Thermodynamic analysis of methane reforming with CO2, CO2 + H2O, CO2 + O2 and CO2 + air for hydrogen and synthesis gas production

Journal of CO2 Utilization 7 (2014) 30–38

Contents lists available at ScienceDirect

Journal of CO2 Utilization journal homepage: www.elsevier.com/locate/jcou

Thermodynamic analysis of methane reforming with CO2, CO2 + H2O, CO2 + O2 and CO2 + air for hydrogen and synthesis gas production Antonio C.D. Freitas, Reginaldo Guirardello * School of Chemical Engineering, University of Campinas (UNICAMP), Av. Albert Einstein 500, 13083-852 Campinas, SP, Brazil

A R T I C L E I N F O

A B S T R A C T

Article history: Received 27 March 2014 Received in revised form 29 May 2014 Accepted 25 June 2014 Available online 18 July 2014

The main objective of this work is performing a thermodynamic evaluation of methane reforming with CO2, CO2 + H2O, CO2 + O2 and CO2 + air. These evaluations were carried out by Gibbs energy minimization, in conditions of constant pressure and temperature, and entropy maximization, at constant pressure and enthalpy, methods, to determine the equilibrium compositions and equilibrium temperatures, respectively. Both cases were treated as optimization problems (using non-linear programming formulation), satisfying the restrictions imposed by atom balance and non-negativity of number of moles. The GAMS123.1 software and the CONOPT solver were used in the resolution of the proposed problems. All calculations performed presented a low computational time (less than 1 s). The calculated results were compared with previously published experimental and simulated data with a good agreement between them for all systems. The H2 and syngas production were favored at high temperature and low pressure conditions. The addition of H2O or O2 proved to be an effective way to reduce the coke formation in the systems. The CO2 reforming presented endothermic behavior, but the addition of O2 or air reduced this trend and in some conditions autothermal behavior was observed. ß 2014 Elsevier Ltd. All rights reserved.

Keywords: Gibbs energy minimization Entropy maximization Methane reforming reactions Hydrogen production Synthesis gas production

1. Introduction In recent years, hydrogen has been attracting great interest as a clean fuel for combustion engines and fuel cells [1]. Among all the potential sources of hydrogen, natural gas, which has methane as main component, has been considered a good option because it is clean, abundant and it can be easily converted to hydrogen [2]. Currently, the main routes to produce hydrogen from methane are the catalytic reforming technologies, such as steam reforming (SR), dry reforming (DR), oxidative reforming (or partial oxidation) (OR) and autothermal reforming (ATR). Among these, the main industrial route to produce hydrogen and syngas from methane is SR, this reaction produces a syngas with a high H2/CO molar ratio (close to three) [3]. The dry reforming process becomes industrially advantageous when compared to steam reforming or partial oxidation for syngas production, as the H2/CO molar ratio in the product is close to 1.0/ 1.0 [4]. This low H2/CO ratio is suitable for further use in Fischer– Tropsch synthesis of long-chain hydrocarbons, dimethyl ether and

* Corresponding author. Tel.: +55 19 3521 3955; fax: +55 19 3521 3910. E-mail addresses: [email protected], [email protected] (Antonio C.D. Freitas), [email protected] (R. Guirardello). http://dx.doi.org/10.1016/j.jcou.2014.06.004 2212-9820/ß 2014 Elsevier Ltd. All rights reserved.

methanol; all of which require lower H2/CO ratios than that obtained by conventional SR process [5–7]. The major drawback of DR is that high temperatures are required to reach high conversion levels due to the highly endothermic nature of the process. These severe operating conditions combined with the tendency of the process to produce large quantities of coke (C(s)) result in deactivation of the catalysts by coke deposition [8,9]. The problem of C(s) deposition can be resolved either (i) by developing catalysts that minimize the rate of coke formation, or (ii) by adding steam [4,10–12], or oxygen [4,13–17] to the feed gas stream. The main possible reactions in CO2, CO2 + H2O and CO2 + O2 reforming’s process are summarized in Table 1. Research on thermodynamic behaviors of reaction systems by calculating equilibrium compositions have been utilized in understanding the feasibility of a variety of reactions [18–24]. The evaluation of the thermodynamic behavior of the reactions provides the first step to analyze the limits of temperature, pressure and feed ratios on equilibrium compositions. In the present work, a complete thermodynamic analysis of CO2, CO2 + H2O and CO2 + O2 reforming of methane were performed. The effect of molar feed compositions, pressure and temperatures were evaluated over the reaction performances. For this, we used the Gibbs energy minimization and entropy maximization methods to

A.C.D. Freitas, R. Guirardello / Journal of CO2 Utilization 7 (2014) 30–38 Table 1 Main reactions in the DR of methane.

Nomenclature Cp equation parameter for component i number of atoms of element m in component i Cp equation parameter for component i Cp equation parameter for component i heat capacity for component i Cp equation parameter for component i Gibbs energy enthalpy enthalpy of component i in phase k number of moles for component i in the phase k initial number of moles for component i number of components in the system number of elements in the system universal gas constant entropy entropy of component i in phase k pressure temperature

Cpai ami Cpbi Cpci Cpi Cpdi G H Hik nki n0i NC NE R S Ski P T

31

Reaction number

Reaction

0 DH298 K (kJ/mol)

1 2 3 4 5 6 7 8 9 10 11

CH4 + CO2 $ 2CO + 2H2 CO2 + H2 $ CO + H2O CH4 $ C(s) + 2H2 2CO $ C(s) + CO2 CO2 + 2H2 $ C(s) + 2H2O H2 + CO $ H2O + C(s) CH4 + 2H2O $ CO2 + 4H2 CH4 + H2O $ CO + 3H2 CH4 + 1/2O2 ! CO + 2H2 CH4 + 2O2 ! CO2 + 2H2O C(s) + 1/2O2 ! CO

247 41 75 172 90 131 165 206 36 802 110

Smith and Missen [25] demonstrated that the stoichiometric formulation is equivalent to the non-stoichiometric one, if all independent reactions are considered. The values of mgi can be calculated from the formation values given at some reference conditions, using the following thermodynamic conditions:   @Hi ¼ C pi i ¼ 1; . . . ; NC (4) @T P

Greek letter chemical potential of component i in the phase k

mki

@  mi  Hi ¼ i ¼ 1; . . . ; NC @T RT P R  T2

Superscripts g gas phase k phase in the element liquid phase l solid phase s

(5)

The CO2 reforming of methane typically occurs in low or moderate pressures (1–15 atm) and high temperatures (above 1000 K) thus, this work considered the hypothesis of ideal gas (fi = 1), the absence of liquid phase and the formation of solid carbon as pure component. Therefore, Eq. (1) can be simplified, and the Gibbs energy can be expressed as follows: 0 0 0 111 NC NC X X g @ g;0 g g n  m þ R  T @lnP þ @lnn  ln n AAA G¼

Subscripts component in the mixture i elements in component i m

i

i

i

i¼1

þ determine the equilibrium compositions and equilibrium temperatures, respectively. 2. Methodology 2.1. Equilibrium at constant P and T: formulation as a problem of minimization of Gibbs energy The thermodynamic equilibrium condition for reactive multicomponent closed system, at constant P and T, with given initial composition, can be obtained by minimization of Gibbs energy (G) of the system, given by: minG ¼

NC X

ngi mgi þ

i¼1g

NC X

nli mli þ

i¼1g

NC X

nsi msi

(1)

i¼1g

While satisfying the restrictions of non-negative number of moles of each component in each phase: ngi ; nli ; nsi  0

(2)

And the restriction of mole balances, given by atom balance for reactive systems:

NC X nsi  ms;0 i

i¼1

i¼1

(3)

(6)

j¼1

The Gibbs energy minimization was used to study the thermodynamic behavior of the system in isothermic conditions. The effects of reaction temperature, pressure and inlet compositions were evaluated under the main products composition. During the process of optimization, utilizing the Gibbs energy minimization method the number of moles of the gaseous ðngi Þ, liquid ðnli Þ and solid ðnsi Þ phase are considered decision variables, while T, P and the chemical potential of the pure component in the reference state ðn0i Þ are considered parameters. 2.2. Equilibrium at constant P and H: formulation as a problem of entropy maximization The thermodynamic equilibrium condition for reactive multicomponent closed systems, at constant P and H, with given initial composition, can be obtained by maximization of the entropy (S) of the system, with respect to nki : maxS ¼

NC NC NC X X X ngi  Sgi þ nli  Sli þ nsi  Ssi i¼1

NC NC X X ami  ðngi þ nli þ nsi Þ ¼ ami  n0i m ¼ 1; . . . ; NE

i

j¼1

i¼1

(7)

i¼1

While satisfying the same previous restrictions, given by Eqs. (2) and (3). Usually, physical properties are given as functions

A.C.D. Freitas, R. Guirardello / Journal of CO2 Utilization 7 (2014) 30–38

32

of composition, pressure and temperature, not enthalpy. Therefore, an additional restriction, referent to enthalpy balance, must be satisfied: NC NC X X ðngi  Hig þ nli  Hil þ nsi  His Þ ¼ ðn0i  Hi0 Þ ¼ H i¼1

(8)

i¼1

The entropy of each component in the mixture and the enthalpy balance can be determined using the following thermodynamic relations: !

@mki Ski ¼  @T Hik T2

¼

i ¼ 1; . . . ; NC

(9)

P;nki

@ mki @T T

! i ¼ 1; . . . ; NC

(10)

P;nki

The hypothesis of ideal gas (fi = 1), the absence of liquid phase and the presence of solid carbon as pure component were considered, so the entropy can be expressed as follows: !!! NC NC NC X X X g g;0 g g ni  Si  R lnP þ lnni  ln n j nsi S¼ þ i¼1

J¼1

i¼1

 Ss;0 i

(11)

The calculation of the thermodynamic equilibrium problem using entropy maximization method, at adiabatic conditions, allows evaluating systems that could result in hot spots in reactors undergoing exothermic reactions. In the maximization of entropy, the variables are: ngi , nli , nsi , T and all quantities that depend on them, such as physical properties of pure components (which depend on temperature) and molar fractions. The parameters are physical properties of pure components at some reference temperature, and the initial molar amount n0i . 2.3. Numerical procedure Although the formulated problem is non-linear, the used methodology guarantees the global optimum, since in this case the problem is convex [26]. The software GAMS1 23.2.1, (General Algebraic Modeling System) with the CONOPT solver was used in the resolution of the combined chemical and phase equilibrium problem. A description of GAMS software can be found in Brooke et al. [27]. The method of thermodynamic analysis by minimization of Gibbs energy is commonly employed and others works use this technique with excellent results [19,20,28]. The entropy maximization methodology is less used, but recent works show good results in applications of this technique [20,28,29]. The solid phase formed was considered as pure component, this consideration showed good results in previous works [18,20,23,24]. For each reaction, a detailed thermodynamic evaluation is presented, based on thermodynamic effects of temperature, pressure and reactant composition. The conditions under each

one of the methane reforming technologies were evaluated are presented in Table 2. All calculations were performed at pressures below 10 bar, mainly because the ideal gas consideration, that was used in the formulation of both models (Gibbs energy minimization and entropy maximization), this consideration restrict the use of these models under low-pressure conditions.

3. Results and discussion A thermodynamic analysis based on Gibbs energy minimization and entropy maximization was carried out for CO2, CO2 + H2O and CO2 + O2 reforming reactions. The thermodynamic equilibrium calculations showed a low computational time, in all cases less than 1 s. The main species in the methane reforming processes are CH4, CO2, CO, O2, H2, H2O and solid carbon (C(s)) [30,31]. The thermodynamic data necessary to perform the simulations was presented in the Table A1 in Appendix A and the data were taken in the literature [32–35]. 3.1. Model validation 3.1.1. Comparison with experimental data Fig. 1 presents a comparison between simulated and experimental data considering the simulations allowing and avoiding the C(s) formation for the CO2 reforming of methane. Fig. 1(a) presents the comparison for the CH4 conversion, the experimental data were taken from Donazzi et al. [36], O’Connor and Rossi [37] for two different catalysts and Khalesi et al. [37]. Fig. 1(b) presents the comparison between the simulated data to H2/CO molar ratio and the experimental data obtained in Khalesi et al. [37]. All simulations were performed at the same conditions at the experimental data were obtained, constant pressure of 1 atm and constant CO2/CH4 molar ratio of 1/1. In Fig. 1(a) and (b), are presented simulated curves allowing and avoiding solid carbon formation in the system. The simulation avoiding solid carbon formation in the system was obtained equaling the number of moles of solid carbon to zero during the optimization. This consideration was included in the simulations to represent the catalytic effect of solid carbon inhibition. Analyzing Fig. 1(a) it is possible to verify that the major part of the experimental data is more close to the limit established by the simulations performed avoiding the C(s) formation. In addition, the experimental data become closer to equilibrium conditions as the temperature increases. This behavior can be explained because the systems become more reactive at high temperatures. Another interesting effect to be emphasized in Fig. 1(a) is the increase in the CH4 conversion with increases in the reaction temperature. In Fig. 2 the effect of CO2/H2O molar ratio in the feed, were evaluated under the H2/CO molar ratio obtained in the products for the CO2 + H2O reforming of methane. The experimental data were obtained in Choudhary and Mondal [4]. The experiments of Choudhary and Mondal [4] were performed at the following conditions: temperature of 850 8C (1123.15 K), atmospheric pressure and with CO2/H2O molar ratios between 0.0 and 2.4/1.0.

Table 2 Conditions evaluated in the different CO2 reforming processes evaluated. Reforming technology

Temperature range (K)

Pressure range (bar)

Feed composition range

Dry reforming (DR) Dry + steam reforming (DRS) Dry autothermal reforming (DAR)

600–1600 600–1600 600–1600

1–10 bar 1–10 bar 1–10 bar

CO2/CH4 – 0.1–1.0/1.0 CO2/H2O/CH4 – 0.1–1.0/0.5–2.0/1.0 CO2/O2/CH4 – 0.1–1.0/0.1–1.0/1.0

A.C.D. Freitas, R. Guirardello / Journal of CO2 Utilization 7 (2014) 30–38

100

80

With C(s)

60 40

Without C(s)

20

(a) 0 600

800

1000

1200

1400

CH4 conversion (%)

CH4 conversion (%)

100

33

80

With C(s)

60 40

Without C(s)

20

Temperature (K)

H2 /CO molar ratio

1.4

0 773.15

With C(s)

1.2

0.6 0.4 0.2

(b)

0 850

1073.15

Fig. 3. Comparison of simulated and experimental data for the CO2 + O2 reforming of methane for the CH4 conversion of methane. Experimental data (&) from Jing et al. [39].

Without C(s)

800

973.15

Temperature (K)

1 0.8

873.15

900

950

1000

1050

1100

Temperature (K) Fig. 1. Validation with the experimental data from (a) (~)) Donazzi et al. [36], (*) O’Connor and Rossi – Pt-ZrO2 [37], (&) O’Connor and Rossi – Pt-Al2O3 [37] and (^) Khalesi et al. [37] for the CH4 conversion in the CO2 reforming of methane and (b) Khalesi et al. [37] Sr0.8; (~)) Khalesi et al. [37] Ca0.2 for H2/CO molar ratio obtained as product in the CO2 reforming of methane.

Analyzing Fig. 2 it is possible to verify that the proposed model presented a good predictive ability against the experimental data. Furthermore, it is interesting to emphasize that H2/CO molar ratio decreased with the elevation of the CO2 concentration in the feed stream, this behavior can be explained by an increase in the carbon ratio in the feed stream. Fig. 3 presents the experimental data of Jing et al. [38] and the simulations performed by the present work for the CH4 conversion in the CO2 + O2 reforming of methane. The simulations were performed at the same conditions of the experiments of Jing et al. [38], atmospheric pressure, CH4/CO2/O2 molar ratio of 1.0/0.4/0.3 and temperatures between 773.15 K and 1073.15 K. Analyzing Fig. 3 it is possible to verify that the experimental data presented a behavior between the predicted by the simulations enabling and avoiding the coke formation. With the increase in the operation temperature it is possible to verify that the behavior simulated by the two curves tend to become equal, this behavior can

be explained by the fact of the coke formation become to be controlled by thermal effects and not by catalytic effects. In a general way, it is possible to verify that the proposed model presented a good predictive ability to represent the behavior of the CO2, CO2 + H2O and CO2 + O2 reforming of methane. It is interesting to emphasize that most of the observed deviations between simulated and experimental data can be explained by the uncertainties related to the experimental data from the literature. 3.1.2. Comparison with simulated data The model was validated with simulated data obtained in the literature to verify the validity of the proposed model in comparison with similar models obtained in the literature. The simulated data was obtained in papers from literature that used Gibbs energy minimization method, for CO2 reforming the simulated data was obtained in the works of Avila-Neto et al. [31] and Akpan et al. [39], for CO2 + H2O reforming in the work of Aydinoglu [40] and for CO2 + O2 reforming in the work of Amin and Yaw [41]. In Fig. 4 the comparison between the simulated data of AvilaNeto et al. [31] and Akpan et al. [39], the data were simulated at the same conditions of these work; temperatures between 673 and 1073 K; atmospheric pressure and with a CH4/CO2/N2 of 2.0/2.0/ 1.0. It is interesting to emphasize that the papers of Avila-Neto et al. [31] and Akpan et al. [39] performed the simulations without considering the formation of coke in the system. Analyzing Fig. 4 it is possible to verify that the simulations performed by the present

4.0 H2/CO molar ratio

CH4 conversion (%)

H2/CO molar ratio

3.5 3.0 2.5 2.0 1.5 1.0 0.5

100 90 80 70 60 50 40 30 20 10 0

With C(s)

Without C(s)

673

0.0 0

0.5

1

1.5

2

2.5

CO2/H2O molar ratio Fig. 2. Comparison between simulated and experimental data from Choudhary and Mondal [38] (^) for the H2/CO molar ratio in the CO2 + H2O reforming of methane.

773

873

973

1073

Temperature (K) Fig. 4. Comparison between the simulations performed by the present work and the simulated data obtained in A´vila-Neto et al. [31] (&) and Akpan et al. [40] (^) for the CH4 conversion in the CO2 reforming of CH4.

A.C.D. Freitas, R. Guirardello / Journal of CO2 Utilization 7 (2014) 30–38 100

(a)

80

60

CH4/CO2/H2O - 1/1/3 CH4/CO2/H2O - 1/1/2 CH4/CO2/H2O - 1/1/1 CH4/CO2/H2O - 1/1/0

40 20 0 473.15

673.15

873.15

1073.15

1273.15

1473.15

100 80

(b)

60 20

CH4/CO2/O2 - 1.0/0.8/0.1 CH4/CO2/O2 - 1.0/0.8/0.2 CH4/CO2/O2 - 1.0/1.0/0.1 CH4/CO2/O2 - 1.0/1.0/0.2

800

0

1.2 1.1 673.15

873.15

1073.15

1273.15

1473.15

Temperature (K) H2O conversion (%)

(a)

-20 -40 473.15 50

(c)

40 30

900

1000 Temperature (K)

1100

1200

900

1000 Temperature (K)

1100

1200

0.9 0.8 0.7 0.6

10

0.5 673.15

873.15

1073.15

1273.15

2.5

800

1473.15

Temperature (K) 3.0

(b)

1.0

20

0 473.15

H2/CO molar ratio

100 90 80 70 60 50 40 30 20 10 0

40

H2/CO molar ratio

CO2 conversion (%)

Temperature (K)

CH4 conversion (%)

CH4 conversion (%)

34

Fig. 6. Comparison between the simulated data of Amin and Yaw [41] and the simulations performed by the proposed model to CH4 conversion (a) and H2/CO molar ratio (b).

(d)

2.0 1.5 1.0 0.5 0.0 873.15

1073.15

1273.15

1473.15

Temperature (K) Fig. 5. Comparison between the simulated data of Aydnoglu et al. [41] and the simulations performed by the proposed methodology. (a) CH4 conversion; (b) CO2 conversion; (c) H2O conversion; (d) H2/CO molar ratio. Symbols: (~)) CH4/CO2/H2O – 1.0/1.0/0.0, (^) CH4/CO2/H2O –1.0/1.0/1.0, (*) CH4/CO2/H2O – 1.0/1.0/2.0 and (&) CH4/CO2/H2O – 1.0/1.0/3.0 to (a), (b) and (c); for (d) (~)) CH4/CO2/H2O – 1.0/ 1.0/0.0, (^) CH4/CO2/H2O – 1.0/0.67/0.33, (*) CH4/CO2/H2O – 1.0/0.5/0.5 and (&) CH4/CO2/H2O – 1.0/0.33/0.67.

work are in excellent agreement with the simulated data of both work obtained in the literature. In Fig. 5 the data simulated by Aydinoglu [40] for the CO2 + H2O reforming of methane were compared with the simulations performed using the model proposed in the present work. The results were compared to CH4 conversion (Fig. 5(a)), CO2 conversion (b), H2O conversion (c) and H2/CO molar ratio (Fig. 5(d)). The results were presented as function of the temperature and the CH4/CO2/H2O molar ratio, and all simulations were performed preventing the coke formation, as well as the paper of Aydinoglu [40]. Analyzing Fig. 5 it is possible to verify that the simulations performed by the present work, for all parameters analyzed are in agreement with the simulated data obtained in the work of Aydinoglu [40]. It is interesting to emphasize that the elevation of the temperature is observed an increase in the CH4 and CO2 conversion. The H2/CO molar ratio showed a direct relationship with the elevation of the molar ratio of water in the feed, increasing

the H2O ratio in the feed, an increase in the H2/CO molar ratio was observed. The CO2 + O2 reforming of CH4 was evaluated too, and the comparison between the simulated data obtained in Amin and Yaw [41] and the simulations performed by the present work were presented in Fig. 6. The simulations were compared at four different CH4/CO2/O2 molar compositions, including: CH4/CO2/O2 – 1/0.8/0.1 (Fig. 6(a)), CH4/CO2/O2 – 1/0.8/0.2 (Fig. 6(b)), CH4/CO2/ O2 – 1/1/0.1 (Fig. 6(c)), CH4/CO2/O2 – 1/1/0.2 (Fig. 6(d)). All simulations were performed at the same conditions of the work of Amin and Yaw [41] and the coke formation was not considered in the simulations. Analyzing Fig. 6 it is possible to verify that in all CH4/CO2/O2 molar ratios evaluated, the simulations of the present work are in good agreement with the data obtained in Amin and Yaw [41]. In a general way it is possible to verify that the proposed model presented a good accuracy against experimental and simulated data, obtained in the literature, for the CO2, CO2 + H2O and CO2 + O2 reforming of methane. 3.2. Reforming’s comparison In the following sections, the results obtained using the Gibbs energy minimization and entropy maximization method for the CO2, CO2 + H2O and CO2 + O2 of methane are presented. 3.2.1. Temperature effects In Fig. 7 a comparison between the three CO2 reforming technologies is presented, with respect to the effect of reaction temperature. The H2 and CO production and the conversion of CH4 and CO2 were evaluated in Fig. 7(a)–(d) respectively. The simulations were performed at the following conditions pressure of 1 atm, temperatures between 600 and 1600 K and CO2/

A.C.D. Freitas, R. Guirardello / Journal of CO2 Utilization 7 (2014) 30–38

3.0

(a)

Moles of H2

2.5

Table 3 H2/CO molar ratio in the CO2, CO2 + H2O and CO2 + O2 reforming of methane.

CO2+H2O

CO2

2.0 1.5

Reaction temperature (K)

H2/CO molar ratio CO2

CO2 + H2O

CO2 + O2

800 1000 1200 1400 1600

4.889 1.159 0.987 0.997 0.999

4.889 1.371 1.263 1.193 1.152

7.830 1.873 1.656 1.664 1.666

CO2+O2

1.0 0.5 0.0 800

600

1000

1200

1400

35

1600

Temperature (K) 2.5

(b)

CO2

Moles of CO

2.0

CO2+H2O

1.5

CO2+O2

1.0 0.5 0.0 600

800

1000

1200

1400

1600

1400

1600

Temperature (K)

CH4 conversion (%)

100 80

CO2+O2

(c)

CO2

60 40 CO2+H2O

20 0 600

800

1000

1200

Temperature (K)

CO2 Conversion (%)

100

(d)

80 60

CO2+O2

CO2

40 20

CO2+H2O

0 -20 -40

600

800

1000

1200

1400

1600

Temperature (K) Fig. 7. Comparison of CH4 reforming of methane with CO2, CO2 + H2O and CO2 + O2 for the (a) number of moles of H2 produced, (b) number of moles of CO produced, (c) CH4 conversion and (d) CO2 conversion.

CH4 molar ratio of 1.0/1.0 for CO2 reforming; CO2/H2O/CH4 molar ratio of 0.5/0.5/1.0 for CO2 + H2O reforming; and CO2/O2/CH4 molar ratio of 0.5/0.5/1.0 for CO2 + O2 reforming. The number of moles of H2 is presented in Fig. 7(a) for the three reforming processes evaluated as a function of the process temperature. It is important emphasize that the highest H2 production is observed in the CO2 + H2O reforming reaction, this behavior can be explained by the presence of water in the feed, since the increase in the ratio of water results in increased rate of hydrogen in the feed. Lower production of H2 was observed in the CO2 + O2 reforming of methane. Fig. 7(b) presents the production of CO in number of moles, higher CO production was observed in the CO 2 reforming of methane, mainly due to the high ratio of carbon in the feed. The addition of H 2 O or O 2 , results in reduced productions of CO, the reduction become most significant when O 2 was used.

Fig. 7(c) presents the CH4 conversion and Fig. 7(d) presents the CO2 conversion, in both cases high conversions are achieved when high temperatures are used in the reaction, for temperatures above 1300 K was observed complete conversion of CH4 for the three cases examined. The CO2 conversion presented a very similar trend but the CO2 + O2 reforming do not presented total conversion of CO2, the maximum observed was 78.8% at 1600 K. Still in Fig. 7(d), it is possible to verify an anomalous behavior for CO2 conversion at 800 K. At this temperature, there is observed a decrease in CO2 conversion, to temperatures that exceed 800 K an elevation in CO2 conversion was observed again. This behavior can be explained by the thermal characteristics of water gas shift (WGS) reaction. The equilibrium of this reaction shows a significant temperature dependence and the equilibrium constant decreases with an increase in temperature, that is, higher CO2 was observed as product in the reaction. However, at higher temperatures, the CO2 added as reagent or produced by WGS reaction becomes consumed to produce CO and H2. Table 3 presented the H2/CO molar ratio for the three reforming processes analyzed. The reactions were evaluated at the same above mentioned conditions, and the results are presented in Table 3 as function of the temperature. It is interesting to emphasize that the lower H2/CO molar ratios were observed in the CO2 reforming of methane, and ratios close to one are achieved for temperatures above 1200 K. The addition of water results in the elevation of the H2/CO molar ratio, an increase of approximately 40% was observed in all temperature range. The addition of O2 also resulted in increased H2/CO molar ratio, although this was less pronounced, a reduction around 18% was observed for temperatures above 1000 K. 3.2.2. Pressure effects In Fig. 8 the effect of pressure was presented for the CH4 conversion, H2 and syngas production for three pressures, 1, 5 and 10 atm. The simulations were performed at the following conditions: constant temperature of 1200 K and constants feed molar ratios of: CO2/CH4 – 1.0/1.0; CO2/H2O/CH4 – 0.5/0.5/1.0 and CO2/O2/CH4 – 0.5/0.5/1.0. Analyzing Fig. 8(a) it is possible to verify that the elevation of the reaction pressure results in the reduction of the CH4 conversion in all reforming process analyzed, the greatest reduction was observed for the CO2 + H2O reforming and the observed reduction was 18%. Another interesting effect is that the addition of H2O resulted in a reduction in the CH4 conversion when compared with the CO2 reforming process. On the other hand, the addition of O2 resulted in the elevation of the CH4 conversion. In Fig. 8(b) the number of moles of H2 produced in the three reforming process analyzed are presented. The elevation of the pressure resulted in a reduction in the number of moles of H2 in all reforming processes analyzed. Similar behavior was observed in the partial oxidation of methane [20].

36

A.C.D. Freitas, R. Guirardello / Journal of CO2 Utilization 7 (2014) 30–38

Fig. 9. Moles of C(s) produced in the CO2, CO2 + H2O and CO2 + O2 reforming of CH4 as function of temperature (a) and pressure (b).

In Fig. 8(c) the number of moles of syngas (H2 + CO) produced in the three reforming process analyzed are presented. It is important to emphasize that syngas production decreased with increasing pressure, but the H2/CO molar ratio observed in the products (see Fig. 8(d)) do not show significant modifications with the variation of the pressure.

analyzed separately on coke formation in the CO2, CO2 + H2O and CO2 + O2 reforming reactions. In Fig. 9 the formation of C(s) are presented as function of the temperature (Fig. 9(a)) and as function of the pressure (Fig. 9(b)). In Fig. 9(a) the simulations were performed at constant pressure of 1 atm and in the following compositions: CO2/CH4 – 1.0/1.0 for the CO2 reforming, CO2/H2O/CH4 – 0.5/0.5/1.0 for the CO2 + H2O reforming and CO2/O2/CH4 – 0.5/0.5/1.0 for the CO2 + O2 reforming and with temperatures between 600 and 1200 K. In Fig. 9(b) the simulations were performed at the following conditions: constant temperature of 1200 K, pressures of 1, 5 and 10 atm and at the same composition range utilized in Fig. 9(a). It is interesting to emphasize that the addition of both, H2O and O2, resulted in the reduction of C(s) formation and this effect can be seen in Fig. 9(a) and (b). The elevation of the temperature resulted in the complete elimination of C(s) formation to temperatures greater than 1200 K in all reforming analyzed, this effect can be seen in Fig. 9(a). The elevation of the pressure presented a negative effect for the system, because higher pressures favor the formation of C(s) this effect was observed on the three reforming process analyzed, and can be seen in Fig. 9(b). It is interesting to emphasize that in the CO2 + O2 reforming no coke formation was observed at 1 atm, but at 5 and 10 atm significant amounts of coke were formed. The simulation considering air as oxidant agent presented a very similar behavior to that observed when O2 was used, this behavior can be explained due the fact that at isothermic conditions the N2 form air act as an inert in the reaction, thus not affecting the composition of the products.

3.2.3. Solid carbon formation The coke formation is reported as one of the major drawbacks to application of the CO2 reforming processes in large scale. In this part of the paper the effects of changes in operating variables were

3.2.4. Equilibrium temperatures Fig. 10 presents the results obtained using the entropy maximization model to predict the final temperatures in the CO2, CO2 + H2O, CO2 + O2 and CO2 + air reforming reactions. The

Fig. 8. Effect of the pressure under the (a) CH4 conversion, (b) H2 production, (c) syngas production and (d) H2/CO molar ratio for the CO2, CO2 + H2O and CO2 + O2 reforming’s of CH4.

A.C.D. Freitas, R. Guirardello / Journal of CO2 Utilization 7 (2014) 30–38

37

This behavior can be explained by the presence of the nitrogen in the air. Similar characteristics for the air use were observed in the partial oxidation of methane [20]. The CO2 reforming and CO2 + H2O reforming of methane presented a strongly endothermic behavior and it is important to emphasize that the addition of water further emphasized the system endothermicity. In Fig. 10(b) the effect of the pressure under the thermal behavior of the systems was presented. The simulations were performed at constant initial temperature of 1200 K and at the same molar ratios in the feed of Fig. 10(a). Three pressures were studied 1, 5 and 10 atm. For all reforming process analyzed the elevation of the pressure resulted in the elevation of the equilibrium temperatures. In Fig. 10(b) it can be seen clearly that the reforming reactions with O2 or air presented exothermic behavior (final temperatures higher than initial temperatures) and the CO2 and CO2 + H2O reforming presented a highly endothermic behavior (final temperatures below the initial temperatures).

4. Conclusion

Fig. 10. Equilibrium temperatures for the CO2, CO2 + H2O, CO2 + O2 and CO2 + air reforming reactions as function of (a) initial temperatures and (b) pressure at constant initial temperature of 1000 K.

use of air was considered by the inclusion of N2 in the simulations the air was simulated like having a fixed composition (80% of N2 and 20% of O2). In Fig. 10(a) the effect of the initial temperature are presented. The simulations were performed at constant pressure (1 atm) and at the following molar ratios in the feed: CO2/CH4 – 1.0/1.0 for the CO2 reforming; CO2/H2O/CH4 – 0.5/0.5/1.0 for the CO2 + H2O reforming; CO2/O2/CH4 – 0.5/0.5/1.0 for the CO2 + O2 reforming and CO2/O2/N2/CH4 – 0.5/0.5/2.0/1.0 for the CO2 + air reforming. The elevation of the initial temperature resulted in the elevation of the equilibrium temperatures in all reforming processes analyzed, the CO2 + O2 and CO2 + air presented exothermic behavior in all temperature range analyzed, and the use of air proved to be efficient to control the exothermicity of the reaction.

The calculated results were compared with previously published experimental and simulated data with a good agreement between them for all systems. The H2 and syngas production were favored at high temperature and low pressure conditions. The addition of H2O or O2 proved to be an effective way to reduce the coke formation in the systems. The CO2 reforming presented endothermic behavior, but the addition of O2 or air reduced this trend and in some conditions autothermal behavior was observed. The addition of H2O increased the endothermic behavior of the CO2 reforming reaction. The GAMS123.1 software and the CONOPT solver were used in the resolution of the proposed problems and all calculations performed presented a low computational time (less than 1 s). Acknowledgements The authors gratefully acknowledge the financial support from CAPES – Coordenac¸a˜o de Aperfeic¸oamento de Pessoal de Nı´vel Superior, FAPESP – Fundac¸a˜o de amparo a` pesquisa de estado de Sa˜o Paulo (Process 2011/20666-8) and CNPq – Conselho Nacional de Desenvolvimento Cientı´fico e Tecnolo´gico, Brazil. Appendix A. Thermodynamic data Table A1

Table A1 Thermodynamic data for the components considered in the simulations. Component

Cpai

103 Cpbi

106 Cpci

105.Cpdi

Tmax

DH0f (J/mol)

DH0f (J/mol)

CH4 O2 CO2 CO H2O H2

1.702 3.639 5.457 3.376 3.470 3.249

9.081 0.506 1.045 0.557 1.450 0.422

2.164 – – – – –

– 0.227 1.157 0.031 0.121 0.083

1500 2000 2000 2500 2000 3000

74,520 0 393,509 110,525 241,818 0

50,460 0 394,359 137,169 228,572 0

Tmax, temperature limit of C pgi expression. C pgi =R ¼ Ai þ Bi :T þ C i :T 2 þ Di :T 2 :

38

A.C.D. Freitas, R. Guirardello / Journal of CO2 Utilization 7 (2014) 30–38

References [1] S. Ayabe, H. Omoto, T. Utaka, R. Kikuchi, K. Sasaki, Y. Teraoka, K. Eguchi, Appl. Catal. A: Gen. 241 (2003) 261–269. [2] J.A.C. Dias, J.M. Assaf, J. Power Sources 137 (2004) 264–268. [3] T. Rostrup-Nielsen, Catal. Today 106 (2005) 293–296. [4] V.R. Choudhary, K.C. Mondal, Appl. Energy 83 (2006) 1024–1032. [5] M.C.J. Bradford, M.A. Vannice, J. Catal. 173 (1998) 157–171. [6] M.M.V.M. Souza, D.A.G. Aranda, M. Schmal, J. Catal. 204 (2001) 498–511. [7] S.M. Stagg-Williams, F.B. Noronha, G. Fendley, D.E. Resasco, J. Catal. 194 (2000) 240–249. [8] A. Djaidja, S. Libs, A. Keinnemann, A. Barama, Catal. Today 113 (2006) 194–200. [9] K. Nagaoka, K. Seshan, K. Aika, J.A. Lercher, J. Catal. 197 (2001) 34–42. [10] V.R. Choudhary, A.M. Rajput, Ind. Eng. Chem. Res. 35 (1996) 3934–3939. [11] M.E.S. Hegarty, A.M. O’Connor, J.R.H. Ross, Catal. Today 42 (1998) 225–232. [12] F.B. Noronha, A. Shamsi, C. Taylor, E.C. Fendley, S. Stagg-William, D.E. Resasco, Catal. Lett. 90 (2003) 13–21. [13] P.D.F. Vernon, M.L.H. Green, A.K. Cheetham, A.T. Ashcroft, Catal. Today 13 (1992) 417–426. [14] V.R. Choudhary, A.M. Rajput, B. Prabhakar, Catal. Lett. 32 (1995) 391–396. [15] V.R. Choudhary, A.M. Rajput, B. Prabhakar, Angew. Chem. Int. (Engl. Ed.) 33 (1994) 2104–2106. [16] V.R. Choudhary, B.S. Uphade, A.S. Mamman, Appl. Catal. A: Gen. 168 (1998) 33– 46. [17] A.M. O’Connor, J.R.H. Ross, Catal. Today 46 (1998) 203–210. [18] D. Castello, L. Fiori, Bioresour. Technol. 102 (2011) 7574–7582. [19] J. Castillo, I.E. Grossmann, Comput. Chem. Eng. 5 (1981) 99–108. [20] A.C.D. Freitas, R. Guirardello, J. Nat. Gas Chem. 21 (2012) 571–580. [21] Y. Lu, X. Guo, X. Zhang, Q. Yan, Chem. Eng. J. 131 (2007) 233–244.

[22] D.V. Nichita, S. Gomez, E. Luna, Comput. Chem. Eng. 26 (2002) 1703–1724. [23] C.C.R.S. Rossi, M.E. Berezuk, L. Cardozo-Filho, R. Guirardello, Comput. Chem. Eng. 35 (2011) 1226–1237. [24] A.C.D. Freitas, R. Guirardello, Chem. Eng. Trans. 27 (2012) 361–366. [25] W.R.M. Smith, R.W. Missen, Chemical Reaction Equilibrium Analysis: Theory and Algorithms, Original Edition by John Wiley & Sons, New York, 1982. [26] C. McDonald, C.A. Floudas, Ind. Eng. Chem. Res. 34 (1995) 1674–1687. [27] K. Brooke, D. Meeraus, R. Raman, GAMS – A User’s Manual, 1998. [28] A.C.D. Freitas, R. Guirardello, Chem. Eng. Trans. 32 (2013) 553–558. [29] M. Castier, Fluid Phase Equilib. 276 (2009) 7–17. [30] S. Rakass, H. Oudghiri-Hassani, P. Rowntree, N. Abatzoglou, J. Power Sources 158 (1) (2006) 485–496. [31] C.N. Avila-Neto, S.C. Dantas, F.A. Silva, T.V. Franco, L.L. Romanielo, C.E. Hori, A.J. Assis, J. Nat. Gas Sci. Eng. 1 (2009) 205–215. [32] B.E. Poling, J.M. Prausnitz, P.J. O’Connell, The Properties of Gases and Liquids, 5th ed., McGraw Hill, New York, 2000. [33] R.C. Reid, J.M. Prausnitz, B.E. Poling, The Properties of Gases and Liquids, 4th ed., McGraw Hill, New York, 1987. [34] R.C. Reid, J.M. Prausnitz, B.E. Poling, The Properties of Gases and Liquids, 3rd ed., McGraw Hill, New York, 1977. [35] DIADEM Public v.1.2C DIPPR Design Institute for Physical Property Data. Information and Data Evaluation Manager, 2000. [36] A. Donazzi, A. Beretta, G. Groppi, P. Forzatti, J. Catal. 255 (2008) 259–268. [37] A. Khalesi, H.R. Arandiyan, M. Parvari, Chin. J. Catal. 29 (2008) 960–968. [38] Q. Jing, H. Lou, L. Mo, J. Fei, X. Zheng, J. Mol. Catal. A: Chem. 212 (2004) 211–217. [39] E. Akpan, Y. Sun, P. Kumar, H. Ibrahim, A. Aboudheir, R. Idem, Chem. Eng. Sci. 62 (2007) 4012–4024. [40] S.O. Aydinoglu, Int. J. Hydrog. Energy 35 (2010) 12821–12828. [41] N.A.S. Amin, T.C. Yaw, Int. J. Hydrog. Energy 32 (2007) 1789–1798.