ISSN 1068-364X, Coke and Chemistry, 2017, Vol. 60, No. 4, pp. 140–143. © Allerton Press, Inc., 2017.
COKE
Non-Isothermal Kinetics of Metallurgical Coke Gasification by Carbon Dioxide1 Zhongsuo Liu* and Qi Wang School of Materials and Metallurgy and Key Laboratory of Chemical Metallurgy Engineering, University of Science and Technology Liaoning, Anshan 114051, Liaoning, China *e-mail:
[email protected] Received March 20, 2017
Abstract⎯Under non-isothermal conditions, thermogravimetric analysis was applied to study carbon dioxide gasification of three metallurgical cokes. The cokes selected for the study were named Coke A, Coke B and Coke C. The experimental data are fitted using four common gas-solid kinetic models: the homogeneous model, the sharp interface model, the traditional model and the random pore model. It is found that the random pore model most closely reflects the kinetic behavior of coke gasification characteristics. Using the random pore model, the apparent activation energies for gasification of Coke A, Coke B, and Coke C were calculated to be 139.08, 127.78, and 116.32 kJ mol–1, respectively. Keywords: metallurgical coke, carbon dioxide gasification, kinetic model, thermogravimetric analysis DOI: 10.3103/S1068364X17040020
The carbon (coke) gasification reaction is a typical non-catalytic gas–solid reaction. To build a mathematical model which can describe quantitatively the carbon gasification process is an important topic in the research of carbon gasification kinetics. Involving the transformation of a solid reactant into gaseous products by means of a solid-gas reaction, carbon gasification is a very complex process [1]. The carbon gasification process corresponds to different kinetic parameters under different gasification conditions. The kinetic parameters calculated using different gasification models are also different from each other. In this work, we apply thermogravimetric analysis to study gasification of metallurgical cokes in carbon dioxide. We also compare the simulation results obtained using four common kinetic models and select the most suitable model for determining kinetic parameters. 1. KINETIC MODELS On the basis of different experiment conditions and samples, different gasification models have already been proposed. The most commonly used models include the homogeneous model (HM) [2], the sharp interface model (SIM) [3], the traditional model (TM) [4] and the random pore model (RPM) [5, 6]. The HM assumes that the reaction occurs within a particle. When the chemical reaction takes place, the 1 The article is published in the original.
size of the solid particle remains constant, but its density changes uniformly. Based on this assumption, the reaction rate of the first order reaction can be expressed as follows:
dx = k (1) HM (1 − x ) , dt where kHM is the reaction rate constant, related primarily to gasification agent concentration and gasification temperature. This model is concise and therefore widely applied. The SIM is also called the unreacted-core shrinking model [7]. The basic idea of this model is that the reaction is thought to occur on the geometric interface. The thickness of the geometric interface is zero. As the reaction progresses, the reaction interface gradually moves into the solid reactant. Its expression can generally be shown as: 2/3 dx = k (2) . SIM (1 − x ) dt The TM combines the HM and the SIM. The TM considers the experience factor and the physical meaning of some parameters. It is usually represented as follows: n dx = k (3) TM (1 − x ) . dt The RPM, proposed by Bhatia and Perlmutter, considers physical structural changes during the gasification reaction. This model assumes that the coal particle has numerous cylindrical pores of various
140
NON-ISOTHERMAL KINETICS OF METALLURGICAL COKE GASIFICATION
141
Table 1. Proximate and ultimate analysis of samples (wt %, dry basis) Proximate analysis
Ultimate analysis
Sample Coke A Coke B Coke C
A
M
V
FC
C
H
N
St
O
11.43 11.70 12.61
0.16 0.28 0.34
1.11 0.84 0.79
87.30 87.18 86.26
97.89 97.28 96.98
0.36 0.73 0.89
1.10 1.19 1.32
0.62 0.56 0.53
0.40 0.31 0.28
diameters. The reaction occurs mainly on the inner surface of the cylindrical pores. The random growing and overlapping of the pores increases and decreases the effective reaction area. The RPM can be expressed as:
dx = k (4) RPM (1 − x ) 1 − ψ ln (1 − x ), dt where ψ is the pore structure parameter, defined as follows: ψ=
4πL0 (1 − ε 0 )
(5) , S 02 where S0, L0, and ε0 represent the initial pore specific surface area, initial pore length, and initial solid porosity, respectively.
Figure 1 shows how carbon conversion degree changes with temperature. Reaction rate clearly increases with the increase of the temperature. Because the carbon gasification by carbon dioxide is a highly endothermic reaction process, the conversion rate is promoted by the rising temperature. Comparing the reaction rates of the samples, it is clear that the order is Coke A > Coke B > Coke C. Volatile matter has some impact on the microstructure of the cokes. During heating, volatile matter ejection can increase porosity and enlarge pore specific areas to enhance the gasification reaction. Coke A has a high volatile matter content; thus, the reaction rate of Coke A is the fastest. We used the correlation coefficient and the average standard deviation to evaluate the fitness of every model. The average standard deviation (e) is defined as: N
2. EXPERIMENTAL SECTION
3. RESULTS AND DISCUSSION The carbon conversion degree (x) can be defined as follows [8]:
m0 − mt (6) , m0 − m∞ where m0 is the sample mass at the start of the gasification, mt is the sample mass at reaction time t and m is the mass of the sample at the end of the reaction. x=
COKE AND CHEMISTRY
Vol. 60
No. 4
2017
e=
i exp
i − x pred
i =1
)
i x exp
N
(7)
,
i where x exp denotes the experimentally obtained car-
i bon conversion degree at point i, x pred denotes the predicted carbon conversion degree based on the selected model at point i and N denotes the total number of points in the gasification experiment. The average standard deviations of the four models fitting the sample gasification reaction are listed in
1.0 Conversion degree x
Three different kinds of cokes obtained from Ansteel were selected for this study. The samples had a mean particle size of 200 μm. The results of proximate and ultimate analyses of these samples are shown in Table 1. The gasification experiments were conducted using a Setsys Evolution thermogravimetric analyzer (Setaram Co., France) at atmospheric pressure. The sample (8 ± 0.5 mg) was placed in a platinum crucible with diameter 3 mm and height 1.5 mm. The temperature was recorded from 383 to 1750 K, with a constant heating rate of 5 K min–1, under high-purity carbon dioxide flowing at 40 mL min–1. To dry the samples, they were placed in an electric stove and heated at 383 K for 8 h. Carbon dioxide was passed through deoxidizing and dehydrating agents, and then entered the thermogravimetric analyzer.
∑ (x
0.8
Coke A Coke B Coke C
0.6 0.4 0.2 0 800
1000
1200 1400 Temperature, K
1600
1800
Fig. 1. Conversion degree and temperature for three cokes.
142
ZHONGSUO LIU, QI WANG
Table 2. Average standard deviations of predictions made by four kinetic models (%) Coke samples Model HM SIM TM RPM
coke A
coke B
coke C
5.520 5.825 2.287 1.137
5.812 5.126 2.869 1.096
5.918 5.674 2.438 1.261
Table 3. Kinetic parameters of three cokes, estimated from the random pore model Sample
Ea, kJ mol–1
k0, s–1
R2
Coke A
139.08
8.83 × 10
8
Coke B
127.78
2.36 × 10
7
0.9972
Coke C
116.32
5.71 × 106
0.9951
0.9944
Table 2. The best fitted theoretical model for the carbon gasification is the RPM, followed by the TM. Neither the HM nor the SIM is a sufficiently good fit for the experimental data in this study. The curves of carbon conversion degree with time (or temperature) are S-shaped. The gasification reaction rate increases first and then slows with increasing carbon conversion degree. The HM, the SIM and the TM do not describe this trend. In contrast, the RPM describes it well, allowing interpretation of the growing and cross-linking of the pores. At high temperatures, the reaction rate (dx/dt) varies greatly because the exponential term of (1 – x) becomes large. The exponential term of 0 Coke A Coke B Coke C
The Arrhenius plots of coke gasification reactions using the RPM are shown in Fig. 2. The Arrhenius plots of all samples show a linear relationship, suggesting that the RPM can well reflect the coke gasification reactions. The Arrhenius plots in Fig. 2 also provide the apparent activation energy and frequency factor of the coke gasification. The kinetic parameters determined by the RPM are listed in Table 3. The apparent activation energies of Coke A, Coke B, and Coke C are 139.08, 127.78, and 116.32 kJ mol–1, respectively. The apparent activation energy and frequency factor depend mostly on the rate constant (k) at different temperatures. We can see from Table 3, the order of apparent activation energy is Coke C > Coke B > Coke A, which reverses the order of the reaction rate. This shows that there is a kinetic compensation effect between activation energy and frequency factor. 4. CONCLUSIONS In this work, the carbon dioxide gasification of three metallurgical cokes under non-isothermal conditions was studied. The main conclusions are as follows. (1) In non-isothermal experiments, the reaction rates of the cokes increased with increasing temperature. The gasification reactivities of three cokes under carbon dioxide were investigated. Results show that the order is Coke A > Coke B > Coke C. (2) By comparing the models used in this study, it is found that the RPM is the best fitted theoretical model. This is because the change in pore structure has a strong effect on the gasification process.
lnk [min–1]
–1
(1 – x) may not be 1 or 2/3; therefore, the HM and the SIM do not describe the gasification process well. The TM is a semi-empirical model with better fitting precision than that of the HM or the SIM. Compared with the RPM, however, the fitting precision of the TM is not good enough. The RPM can satisfactorily fit the data of all samples in the present study.
–2
(3) Using the RPM, the apparent activation energies of gasification of Coke A, Coke B, and Coke C under carbon dioxide are 139.08, 127.78, and 116.32 kJ mol–1, respectively. The order of apparent activation energy is Coke C > Coke B > Coke A.
–3
–4
ACKNOWLEDGMENTS –5
6.5
7.0 7.5 –1 –4 –1 T , 10 K
8.0
Fig. 2. Arrhenius plots of coke gasification reactions for three cokes.
8.5
This work was financially supported by the National Natural Science Foundation of China (project no. U1361212) and the Science and Technology Program Project of Anshan City, China (project no. 3750). COKE AND CHEMISTRY
Vol. 60
No. 4
2017
NON-ISOTHERMAL KINETICS OF METALLURGICAL COKE GASIFICATION
REFERENCES 1. Silva, I.S., Palma, C., Lemos, F., et al., Non-catalytic carbon gasification modeling, Stud. Surf. Sci. Catal., 1997, vol. 109, no. 97, pp. 535–540. 2. Ishida, M. and Wen, C.Y., Comparison of zone-reaction model and unreacted-core shrinking model in solidgas reactions: I. Isothermal analysis, Chem. Eng. Sci., 1971, vol. 26, no. 7, pp. 1031–1041. 3. Yagi, S. and Kunii, D., Studies on fluidized-solids reactors for particles with decreasing diameters, Chem. Eng., Jpn., 1955, vol. 19, pp. 500–506. 4. Zhou, J., Zhou Z.-J., Gong, X., et al., Study of charCO2 gasification (I) by isothermal thermogravimetry, Coal Convers., 2002, vol. 25, no. 4, pp. 66–67.
COKE AND CHEMISTRY
Vol. 60
No. 4
2017
143
5. Bhatia, S.K. and Perlmutter, D.D., A random pore model for fluid-solid reactions: I. Isothermal kinetic control, AIChE J., 1980, vol. 26, no. 3, pp. 379–385. 6. Bhatia, S.K. and Perlmutter, D.D., A random pore model for fluid-solid reactions: II. Diffusion and transport effects, AIChE J., 1981, vol. 27, no. 2, pp. 247– 254. 7. Wen, C.Y., Noncatalytic heterogeneous solid-fluid reaction models, Ind. Eng. Chem., 1968, vol. 60, no. 9, pp. 34–54. 8. Liu, Z., Wang, Q., Zou, Z., et al., Arrhenius parameters determination in non-isothermal conditions for the uncatalyzed gasification of carbon by carbon dioxide, Thermochim. Acta, 2011, vol. 512, no. 1, pp. 1–4.