these species on soot nuclei precursors. So, the increase in OH enhances soot oxidation reducing soot volume fraction with increasing CO2 additive. It is certain that all factors would contribute to soot emission reduction chemically. According to [39], the presence of CO2 will compete with O2 for atomic hydrogen and lead to the forZZX CO mation of CO through the reaction CO2 + H YZZ + OH (R2). Reactions of CO2 with hydrocarbon radicals may also contribute to CO formation. So, it can be found that CO increases with increasing CO2 addition in Fig. 8 (a). The result provides a favorable condition

to further reduce the soot nucleation rate. From Figs. 6 to 8, it can be found that the effects of CO2 are to increase the OH/H ratio, lower the overall concentration of the O/H radical pool, and increase the availability of CO. It was also observed that the concentration of acetylene (C2H2), the dominant soot precursor species, also decreases as a result of CO2 addition [see Fig. 9 (a)]. From Fig. 9 (b), the increase in CO2 addition enhances molar conversion of C2H2, which not only is converted into soot precursors (C3H3, A1), but also converted into other species. A1 molar conversion decreases with increasing CO2 addition [Fig. 10 (a)]. So the increase in CO2 addition enhances C2H2 conversion into non-soot-precursor species. Single-ring aromatic benzene (C6H6, A1) and polycyclic aromatic hydrocarbon pyrene (C16H10, A4) are two important soot precursor species. From Figs. 10 and 11, it can be seen that both species and A1 molar conversion decrease with increasing CO2 addition because [16, 17] the endothermic process CO2 + ZZX CO + OH (R2) is responsible for the reducH YZZ tion of the hydrocarbon intermediates in the CO2 added flames through the supplementary formation of hydroxyl radicals. It has been demonstrated that such

1278

Chin. J. Chem. Eng., Vol. 21, No. 11, November 2013

(b) Molar conversion of C2H2 (a) C2H2 mole fraction Figure 9 Comparison of C2H2 mole fraction and molar conversion with different CO2 addition in the C2H4/O2/CO2/N2 flames at Φ = 2.2, T = 1620 K O2/N2 (20/80); O2/CO2/N2 (20/20/60); O2/CO2/N2 (20/40/40); O2/CO2/N2 (20/60/20); O2/CO2 (20/80)

(b) Molar conversion of C6H6 (A1) (a) C6H6 (A1) mole fraction Figure 10 Comparison of C6H6 (A1) mole fraction and molar conversion with different CO2 addition in the C2H4/O2/CO2/N2 flames at Φ = 2.2, T = 1620 K O2/N2 (20/80); O2/CO2/N2 (20/20/60); O2/CO2/N2 (20/40/40); O2/CO2/N2 (20/60/20); O2/CO2 (20/80)

Figure 11 Comparison of C18H10 (A4) mole fraction with different CO2 addition in the C2H4/O2/CO2/N2 flames at Φ = 2.2, T = 1620 K O2/N2 (20/80); O2/CO2/N2 (20/20/60); O2/CO2/N2 (20/40/40); O2/CO2/N2 (20/60/20); O2/CO2 (20/80)

processes come into play at the end of the front and become more efficient in the burnt gases region. However, it can be seen that molar conversion is negative when CO2 increases to a certain component [shown in Fig. 10 (b)]. And C3H3 involved reactions are C3H3 + ZZX C2H3 + HCO and C2H2 + CH2 YZZ ZZX C3H3 + H, OH YZZ the first reaction consumes C3H3 and the latter produces C3H3. But the first reaction is dominant between

the two reactions. In the first reaction, OH compete C3H3 produces C2H3 and HCO resulting in the formation decreases of A1. It can be speculated that this trend will be more obvious with CO2 increasing in the flame. This may be the reasons A1 molar conversion rate is negative shown in Fig. 10 (b). Figure 12 shows the soot volume fraction with CO2 replacing part of the N2 in C2H4/O2/N2 flame at

Chin. J. Chem. Eng., Vol. 21, No. 11, November 2013

1279

Figure 12 Comparison of prediction soot volume fractions in the C2H4/O2/CO2/N2 flames at Φ = 2.2, T = 1620 K O2/N2 (20/80); O2/CO2/N2 (20/20/60); O2/CO2/N2 (20/40/40); O2/CO2/N2 (20/60/20); O2/CO2 (20/80)

Figure 13 Comparison of maximum prediction mole profiles of main C3-C10 species (C3H3, C6H6, C6H6O, C8H8, C10H8) with experimental profiles in the C2H4/O2/Ar and C2H4/O2/Ar/CO2 flames at Φ = 2.5

Φ = 2.2. The increasing CO2 addition decreases soot volume fraction, as seen in Fig. 12. From Figs. 9 and 10, some important soot precursor species (C2H2, A1, A4, etc.) decrease when the increasing CO2 addition caused nucleation reduction. Also the residence time increases with increasing CO2 addition (see Fig. 7) significantly leading to a significant reduction of the fuel burning rate in soot formation. Also, an increase in OH enhances soot oxidation to reduce soot volume fraction with increasing CO2 addition. It is worth noting that the reduction of soot emission by CO2 addition is partially caused by the flame temperature reduction because of enhanced thermal radiation heat loss, in addition to the mechanisms discussed in the literature. 4.3

Validation of the updated mechanism

The experimental data of rich premixed C2H4/O2/Ar flames for one without any additive (labeled as Φ2.50) and one with 15% of CO2 replacing the same quantity of argon (labeled as Φ2.50C) were obtained by Renard et al. [16] to validate the updated mechanism. In this section, two calculations were respectively

conducted in C2H4/O2/Ar and C2H4/O2/Ar/CO2 flames at Φ = 2.5. And calculation conditions were as follows: Case 1 P = 0.1 MPa, Φ = 2.5 in C2H4/O2/Ar flame without CO2 addition. Calculation conditions were: equivalence ratio Φ = 2.5, reactant components and reaction conditions were: X(C2H4) = 0.3300, X(O2) = 0.4000, X(Ar) = 0.27, X(CO2) = 0, P = 0.1 MPa, v0 = 40.3 g·s−1. Case 2 P = 0.1 MPa, Φ = 2.5 in C2H4/O2/CO2/N2 flame with CO2 replacing part of the Ar. Calculation conditions were: equivalence ratio Φ = 2.5, reactant components and reaction conditions were: X(C2H4) = 0.3300, X(O2) = 0.4000, X(Ar) = 0.12, X(CO2) = 0.15, P = 0.1 MPa, v0 = 40.3 g·s−1. Calculations are performed using the PREMIX model. The reverse rate constants were taken from the same source as the mechanism [6]. In this section, only mole fraction profiles of important species involving soot formation are discussed in detail. Figure 13 shows the maximum mole fractions of some C3 to C10 species at equivalence ratio of 2.50. The simulated results agree well with experimental data, even if the prediction of C6H6O is a little higher than the experiment. CO2 addition usually leads to a decrease by only about 5% to 20% of the maximum

1280

Chin. J. Chem. Eng., Vol. 21, No. 11, November 2013

However, despite the extensive work on the elementary reactions leading to the first aromatic ring, neither the dominant benzene (C6H6, A1) formation pathway nor the sensitivity of benzene (C6H6, A1) in C2H4/O2/N2 and C2H4/O2/CO2 atmosphere is yet well understood. Figure 14 shows the results of C6H6 (A1) normalized sensitivity coefficients of 17 reactions for C2H4/O2/N2 [Fig. 14 (a)] and C2H4/O2/CO2 [Fig. 14 (b)] flames at 0.1 MPa and temperature 1630 K. There are three reactions (R234, R224 and R193) at 1630 K:

mole fractions of some C3 to C10 hydrocarbon intermediates (C3H3, C6H6, C6H6O, C8H8, C10H8). 4.4

Sensitivity analysis and reaction path analysis

The important steps in soot formation from gas phase hydrocarbons are believed to be formation of the first ring, formation of polycyclic aromatic hydrocarbons (PAHs), soot inception, and subsequently soot growth. It is now widely accepted that benzene and phenyl formation constitutes the first step in this growth process that lead to PAH and ultimately soot particles.

2C3H3 ⎯⎯ → A1 ZZX C2H3 + HCO C3H3 + OH YZZ

(R234) (R224)

(a)

(b) Figure 14 Comparison of sensitivity analysis without and with CO2 additive in the C2H4/O2/N2 flames at Φ = 2.2

Chin. J. Chem. Eng., Vol. 21, No. 11, November 2013

ZZX C3H3 + H C2H2 + CH2 YZZ

1281

Finally, the typical reaction-path diagrams in Fig. 15 represent the main reactions and species leading to benzene formation in C2H4/O2/N2 atmosphere [Fig. 14 (a)] and C2H4/O2/CO2 atmosphere [Fig. 15 (b)]. Such diagrams, resulting from combining Fig. 14, are useful in identifying the difference in benzene formation in the different atmospheres. In Figs. 15 (a) and 15 (b), the hydroxide radical (OH), oxygen radical (O) and other side species (H) are chosen as side species. Any reaction with the hydroxide radical is dot dash line, any reaction with the oxygen radical is dashed line,

(R193)

which have absolute values of sensitivity coefficients that are larger in both flames. However, there are some reactions that are different in the two flames; such as R451, R375, R239, R369, R152, R89, R69, R60 and R2. These nine reactions appear in Fig. 14 (a) but are replaced by R257, R205, R85, R76, R75, R73, R31, R3 and R1 in Fig. 14 (b). So, the C2H4 oxidation is changed due to CO2 addition. Some reactions that have higher sensitivity coefficients are not presented here.

(a)

(b) Figure 15 Comparison of reaction analysis without and with CO2 additive in the C2H4/O2/N2 flames at Φ = 2.2

1282

Chin. J. Chem. Eng., Vol. 21, No. 11, November 2013

and any reaction with H is solid line. Each substantial reaction is represented by an arrow, the thickness of which is representative of the relative importance of that particular reaction of formation. The diagrams are not representative of all reactions in the mechanism but only concern the main pathways leading to benzene formation. From Fig. 15, benzene formation is mainly achieved in two ways: C2H2 → C3H3 → A1(C6H6) and C2H2 → A1(C6H6) in both C2H4/O2/N2 and C2H4/O2/CO2 atmosphere. From Fig. 15 (a), besides conversion of small amounts to CO, most of the C3H3 converts to A1 in C2H4/O2/N2 atmosphere. However, besides most of the C3H3 converting to A1 and CO, a certain amount of C3H3 converts to HCO in Fig. 15 (b). So, this is a factor that the soot volume fraction decreases in the C2H4/O2/CO2 flame compared to the C2H4/O2/N2 flame. Comparing Figs. 15 (a) and 15 (b), there are different main paths between the two. They are C2H4 → C2H2 → C4H2 → i-C4H2 → CH2CO → CO2, C2H4 → C2H2 → C4H2 → i-C4H2 → CH2CO → CH3 → CH4 in the C2H4/O2/N2 flame [Fig. 15 (a)] and C2H4 → C2H2 → C4H2 → H2C4O → CH2CO → CH3 CH2O → HCO in the C2H4/O2/N2 flame [Fig. 15 (b)]. It can be concluded that CO2 in the atmosphere changed the path of C2H4 oxidation. 5

to a decrease by only about 5% to 20% of the maximum mole fractions of some C3 to C10 hydrocarbon intermediates. The sensitivity analysis and reaction-path analysis on JSR/PFR reactor for C2H4/O2/N2 and C2H4/O2/N2/CO2 flames show that C2H4 reaction path and products are altered due to the CO2 addition. REFERENCES 1 2 3 4 5 6 7

CONCLUSIONS 8

A kinetic modeling study was carried out to investigate the mechanism of the chemical effects of CO2 addition on the soot formation in a JSR/PFR reactor under atmospheric pressure at equivalence ratios Φ = 2.2. An updated mechanism, which was applied and shown to reproduce reasonably well the concentration profiles of major, intermediate, and aromatic species in ethylene flames, focused on the effect of the presence of an O2/CO2 atmosphere instead of air. Numerical results were verified against experimental data. Modeling predictions of reactants, stable intermediates, and PAH compounds generally showed good agreement with experimental data. The simulation shows that the endothermic procZZX CO + OH (R2), which is the most esses CO2 + H YZZ important reaction associated with the direct chemical participation of CO2, is comparatively fast even at medium temperature and is responsible for the promoted concentration of hydroxyl. The competition of CO2 for H radical through the forward reaction of R2 with the single most important chain branching reacZZX O + OH significantly reduces the tion H + O2 YZZ fuel burning rate. The effect of CO2 is to increase the OH/H ratio, lower the overall concentration of the O/H radical pool, and increase the availability of CO. So, the chemical effects of CO2 cause significant increases in residence time and mole fractions of CO and OH, and significant decreases in some intermediates (H, C2H2), polycyclic aromatic hydrocarbons (PAHs, C6H6 and C16H10) and soot volume fraction. The results also show that CO2 addition usually leads

9 10

11 12 13 14 15

16

17

18

19

Ce'cile, C., Christian, C., Iskender, G.Ö., Dilek, F.K., “CO2 addition and pressure effects on laminar and turbulent lean premixed CH4 air flames”, Proc. Combust. Inst., 32, 1803-1810 (2009). Kwang, C.O., Hyun, D.S., “The effect of oxygen and carbon dioxide concentration on soot formation in non-premixed flames”, Fuel, 85, 615-624 (2006). Zhang, Y.D., Zhou, H.C., Xie, M.L., Fang, Q.Y., Wei, Y., “Modeling of soot formation in gas burner using reduced chemical kinetics coupled with CFD code”, Chin. J. Chem. Eng., 18, 967-978 (2010). Maricq, M.M., “A comparison of soot size and charge distributions from ethane, ethylene, acetylene, and benzene/ethylene premixed flames”, Combust. Flame, 144, 730-743 (2006). Marr, J.A., “PAH chemistry in a jet stirred/plug flow reactor system”, Ph.D. thesis, Massachusetts Institute of Technology, Cambridge, USA (1993). Kazakov, A., Wang, H., Frenklach, M., “Detailed modeling of soot formation in laminar premixed ethylene flames at a pressure of 10 Bar”, Combust. Flame, 100, 111-120 (1995). Zhang, Y., Lou, C., Xie, M., Fang, Q., Zhou, H., “Computation and measurement for distributions of temperature and soot volume fraction in diffusion flames”, Journal of Central South University of Technology, 18, 1263-1271 (2011). Li, S.L., Jiang, Y., Chen, W.T., “Numerical analysis on the characteristics of soot particles in C2H4/CO2/O2/N2 combustion”, Chin. J. Chem. Eng., 21, 238-245 (2013). Chen, Y., Zhang, J., “Effects of gas temperature fluctuation on the soot formation reactions”, Chin. J. Chem. Eng., 21, 25-30 (2013). Kazakov, A., Frenklach, M., “Dynamic modeling of soot particle coagulation and aggregation: Implementation with the method of moments and application to high-pressure laminar premixed flames”, Combust. Flame, 114, 484-501 (1998). Violi, A., D’Anna, A., D’Alessio, A., “Modeling of particulate formation in combustion and pyrolysis”, Chem. Eng. Sci., 54, 3433-3442 (1999). Frenklach, M., “Reaction mechanism of soot formation in flames”, Phys. Chem. Chem. Phys., 4, 2028-2037 (2002). Du, D.X., Axelbaum, R.L., Law, C.K., “The influence of carbon dioxide and oxygen as additives on soot formation in diffusion flames”, Proc. Combust. Inst., 23, 1501-1507 (1990). Ni, T., Gupta, S.B., Santoro, R.J., “Suppression of soot formation in ethene laminar diffusion flames by chemical additives”, Proc. Combust. Inst., 25, 1379-1385 (1994). Liu, F., Guo, H., Smallwood, G.J., Gulder, Ö., “The chemical effects of carbon dioxide as an additive in an ethylene diffusion flame: Implications for soot and NOx formation”, Combust. Flame, 125, 778-787 (2001). Renard, C., Dias, V., van Tiggelen, P.J., Vandooren, J., “Flame structure studies of rich ethylene-oxygen-argon mixtures doped with CO2, or with NH3, or with H2O”, Proc. Combust. Inst., 32, 631-637 (2009). Vandooren, J., Thill, L., Musick, M., van Tiggelen, P.J., “Depletion of soot precursors by CO2 addition to rich hydrocarbon flames”, In: 20th Task Leaders Meeting of the IEA Implementing Agreement, Energy Conservation and Emission Reduction in Combustion, Ottawa, 26-29 (1998). Gordon, S., McBride, B.J., Computer Program for Calculation of Complex Chemical Equilibrium Compositions, Rocket Performance, Incident and Reflected Shocks and Chapman-Jouget Detonations, National Aeronautics and Space Administration, Washington (1971). Troe, J., “Thermal dissociation and recombination of polyatomic molecules”, Proc. Combust. Inst., 15, 667-680 (1975).

Chin. J. Chem. Eng., Vol. 21, No. 11, November 2013 20

21 22 23 24

25

26 27 28

Kee, R.J., Dixon-Lewis, G., Warnatz, J., Coltrin, M.E., Miller, J.A., “A Fortran computer code package for the evaluation of gas-phase, multicomponent transport properties”, Sandia Report #SAND 86-8246, Sandia National Laboratories (1986). Frenklach, M., Wang, H., “Detailed mechanism and modeling of soot particle formation”, In: Soot Formation in Combustion, Springer Berlin Heidelberg, Germany vol. 59, 165-192 (1994). Allen, M.T., Yetter, R.A., Dryer, F.L., “High pressure studies of moist carbon monoxide/nitrous oxide kinetics”, Combust. Flame, 109, 449-470 (1997). Westmoreland, P.R., Howard, J.B., Longwell, J.P., “Prediction of rate constants for combustion and pyrolysis reactions by bimolecular QRRK”, AIChE J., 32, 1971-1979 (1986). Rasmussen, C.L., Hansen, J., Marshall, J., Glarborg, P., “Experimental measurements and kinetic modeling of CO/H2/O2/NOx conversion at high pressure”, Int. J. Chem. Kinet., 40 (8), 454-480 (2008). You, X., Wang, H., Goos, E., Sung, C.J., Klippenstein, S.J., “Reaction kinetics of CO + HO2→ products: Ab initio transition state theory study with master equation modeling”, J. Phys. Chem. A., 111, 4031-4042 (2007). Yu, H.G., Muckerman, J.T., Francisco, J.S., “Direct ab initio dynamics study of the OH + HOCO reaction”, J. Phys. Chem. A., 109, 5230-5236 (2005). Nolte, J., Grussdorf, J., Temps, F., Wagner, H.G., “Kinetics of the reaction HOCO + O2 in the gas phase”, Z. Naturforsch. A., 48, 1234-1238 (1993). Baulch, D.L., Bowman, C.T., Cobos, C.J., Cox, R.A., Just, T., Kerr, J.A., Pilling, M.J., Stocker, D., Troe, J., Tsang, W., Walker, R.W., Warnatz, J., “Evaluated kinetic data for combustion modeling: Supplement II”, J. Phys. Chem. Ref. Data, 34, 757-1397 (2005).

29 30 31 32 33

34 35 36 37 38 39

1283

Wang, H., Frenklach, M., “Calculations of rate coefficients for the chemically activated reactions of acetylene with vinylic and aromatic radicals”, J. Phys. Chem., 98, 11465-11478 (1994). Glarborg, P., Miller, J.A., Kee, R.J., “Kinetic modeling and sensitivity analysis of nitrogen oxide formation in well-stirred reactors”, Combust. Flame, 65, 177-202 (1986). Laufer, A.H., Bass, A.M., “Rate constants of the combination of methyl radicals with nitric oxide and oxygen”, Chem. Phys. Lett., 16, 151-158 (1977). Koch, M., Temps, F., Wagener, R., Wagner, H.G., “Kinetics of the reactions of CH2 (ã1A1) with CH3C2H, HCN, CO2, N2O and COS”, Ber. Bunsenges Phys. Chem., 94, 645-650 (1990). Mehlmann, C., Frost, M.J., Heard, D.E., Orr, B.J., Nelson, P.F., “Rate constants for removal of CH (D)(v = 0 and 1) by collisions with N2, CO, O2, NO and NO2 at 298 K and with CO2 at 296≤T/K ≤873”, J. Chem. Soc. Faraday Trans., 92, 2335-2341 (1996). Frenklach, M., Yuan, T., Ramachandra, M.K., “Soot formation in binary hydrocarbon mixtures”, Energy Fuels, 2, 462-473 (1988). Smoluchowski, M.V., “Engineering biological structures of prescribed shape using self-assembling multi-cellular systems”, Z. Phys. Chem., 92, 129-132 (1971). Frenklach, M., Harris, S.J., “Aerosol dynamics modeling using the method of moments”, J. Colloid Interface Sci., 118, 252 (1987). Appel, J., Bockhorn, H., Frenklach, M., “Kinetic modeling of soot formation with detailed chemistry and physics: Laminar premixed flames of C2 hydrocarbons”, Combust. Flame, 121, 122-136 (2000). Harris, S.J., Weiner, A.M., Blint, R.J., “Formation of small aromatic molecules in a sooting ethylene flame”, Combust. Flame, 72, 91-109 (1988). Mendiara, T., Glarborg, P., “Re-burn chemistry in oxy-fuel combustion of methane”, Energy Fuels, 23, 3565-3572 (2009).