Kinetic-Operational Mechanism to Autothermal Reforming of Methane

Aug 9, 2010 - Nelson M. Lima Filho, and Cesar A. M. Abreu. DEQ, CTG, UFPE, Av. Prof. Artur de Sá, s/n° - Cidade Universitária, Recife-PE - 50740-52...
0 downloads 0 Views 5MB Size
ARTICLE pubs.acs.org/IECR

Kinetic-Operational Mechanism to Autothermal Reforming of Methane Aleksandros E. A. M. Souza,* Leonardo J. L. Maciel, Valderio O. Cavalcanti-Filho, Nelson M. Lima Filho, and Cesar A. M. Abreu DEQ, CTG, UFPE, Av. Prof. Artur de Sa, s/n - Cidade Universitaria, Recife-PE - 50740-521, Brazil ABSTRACT: A reforming experiment was performed using a nickel catalyst, with the goal of suggesting a mechanism for the autothermal reforming of methane (ATR), under thermally neutral conditions. To bring the process to future scale-up at a fixed-bed reactor, kinetic-operational evaluations were made while taking into account the parameters of the gas phase flow and the temperature. From these evaluations, based on conversions, yields, and selectivities, it was possible to develop a novel approach to a descriptive mathematical model of the process behavior. A thermodynamic evaluation of the ATR was performed to measure the effects of operational conditions (temperature, pressure, and composition of the feed), in relation to the established limit values of the chemical equilibrium.

’ INTRODUCTION Natural gas (NG), having a higher content of methane (66.0-97.8%, v/v) in its composition, can be processed in the presence of steam and oxygen via autothermal catalytic reforming, producing hydrogen and synthesis gas. Syngas from NG, a mixture of carbon monoxide and hydrogen, may be used to produce high-value chemicals such as hydrocarbons, fuels,1 and oxygenated compounds.2 In gas-to-liquids (GTL) processes, where natural gas is first converted to syngas, 60-70% of the costs of the overall process are associated with the syngas production. Nickel catalysts, traditionally used in industrial steam reforming processes of NG, have shown good performance in terms of conversion and selectivity to obtain syngas by methane reforming, although these systems are sensitive to coke formation. In this work, following the classical methodologies of catalysis and of chemical reaction engineering, a nickel catalyst was prepared, characterizing it and using it in the development of the methane ATR process, with the objective of using it in the transformation of the NG with the goal of producing syngas and/or hydrogen. For this reason, the thermodynamics of the process reactions were evaluated, while establishing the operational conditions for this to occur. In consideration of future scale-ups at a fixed-bed reactor, kinetic-operational evaluations were made in terms of the parameters of the gas phase flow and the temperature. The foundations obtained through these evaluations, in terms of conversions, yields, and selectivities, were useful to elaborate a descriptive mathematical model of the process behavior, whose validations were able to lead to simulations of the profiles of the components’ concentration in the reaction environment. Reforming Processes Technologies. NG is a designation given to a mixture of light hydrocarbons stored in porous formations of the subsoil, whose composition varies depending on the region where it is produced. It is a colorless, odor-free, and nontoxic product which remains in a gaseous state under normal conditions of surrounding temperature and atmospheric pressure, r 2010 American Chemical Society

being basically composed of methane. Synthesis gas (syngas), a mixture of hydrogen and carbon monoxide, which is fundamental to the chemical industry of synthesis, can be produced from it. Nowadays, there are at least six technologies available to produce syngas:1 the noncatalytic partial oxidation of methane (NCPOM or POX), catalytic partial oxidation of methane (POM), dry reforming of methane (DRM), conventional steam reforming of methane (SRM), autothermal reforming of methane (ATR), and combined reforming of methane (CRM). There are several catalysts used in the reforming of natural gas, just as with the reforming of alcohol and ether. The great majority of them use a group VIII metal and a supporting oxide to function as an acid site. Due to the costs involved,2-5 industry prefers to use nickel as the active phase, supported in alumina as the passive phase.

’ THEORETICAL CALCULATIONS Thermodynamics and Kinetics of Reforming Processes. ATR is by definition a combination of the conventional SRM and the NC-POM or POX,1,2 under thermally neutral conditions, taking into consideration the heat loss due to the surroundings. In this process, a mixture of natural gas, oxygen, and water is first submitted to a noncatalytic partial oxidation. Next, the products from this first step are passed through a nickel catalyst, completing the reforming reaction and obtaining syngas. This process creates problems in the formation and accumulation of coke, which can be minimized through the introduction of steam.1 Thermodynamics of the Process Reactions. Thermodynamic evaluation of the ATR is necessary to measure the effects of Special Issue: IMCCRE 2010 Received: March 15, 2010 Accepted: July 22, 2010 Revised: July 21, 2010 Published: August 9, 2010 2585

dx.doi.org/10.1021/ie100637b | Ind. Eng. Chem. Res. 2011, 50, 2585–2599

Industrial & Engineering Chemistry Research

ARTICLE

reaction.6

operational conditions (temperature, pressure, and composition of the feed), relating them to the established limit values of chemical equilibrium. To determine the composition of the feed, with the possibility of establishing chemical equilibrium, the minimization conditions of Gibbs energy can be used, whose solutions to thermodynamic equations of minimization use the Lagrange multipliers, λk, where k is each element (C, H, O) of the

ΔGf i

  X P þ RT ln yi ji λk aik ¼ 0 þ P k X

ð1Þ

ni aik - Ak ¼ 0

ð2Þ

i

Table 1. Possible Reactions in the ATR and Equilibrium Constants at 1023, 1073, and 1123 K Gibbs energy7,8a

reaction

K1023K

K1073K

K1123K

CH4 a C þ 2H2

ð3Þ

ΔG ¼ þ 74:52 - 0:081T kJ = mol

T > 920 K

13.12

21.42

33.54

1 CH4 þ O2 a C þ H2 þ H2 O 2 1 CH4 þ O2 a CO þ 2H2 2

ð4Þ

ΔG ¼ - 167:29 - 0:037T kJ = mol

"T

7.7  1010

3.2  1010

1.5  1010

ð5Þ

ΔG ¼ - 36:00 - 0:170T kJ = mol

"T

2.8  1011

2.5  1011

2.2  1011

CH4 þ O2 a C þ 2H2 O

ð6Þ

ΔG ¼ - 409:12 þ 0:008T kJ = mol

T < 51 140 K

4.5  1020

4.8  1019

6.3  1018

CH4 þ O2 a CO þ H2 þ H2 O

ð7Þ

ΔG ¼ - 277:82 - 0:126T kJ = mol

"T

1.6  1021

3.7  1020

9.5  1019

CH4 þ O2 a CO2 þ 2H2 3 CH4 þ O2 a CO þ 2H2 O 2 3 CH4 þ O2 a CO2 þ H2 þ H2 O 2

ð8Þ

ΔG ¼ - 318:99 - 0:084T kJ = mol

"T

2.2  1021

4.1  1020

8.9  1019

ð9Þ

ΔG ¼ - 519:64 - 0:081T kJ = mol

"T

9.5  1030

5.5  1029

4.1  1028

ð10Þ

ΔG ¼ - 560:81 - 0:039T kJ = mol

"T

1.3  1031

6.1  1029

3.9  1028

CH4 þ 2O2 a CO2 þ 2H2 O

ð11Þ

ΔG ¼ - 802:62 þ 0:005T kJ = mol

T < 160526 K

7.4  1040

9.1  1038

1.7  1037

CH4 þ H2 O a CO þ 3H2

ð12Þ

ΔG ¼ þ 205:81 - 0:215T kJ = mol

T > 957 K

47.67

164.16

507.59

CH4 þ 2H2 O a CO2 þ 4H2

ð13Þ

ΔG ¼ þ 164:65 - 0:173T kJ = mol

T > 952 K

63.41

181.10

473.37

CH4 þ CO2 a 2C þ 2H2 O

ð14Þ

ΔG ¼ - 15:61 þ 0:011T kJ = mol

T < 1419 K

2.72

2.53

2.38

CH4 þ CO2 a 2CO þ 2H2

ð15Þ

ΔG ¼ þ 246:98 - 0:257T kJ = mol

T > 961 K

35.84

148.81

544.28

CH4 þ 2CO2 a C þ 2CO þ 2H2 O

ð16Þ

ΔG ¼ þ 156:85 - 0:165T kJ = mol

T > 951 K

7.42

17.60

38.57

CH4 þ 2CO2 a 3CO þ H2 þ H2 O

ð17Þ

ΔG ¼ þ 288:14 - 0:299T kJ = mol

T > 964 K

26.94

134.89

583.63

CH4 þ 3CO2 a 4CO þ 2H2 O

ð18Þ

ΔG ¼ þ 329:31 - 0:341T kJ = mol

T > 966 K

20.25

122.27

625.82

CH4 þ CO a 2C þ H2 þ H2 O

ð19Þ

ΔG ¼ - 56:77 þ 0:053T kJ = mol

T < 1071 K

3.61

2.79

2.22

CH4 þ 2CO a 3C þ 2H2 O

ð20Þ

ΔG ¼ - 188:07 þ 0:186T kJ = mol

T < 1011 K

0.99

0.36

0.15

CH4 þ 2CO a 2C þ CO2 þ 2H2

ð21Þ

ΔG ¼ - 97:94 þ 0:095T kJ = mol

T < 1031 K

4.80

3.08

2.07

CH4 þ 3CO a 3C þ CO2 þ H2 þ H2 O ð22Þ

ΔG ¼ - 229:23 þ 0:228T kJ = mol

T < 1005 K

1.32

0.40

0.14

ð23Þ

ΔG ¼ - 110:52 - 0:089T kJ = mol

"T

2.1  1010

1.1  1010

6.6  109

C þ O2 a CO2

ð24Þ

ΔG ¼ - 393:51 - 0:003 kJ = mol

"T

1.6  1020

1.9  1019

2.7  1018

C þ H2 O a CO þ H2

ð25Þ

ΔG ¼ þ 131:29 - 0:134T kJ = mol

T > 980 K

3.63

7.66

15.13

C þ 2H2 O a CO2 þ 2H2

ð26Þ

ΔG ¼ þ 90:13 - 0:092T kJ = mol

T > 980 K

4.83

8.46

14.11



1 O2 a CO 2

C þ CO2 a 2CO



ð27Þ

ΔG ¼ þ 172:46 - 0:176T kJ = mol

T > 980 K

2.73

6.95

16.23

1 O2 a CO2 2

ð28Þ

ΔG ¼ - 282:98 þ 0:086T kJ = mol

T < 3290 K

7.8  109

1.7  109

4.0  108

CO þ H2 O a CO2 þ H2 1 H2 O a H2 þ O2 2

ð29Þ

ΔG ¼ - 41:17 þ 0:042T kJ = mol

T < 980 K

1.33

1.10

0.93

ð30Þ

ΔG ¼ þ 241:82 - 0:044T kJ = mol

T > 5496 K

1.7  10-10

6.7  10-10

2.3  10-9

CO þ

a

spontaneity

Selected data based on Kelley,9 Takahashi and Westrum,10 Spencer,11 Pankratz,12 Wagman et al.,13 and Chase.14 2586

dx.doi.org/10.1021/ie100637b |Ind. Eng. Chem. Res. 2011, 50, 2585–2599

Industrial & Engineering Chemistry Research

ARTICLE

Equation 1 represents the N equations for equilibrium, one for each chemical species i, and eq 2 represents w equations representing mass balances, one for each element k. The total number of equations, N þ w, is sufficient to determine all the P unknowns of these equations, which are the ni0 s for eq 2 (yi = ni/ i ni) and λk0 s for eq 1. Reaction Mechanism. The designation of the reaction mechanism of the ATR process comes from the evaluation of all the reactions theoretically possible, whose verification is derived from the definition of the operational conditions (spontaneity; Table 1). Mathematical Modeling of Reactive Evolutions. With the objective of describing the reactional behavior of the process operated on a fixed-bed reactor, a mass balance was developed for each component under steady state conditions. For a component of the process i, with molar concentration Ci (mol/m3), operating with reaction rate Ri (mol 3 kg-1 cat 3 s-1), the balance equation is as follows: dCi þ Ri ¼ 0 ð31Þ dτ With the fluid-solid contact spatial time, τ (kg cat 3 s/m3), a system of equations was formed for the reagents and products involved in the operation.

’ MATERIALS AND METHODS Catalyst Preparation. The preparation method utilized incipient wetness where alumina was impregnated with the nickel nitrate solution (Ni(NO3)2 3 6H2O, Sigma), as the precursor, over γ-alumina (γ-Al2O3, Degussa) as a catalytic support. First, the impregnated solution was evaporated to dryness. Then, the solid was dried at 393 K for 12 h and calcinated at 873 K in an air flow over 5 h. Finally, to promote the activation, the material was reduced in a hydrogen atmosphere at 1073 K for 2 h. Catalyst Characterization. The nickel catalyst was characterized by atomic absorption spectroscopy (AAS; Varian AA 220 FS), BET-N2 (Micromeritics ASAP 2010), X-ray diffraction (XRD; Siemens D5000), infrared spectrometry (FTIR; Bruker IFS 66), scanning electron microscopy (SEM; Jeol JSM-6360), thermal analyses (TG, DTG, and DTA; Shimadzu TGA-50), and elementary carbon analysis. Catalyst Evaluations. The autothermal reforming experiments were carried out in a fixed-bed reactor (Idelglass; overall length, Lr = 64.0 cm; length of bed useful, Lb = 12.0 cm; inner diameter, dr = 1.65 cm; pore diameter of sintered plate in the bed, ds = 40-100 μm; inner diameter of the flow of incoming and outgoing, dio; and inner diameter of the pit to the thermocouple, dt = 6.0 mm) and packed with a nickel catalyst (dp = 212 μm, m = 60 mg) under atmospheric pressure and at three different temperatures, 1023, 1073, and 1123 K.15 The reactants were fed into the reactor with a gas mixture with a molar ratio of 5:1:9 CH4/O2/H2O and diluted in argon at 56% (v/v) with a flow rate in the range of 150-400 cm3/min (STP). The reactor effluents, residual reagents, and products were analyzed online by gas chromatograph (GC, Saturn 1200, Varian; TCD, Carbosphere/Porapak-Q). Operational Conditions to Conduct the Process. Aiming toward a maximization of hydrogen production and avoiding carbon deposition, the catalyst systems were evaluated through ATR, eq 32, establishing conditions under a stoichiometric analysis.

CH4 þ aO2 þ bH2 O f cCO þ dCO2 þ eH2

ð32Þ

Figure 1. Stoichiometric analysis of the ATR to maximize H2 production. Identification of the conditions required to not produce coke.16

The stoichiometric adjustment leads to the following relationships: b = e - 2, c = 2 - 2a - b, and d = 2a þ b - 1. To maximize the H2 production (c = 0), d = 1 and a = 1 - b/2 or b = 2 - 2a. For a g 0 f b e 2, a straight line describes the possibilities of the composition of the feed to ATR (Figure 1). In both highlighted lines in Figure 1, the carbon deposits should be null.16 Thus, to avoid coke deposition, the feed was carried out at a molar ratio of 5:1:9 CH4/O2/H2O, in the avoidable coking region of the SRM. On the basis of the molar ratio of the feed, a condition was obtained, eq 33, adopting a stoichiometric steam excess as indicated:17 CH4 þ 0:2O2 þ 1:8H2 O f CO2 þ 3:6H2 þ 0:2H2 O ΔH ¼ þ 109 kJ=mol ð33Þ 18-20

This condition overcame the athermicity of the reaction (exothermicity), producing a sufficient quantity of energy to reach the autothermicity. The operated GHSV range (defined below) suggests there are no limitations to external mass transfer (Ma et al.).21

’ RESULTS AND DISCUSSION Catalyst Characterization. The nickel content and its surface area, characterized by AAS and BET-N2, were 5.75% by weight and 107 m2/g, respectively. Through textural analyses (Table 2), reductions of specific surface area during the catalyst preparation up to the calcination step were observed, while pore volume was almost constant. This can be explained by the metal deposition on the support. During the reduction step, the micropore volume increased, indicating NiO conversion into nickel. X-ray diffractograms of the catalyst preparation steps allowed identification of the solid phases. From the crystallographic patterns of alumina, it was possible to identify a γ-alumina phase on the support in natura and on the support submitted to the thermal pretreatment. There was no evidence of a γ-alumina transition to δ-alumina. XRD analysis of the catalyst, performed after the calcination step, identified nickel oxide as well as nickel aluminates NiAl2O4 (2θ = 37.08 and 59.78)22 and NiAl10O16 (2θ = 19.18, 45.58, and 66.88). After reduction step, the analysis also identified the presence of reduced nickel (2θ = 44.58 and 51.88; Figure 2). 2587

dx.doi.org/10.1021/ie100637b |Ind. Eng. Chem. Res. 2011, 50, 2585–2599

Industrial & Engineering Chemistry Research

ARTICLE

Table 2. BET Analysis of Precatalytic Material and Catalyst

a

material

BET area (m2/g)

pore diameter (Å)

pore volume (cm3/g)

micropore volume (cm3/g)

alumina in natura

226

68

0.4

n.d.a

pretreated alumina

145

102

0.4

0.0006

calcinated catalyst

105

125

0.3

0.0026

reduced catalyst

107

129

0.3

0.0034

n.d. = not determinated.

Figure 2. Diffractograms of the precatalytic material and catalyst. Identification of the phases.

Figure 3. FTIR spectra of the solid phases of the nickel catalyst preparation steps. Catalyst: Ni(5.75 wt %)/γ-Al2O3.

The IR spectra of catalyst Ni(5.75 wt%)/γ-Al2O3 preparation steps are shown in Figure 3, where it is possible to identify the presence of crystallization water and hydrogen bonds (OH-; 3451 and 1637 cm-1). The quantity of crystallization water (peak intensity) decreases as the catalyst reaches its final preparation. The presence of -NO2 groups (1631 cm-1) and nitrate anions (1383 and 827 cm-1) are perceptible, mainly on the impregnated support. Figure 4 presents thermal analyses (TG, DTG, and DTA) of the catalyst during the preparation steps. The TG plot exhibited the initial calcination temperature of the precursor. The DTG profile presented a peak at 346.9 K, indicating a loss of water of 13.0% between 298.7 and 427.9 K. A second peak was observed at 537.3 K, indicating a loss of NO2 of 10.0%, between 427.9 and 814.5 K. There is also an indication that both events overlap in

Figure 4. Thermal analysis of the catalyst during the preparation steps.

Figure 5. Thermogravimetric (TG) and derivate thermogravimetric (DTG) analysis of the nickel catalyst. Determination of mass loss of deposited carbon in ATR. Operating conditions: 1123 K, 1.00 bar, total feed flow rate of 400 cm3/min, feeding molar ratio 5:1:9 CH4/O2/H2O, Ni(5.75 wt %)/γ-Al2O3.

the range of 298.7-814.5 K. Losses progressed to 1273 K possibly due to NOx emissions. The carbon deposited during the reaction process was obtained from thermal analyses (TG and DTG, Figure 5) of the catalyst. They showed a carbon mass loss of 4.2% between 430.7 and 1081.5 K. Higher carbon deposit levels on the catalyst (>70% in weight) were found under drastic operating conditions.23,24 A mass loss of water of 7.2% occurred in the range 300.1-430.7 K (endothermic peak). SEM analyses on the catalyst were performed before and after the reforming operations and allowed the identification of coking formation as a whisker (filamentous carbon; Figure 6). 2588

dx.doi.org/10.1021/ie100637b |Ind. Eng. Chem. Res. 2011, 50, 2585–2599

Industrial & Engineering Chemistry Research

ARTICLE

Figure 6. Electronic micrographies of Ni(5.75 wt %)/γ-Al2O3 catalyst. (a) Before autothermal reforming (400) and (b) after autothermal reforming (2500).

Table 3. Equilibrium Compositions 1023 K

1073 K

1123 K

yCH4

0.33%

0.10%

0.03%

yO2 yH2O

0.00% 19.30%

0.00% 19.61%

0.00% 20.05%

yH2

59.73%

59.51%

59.11%

yCO

14.44%

15.24%

15.82%

yCO2

6.20%

5.54%

N (mols)

23.84

23.95

5.00% 23.98

Evaluation of the Autothermal Reforming Processes. The ATR was evaluated in terms of methane conversion, product yield and selectivity, and the molar ratio of H2/CO production. Determination of the Operation Temperature Range. By evaluating the values of the Gibbs energy of every possible reaction of the ATR process (Table 1), it was possible to foresee the following: The methane cracking occurs above 920 K (eq 3). All of the methane oxidations are possible (eqs 4-11) at any temperature. All of the steam reformings (eqs 12 and 13) become possible above 957 K. To make dry reforming possible, CO2 reforming (eqs 14-18), decreasing the CO2 quantity as the final product in ATR, must work between 966 and 1419 K. Above 1005 K, the reactions between methane and carbon monoxide (eqs 19-22) are preferred, provoking the formation of coke, which is why working at higher temperatures is recommended. The partial regeneration of the catalyst through coke elimination (eqs 23-27) is made much easier when performed above 980 K. The CO2 disproportionalization easily occurs at any temperature (eq 28). The water-gas shift reaction (WGSR) takes place below 980 K, consuming carbon monoxide but producing hydrogen (eq 29). The water does not decompose except above 5499 K (eq 30). Therefore, an operating temperature range between 1005 and 1419 K is strongly recommended. Under these conditions, methane cracking and combustion, dry reformings, and coke elimination will be favored. So, operation temperatures were 1023, 1073, and 1123 K. Equilibrium constants (K) of possible reactions on ATR, at the above temperatures, are shown in Table 1. Composition at Thermodynamic Equilibrium. In Table 3, according to eqs 1 and 2, equilibrium compositions at the operational temperatures are presented, considering gaseous chemical species in the ATR and composition of the feed.

Catalyst Evaluations. Experimental evaluations of the ATR process, performed in a fixed-bed continuous reactor, as a function of the operational time, showed that steady states of concentration were reached after 100 min of operation. Under the same pressure conditions (1.00 bar) and feed compositions (molar ratio 5:1:9 CH4/O2/H2O), for each temperature, flow rates in the range of 150-400 cm3/min in the presence of the nickel catalyst (60 mg, Ni(5.75 wt %)/γ-Al2O3) were employed. As a function of the utilized flow rates, the process was operated with different contact spatial times (9-24 kg 3 s/m3, or GHSV = 331.5  103 to 884.0  103 h-1). The contact spatial time (τ) is defined as the ratio of the mass of the catalyst to the volumetric flow of reactants at standard conditions (298 K and 1.00 bar): ! mcat ð34Þ τ¼ Q_ inlet STP

where mcat is the mass of the catalyst, in kilograms, and Q_ inlet is the volumetric flow rate of the reactants, in cubic meters per hour. The gas hours space velocity (GHSV) is defined as the ratio of the volumetric flow of reactants at standard conditions (298 K and 1.00 bar) to the total catalyst volume and has units of inverse time, or ! Fcat  Q_ inlet ð35Þ GHSV ¼ mcat STP

where Fcat is the density of the catalyst, in kilograms per cubic meter, and Q_ inlet and mcat have already been defined. Comparison of Thermodynamic Predictions. The ATR process was evaluated through operations at 1023, 1073, and 1123 K, under 1.00 bar, studying effects of fluid-solid contact spatial time, relating them to the equilibrium compositions. In Figure 7, it is possible observe that, with the temperature increase, the molar ratio of methane on effluents tends to decrease in the direction of the thermodynamically foreseen equilibrium composition. It is also possible to observe that, with the increase of fluid-solid contact spatial time, to any of the temperatures studied, the methane composition in a steady state approaches the thermodynamic composition of equilibrium without actually reaching it. The evolution of composition of effluents, related to the decreasing of the molar fraction of methane, in the sense of the 2589

dx.doi.org/10.1021/ie100637b |Ind. Eng. Chem. Res. 2011, 50, 2585–2599

Industrial & Engineering Chemistry Research

ARTICLE

Figure 7. Evolution of the methane composition in the ATR. Contact spatial time and temperature effect. Operating conditions: feeding molar ratio 5:1:9 CH4/O2/H2O; 1.00 bar; Ni(5.75 wt %)/γ-Al2O3; 2 h of operation.

Figure 8. Evolution of the hydrogen composition in the ATR. Contact spatial time and temperature effect. Operating conditions: feeding molar ratio 5:1:9 CH4/O2/H2O; 1.00 bar; Ni(5.75 wt %)/γ-Al2O3; 2 h of operation.

approaching of the thermodynamic equilibrium composition, with increasing temperature, suggests an increase of methane conversion through endothermic reactions, such as cracking (eq 3), steam reforming (eqs 12 and 13), or even dry reforming (eqs 15-18), when carbon dioxide is present in the reactional system due to the combustion, either from methane or due to the deposited carbon or carbon monoxide, all of them produced in other parallel reactions. All of these reactions are spontaneous from 966 K (Table 1), when the equilibrium constants increase with the increase in temperature (Table 1), favoring methane consumption. Evaluating the hydrogen production, it is possible to affirm that, with the increase of the fluid-solid contact spatial time, at the studied flow rate, it was possible increase its production. The temperature increase also influences increased hydrogen production (Figure 8), nearing that of thermodynamic equilibrium composition, indicating that its formation is favored by endothermic processes, such as the cracking of methane (reaction 3), steam reforming of reactions 12 and 13, dry reforming of reactions 15 and 17, eventually occurring in the presence of carbon dioxide produced in parallel reactions, and gasification of deposited carbon through steam (eqs 25 and 26). The temperature increase contributes to the formation of carbon monoxide, as thermodynamics predicts (Table 3). This contribu-

tion was observed in an accentuated manner in the proposed ATR operation. The fluid-solid contact spatial time positively influences the formation of carbon monoxide (Figure 9). The increase in carbon monoxide, approaching the composition of the thermodynamic equilibrium, with temperature elevation, points to endothermic processes that promote its formation, such as the steam reforming of reaction 12, dry reforming reactions 15-18, and gasification of deposited carbon, either by steam, reaction 25, or by carbon dioxide in reverse Boudouard reaction, eq 27. All of these reactions occur at upper temperatures of 980 K. The WGSR, eq 29, once it is endothermic, operating at upper temperatures of 980 K, also contributes to carbon monoxide formation with elevated temperatures. Notice that the inclinations of curves in Figure 9 show that the influence of the fluid-solid contact spatial time in the formation of carbon monoxide is very discrete. The temperature increase effect is more important, although this effect is lower than what should be if the equilibrium constants of proposed reactions in the paragraph above are considered (Table 1), indicating that the carbon monoxide could be consumed exothermically through other parallel reactions. The formation of carbon dioxide by reaction 28 consumes carbon monoxide through its combustion at any of the operation temperatures of this work. Due to its consumption, reaction 19 decreases the carbon monoxide yield 2590

dx.doi.org/10.1021/ie100637b |Ind. Eng. Chem. Res. 2011, 50, 2585–2599

Industrial & Engineering Chemistry Research

ARTICLE

Figure 9. Evolution of the carbon monoxide composition in the ATR. Contact spatial time and temperature effect. Operating conditions: feeding molar ratio 5:1:9 CH4/O2/H2O; 1.00 bar; Ni(5.75 wt %)/γ-Al2O3; 2 h of operation.

Figure 10. Evolution of the carbon dioxide composition in the ATR. Contact spatial time and temperature effect. Operating conditions: feeding molar ratio 5:1:9 CH4/O2/H2O; 1.00 bar; Ni(5.75 wt %)/γ-Al2O3; 2 h of operation.

at temperatures lower than 1071 K. Also, reaction 21, when the operation temperature is lower than 1031 K, contributes to this effect. However, reactions 19 and 21 aid carbon deposition under these conditions. The composition in the steady state of hydrogen and carbon monoxide, observed in Figures 8 and 9, is the result of that which was established in the previous stoichiometric evaluation, when operation conditions were proposed to maximize hydrogen production to the detriment of carbon monoxide (operational conditions to conduct the process). The forecasts of the reactive conditions, in thermodynamic equilibrium, also point to a decrease of carbon dioxide production with the increase of temperature (Table 3). This fact was verified when autothermal reforming operations were performed (Figure 10). This decrease in the carbon dioxide yield with increasing temperature, moving away from thermodynamic equilibrium composition, suggests that its formation is governed by exothermic processes, such as reactions 8, 10, and 11, all of them reactions of methane combustion. Reactions 24 and 28, the gasification of deposited carbon and carbon monoxide combustion, respectively, contribute to the formation of carbon monoxide. All of these reactions, being exothermic, have equilibrium constants that decrease with increases in temperature (Table 1), and as result, the increasing of the quantity of carbon dioxide that is formed will be proportionally smaller than the increase in temperature.

Figure 11. Evolution of the methane conversion in the ATR. Contact spatial time and temperature effects. Operating conditions: feeding molar ratio 5:1:9 CH4/O2/H2O; 1.00 bar; Ni(5.75 wt %)/γ-Al2O3; 2 h of operation.

There is still reaction 21, which produces carbon dioxide only at lower temperatures of 1031 K, justifying once more why at upper temperatures the carbon dioxide yield decreases. Dry reforming reaction 14 also contributes to reducing the quantity of carbon dioxide in the composition of the effluent, provoking the carbon deposition. 2591

dx.doi.org/10.1021/ie100637b |Ind. Eng. Chem. Res. 2011, 50, 2585–2599

Industrial & Engineering Chemistry Research

ARTICLE

Figure 12. Evolution of the product yields in the ATR. Contact spatial time and temperature effects. Operating conditions: feeding molar ratio 5:1:9 CH4/O2/H2O; 1.00 bar; Ni(5.75 wt %)/γ-Al2O3; 2 h of operation.

Figure 13. Evolution of the syngas yield and H2/CO molar ratio in the ATR. Contact spatial time and temperature effects. Operating conditions: feeding molar ratio 5:1:9 CH4/O2/H2O; 1.00 bar; Ni(5.75 wt %)/γ-Al2O3; 2 h of operation.

At temperatures above 980 K, the WGSR, eq 29, is moved to the direction of carbon dioxide consumption in the process, which is increased with the increase in temperature. Operational Evaluations. Under the same conditions of pressure and temperature, flow rates of 150-400 cm3 were employed, in the presence of a nickel catalyst (60 mg, Ni(5.75 wt %)/γ-Al2O3). This allowed the observation of the operational behavior of the ATR process. For this purpose, on the basis of reactants and products, descriptions of conversions, yields, and selectivities were obtained. The methane conversion was influenced by the temperature as a result of its increase with the increase in temperature (Figure 11). As was commented, this influence indicates the promotion of endothermic reactions, such as methane cracking, reaction 3; its steam reforming through reactions 12 and 13; and/or its dry reforming through reactions 15-18. Hydrogen yields oscillated between 18 and 81%, with a tendency to increase when the fluid-solid contact spatial time increases. Carbon monoxide yields remained at reduced levels, varying between 2% and a maximum of 41%, inside the stoichiometric study predictions performed for the purpose of maximizing hydrogen production (Figure 12). The influence of temperature on yields of each product revealed a slight increase of these yields at upper thermal levels, but not so with carbon dioxide. In terms of syngas yields (Figure 13), the results obtained were between 10 and 61%.

Stable levels were observed of product selectivities of H2, CO, and CO2 at different fluid-solid contact spatial times; at three operation temperatures; and according to methane conversions (Figure 14). The process was selective to average productions of around 81% hydrogen (minimum of 78% and maximum of 84%), around 11% carbon dioxide (minimum of 5% and maximum of 18%), and around 9% carbon monoxide (minimum of 3% and maximum of 14%). Considering the proposal of hydrogen maximization, an elevated H2/CO molar ratio was experimentally demonstrated. In fact, Figure 13 presents this ratio between 5 and 30 to lower contact spatial time (9 kg cat 3 s/m3). The ratios diminished at larger contact spatial times, favoring the formation of monoxide carbon. Anyway, these elevated values revealed that, under the stipulated conditions, the process occurred, maximizing the hydrogen and decreasing the carbon monoxide production. Variations of H2/CO molar ratios, relative to its decrease with increasing temperature, are the result of the influence of reactions that occur in the overall process, indicating that the formation of hydrogen should be a function of both endothermic processes—reactions 3 and 13—or exothermic—reaction 5. Moreover, monoxide carbon formation should be influenced more by endothermic processes—reaction 27 and the reverse of reaction 29—than by exothermic processes— reaction 5—although these processes also are involved. 2592

dx.doi.org/10.1021/ie100637b |Ind. Eng. Chem. Res. 2011, 50, 2585–2599

Industrial & Engineering Chemistry Research

ARTICLE

Figure 14. Evolution of the product selectivities versus methane conversion in the ATR. Temperature effect. Operating conditions: feeding molar ratio 5:1:9 CH4/O2/H2O; 1.00 bar; Ni(5.75 wt %)/γ-Al2O3; 2 h of operation.

Initial Considerations on the Mechanism Being Proposed. Under the given conditions of temperature, pressure, and flow rates and using the nickel catalyst prepared, as well as considering the thermodynamic predictions and operational assessments, 12 of the reactions of Table 1, reactions 3, 5, 11, 12, 13, 15, 23, 24, 25, 27, 28, and 29, were initially selected to be evaluated to compose a proposed mechanism. Therefore, the following approach was developed. Note that, because reactions 5, 11, 23, 24, and 28 are combustion, they are considered irreversible. Reactions 12 and 13 are related to the steam reforming of methane, usually accompanied by reaction 29, known as WGSR. This process is highly endothermic, and Chesnokov et al.25 suggest running in a tubular reactor to 1173 K and 15-30 bar. The steam reforming of methane is an important and wellestablished industrial technology for producing hydrogen and synthesis gas, which can be used to make ammonia, methanol, synthetic fuels, and dimethyl ether and in oxo synthesis. Because it is highly endothermic, it requires large amounts of energy, making it a very expensive process. In industry, this is a catalytic process in which the catalyst is nickel, supported on ceramic materials such as Al2O3, MgAlOx, or Zr2O3, and may also have stabilization promoters, such as CaAl2O4, K2O, and SiO2.26 Reaction 5 represents the NC-POM (or POX). As this is an exothermic reaction, the heat generated can be used for the SRM. Thus, the combination of these processes has led to the development of ATR.27 In autothermal reforming, high rates of conversion of methane can be obtained at lower temperatures than for methane reforming when the partial oxidation of methane occurs spontaneously, which is why it has been studied.28 However, it has been reported that the synthesis gas can be produced in two steps, consisting of the total combustion of methane—reaction 11—followed by steam reforming—reaction 12—and reforming of methane with CO2 (dry reforming)—reaction 15.29 Since the combustion reaction is much faster than the reforming reaction, it is common to do it catalytically near the entrance of the bed, leaving the reforming reaction to occur in the catalytic bed itself, after the oxygen has been consumed.30,31 Normally, the combustion zone does not overlap the reforming area; however, it is normal that the temperature of the catalyst in the combustion zone is very high, which causes a high thermal gradient in the catalyst bed during the POM. In addition, there

are hot spots where the temperature is much higher than in other parts of the reactor, this being one of the main difficulties in conducting the partial oxidation.32,33 In the presence of oxygen, metallic nickel is oxidized to ionic nickel (Ni2þ), and only the combustion reaction occurs. In the absence of oxygen, that is, after all oxygen reacted with methane, the remaining metallic nickel continues its catalytic activity for reforming. This also causes a high temperature gradient region related to the exothermic and endothermic region, once they are separated. To avoid this gradient, studies have been proposed in which partial oxidation and reforming of methane to produce synthesis gas have been carried out in a fluidized bed.34-37 In a fluidized bed, as long as the heat transfer from the exothermic region to the endothermic region is carried out effectively, the formation of hot spots is effectively reduced, if not inhibited. As a result, the stability and security of the operation are achieved successfully. Nevertheless, there are still problems related to the friction of particles of the catalyst. Thus, efforts have been made to produce synthesis gas by a combination of processes, such as studies on the addition of steam and CO2 in the POM (resulting in the CRM), and the effects of the addition of O2 to steam reforming (resulting in ATR) or DRM.32,33,38 When the SRM and DRM are combined with POM, there is the possibility of significant reduction of hot spots. Thus, it is possible to reach thermoneutral conditions by adjusting the partial pressure ratio of CH4, O2, and steam (or CO2).39 Furthermore, in an autothermal reforming process on a large scale, the contact of oxygen with the catalyst for steam reforming should be avoided to reduce the chances of hot spots.40-42 In addition to problems related to high operating temperatures for the ATR, such as the involvement of thermal stability of the catalyst and the tendency of steam to promote the sintering of support and catalyst, the main problem is coke formation or coking, as can be seen from reaction 3 and the reverse reactions 27 and 25. Thus, it is extremely important to reduce or avoid coke formation on the catalyst. If we observe the reverse reaction 25, it is the most logical way to increase the molar ratio of steam, so that the reaction begins to consume the carbon which has formed. However, despite this method being used on an industrial scale, the generation of steam at high temperatures is quite expensive, which tends to reduce the economic advantages arising from the reduction of carbon deposition significantly. Apparently, the carbon deposition during steam reforming of methane occurs predominantly due to the decomposition of methane—reaction 2593

dx.doi.org/10.1021/ie100637b |Ind. Eng. Chem. Res. 2011, 50, 2585–2599

Industrial & Engineering Chemistry Research

ARTICLE

Table 4. Sequence of Reactions in the Autothermal Reforming (proposed practice) reactions

obs. 1 O2 f CO þ 2H2 2

rates

equation

r2

5

2

partial oxidation

CH4 þ

5

steam reforming

CH4 þ 2H2 O f CO2 þ 4H2

r5

13

6

reserve of WGSR

CO2 þ H2 f CO þ H2 O

r6

29

7

cracking

CH4 h C þ 2H2

often ignored (negligible)16

r7

3

8

gasification of C by CO2

C þ CO2 f 2CO

usually discarded if it is assumed that there is no carbon deposition

r8

27

gasification of C by O2

C þ O2 f CO2

r11

24

11

3. In recent years, the production of hydrogen by methane decomposition was used, consisting of the passage of methane over a metal catalyst, in order to produce hydrogen and coke, which was deposited on the catalyst surface, “poisoning” it. After a given time, the catalyst needed to be regenerated, which was done by passing air over it at high temperatures to burn the carbon depositions—reactions 23 and 28.43 On the basis of the previous considerations, to complete the reaction mechanism, it is necessary to remember that, in catalytic reforming, as well as in other processes, the reactions that occur depend on operating conditions. Thus, to reduce the complexity of the mathematical model to be adopted, only the reactions whose rates are significant in terms of experimental results should be considered. Therefore, the ATR, based on its POM and SRM, often is accompanied by the following reactions: total oxidation of methane, reaction 11; partial oxidation of methane, reaction 5; reforming with carbon dioxide (dry reforming), reaction 15; steam reforming, reactions 12 and 13; water-gas shift, reaction 29; cracking of methane, reaction 3; deposition of carbon from carbon monoxide, reactions 25 and 27; gasification of carbon by oxygen, reactions 23, 24, and 28. Gosiewski et al.,44 thermodynamically analyzing the stoichiometry of the ATR, found that the rate of synthesis gas formation from the reforming reaction of methane with CO2 (DRM)— reaction 15—is much smaller than that of complete oxidation— reaction 11—and the reactions of steam reforming—reactions 12 and 13. Chan and Wang16 also performed a thermodynamic analysis of the processing of natural gas for use in fuel cells and found that, as shown in reaction 25, the addition of water to the partial oxidation process is an effective way to suppress the coke formation. Doing simulations, they showed that for a W/F (water/fuel) molar ratio of more than 1.5, regardless of the A/F (air/fuel) molar ratio, the autothermal reforming is free of carbon deposition (see Figure 1). So, in conclusion, the reaction mechanism includes, in principle, the 12 reactions mentioned—3, 5, 11, 12, 13, 15, 23, 24, 25, 27, 28, and 29—whose concentrations of the following gaseous species, methane (CH4), oxygen (O2), carbon dioxide (CO2), carbon monoxide (CO), steam (H2O), and hydrogen (H2), were analyzed. The carbon deposition (C) was quantified by the total mass deposited on the catalyst. During kinetic tests, oxygen was introduced to the feed stream of argon, which is why one has to also consider argon (Ar) as a diluent, which will affect only the gas properties.

Suggested Reaction Scheme. The experiments performed pointed toward selectivity to hydrogen and carbon dioxide, the first between 78 and 84% and the second between 5 and 18%. Carbon monoxide composed between 3% to 14% of the final product. Thus, considering the preferred way for the reactive process to form hydrogen and carbon dioxide, the latter competing with carbon monoxide, the 12 steps of the reaction, initially suggested, were reduced to a smaller number. Thus, considering the nonsupply of carbon dioxide and the least amount of carbon monoxide formed, since the water-gas shift reaction, eq 29, does not occur at the operational temperatures, in regard to consumption of carbon monoxide and steam, reaction 15 should not occur in measurable proportions. Considering the reactions of the steam reforming of methane, reaction 13 requires a smaller amount of energy than reaction 12 (Table 1) and has higher equilibrium constants at temperatures of 1023 and 1073 K (Table 2); therefore, reaction 13 should occur at the expense of reaction 12. Furthermore, the feeding molar ratio in the ATR is 5:1:9 CH4/O2/H2O, a function of the proposed model; i.e., the molar ratio of CH4/H2O is almost 1:2, enough to supply reaction 13. Despite the reactions of coke deposition from the carbon monoxide and methane cracking being ignored by Chan and Wang,16 reaction 3 was accepted, since the presence of coke was detected by performing thermogravimetric analyses (TG and DTG), in addition to the reasons already given in the comments of Figure 7. Once the coke deposition was considered, the reaction mechanism was permitted; among the reactions of carbon gasification by oxygen, only reaction 24 was considered, due to its equilibrium constant, which is much larger than those of reactions 23 and 28 (Table 1). In addition, as per the comments regarding the formation of carbon monoxide, Figure 9, such production should occur through endothermic reactions, and reaction 23 is exothermic and therefore should not take place. In turn, reaction 27, the reverse Boudouard reaction, also was admitted, being endothermic, enabling the formation of carbon monoxide. The reverse of WGSR, reaction 29, occurs at temperatures above 980 K, consuming part of the carbon dioxide and hydrogen formed, increasing the yield of carbon monoxide with an increase in temperature. Considering the high production of hydrogen, together with that of carbon dioxide, followed by the reduced amount of carbon monoxide production, it is suggested that reaction 13 and the reverse reaction 29 are both practically irreversible. Weak adsorptions of CO and H2 were considered in reactions 5 and 13. 2594

dx.doi.org/10.1021/ie100637b |Ind. Eng. Chem. Res. 2011, 50, 2585–2599

Industrial & Engineering Chemistry Research

ARTICLE

Descriptions of the stationary behavior observed during the operations in the fixed bed reactor are represented via balance equations of the gaseous components i = CH4, O2, H2O, H2, CO, and CO2 produced by eq 31 and following the principles adopted by the mechanisms of Langmuir-Hinshelwood, Eley-Rideal, YangHougen for the reaction rates: r2, partial oxidation of methane;29,45,46 r5, steam reforming of methane;47,48 r6, the reverse of WGSR;47,48 r7, cracking of methane;49 r8, the reverse Boudouard reaction;50 and r11, total combustion of carbon.50 - RT

comb KCH pCH4 ðKO2 pO2 Þ2 dpCH4 4 - k2 þ comb dτ ½1 þ KCH pCH4 þ ðKO2 pO2 Þ2 2 4 ! pCH4  p2H2 O p3H2:5 - k5 !2 K H2 O p H2 O 1 þ KCH4  pCH4 þ p H2

Figure 15. Diagram of the suggested reactions for the ATR.

- k7

The total combustion, reaction 11, is unlikely to occur because it is an exothermic process, and as commented upon in Figure 7, the conversion of methane is predominantly by endothermic processes. In addition, the feed does not use oxygen in large quantities, which is one reason not to consider it in the mechanism. Reaction 25 should not occur, due to it having lower equilibrium constants than those of reaction 13, since both share in the consumption of steam. Finally, after the catalyst evaluations were performed, the preferred path to form hydrogen and carbon dioxide, the latter competing with carbon monoxide, was evaluated, in terms of selectivity, leaving only six reaction steps to be considered (Table 4): the cracking of methane, eq 3; the partial combustion of methane, eq 5; steam reforming, eq 13; the combustion of carbon, eq 24; the classic reverse Boudouard reaction, eq 27; and the reverse of WGSR, eq 29. In view of the proposed mechanism, a diagram depicting the reactions that occurred during the ATR under the conditions presented in this work is shown in Figure 15. The numbering in the representation of reaction rates is not random but is according to the 12 reactions initially considered. Representation of the Kinetic Behavior. The rates of formation and consumption of each species involved take into account the reaction rates of these species in each of the six steps suggested. Under the conditions applied, using the catalyst particle size and adequate flow rates, the process was conducted in a kinetic chemical regime. Thus, the reaction rates of the components in the process of ATR were formulated as expressed by relations 36-42. RCH4 ¼ - r2 - r5 - r7

- RT

KCH4  pCH4 ¼0 1 þ KCH4  pCH4

ð43Þ

comb KCH pCH4 ðKO2 pO2 Þ2 dpO2 1 4 - k2 - k11  pO2 ¼ 0 comb 2 ½1 þ KCH dτ pCH4 þ ðKO2 pO2 Þ2 2 4

- RT

dpH2 O - 2k5 dτ

pCH4  p2H2 O p3H2:5 1 þ KCH4  pCH4 þ

ð44Þ

!

KH2 O pH2 O pH2

!2

- k6  pH2  pCO2 ¼ 0

ð45Þ

comb KCH pCH4 ðKO2 pO2 Þ2 dpH2 4 þ2 þ comb dτ ½1 þ KCH pCH4 þ ðKO2 pO2 Þ2 2 4 ! pCH4  p2H2 O p3H2:5 4k5 !2 þ K H2 O p H2 O 1 þ KCH4  pCH4 þ p H2

- RT

k6  pH2  pCO2 þ 2k7

- RT

ð36Þ

ð46Þ

KCH4  pCH4 ¼0 1 þ KCH4  pCH4

comb KCH pCH4 ðKO2 pO2 Þ2 dpCO 4 þ k2 - k6  pH2  pCO2 comb dτ ½1 þ KCH pCH4 þ ðKO2 pO2 Þ2 2 4

þ 2k8  p0CO:52 ¼ 0

ð47Þ

1 ¼ - r2 - r11 2

ð37Þ

RH2 O ¼ - 2r5 - r6

ð38Þ

RH2 ¼ 2r2 þ 4r5 þ r6 þ 2r7

ð39Þ

RCO ¼ r2 - r6 þ 2r8

ð40Þ

RCO2 ¼ r5 þ r6 þ r11 - r8

ð41Þ

þ k6  pH2  pCO2 þ k11  pO2 - k8  p0CO:52 ¼ 0 ð48Þ

ð42Þ

where R is the universal constant of an ideal gas (J mol-1 K-1); T is the operating temperature (K); p is the partial pressure of

R O2

RC ¼ r7 - r11 - r8

- RT

2595

dpCO2 þ k5 dτ

pCH4  p2H2 O p3H2:5 1 þ KCH4  pCH4

!

KH2 O pH2 O þ pH2

!2

dx.doi.org/10.1021/ie100637b |Ind. Eng. Chem. Res. 2011, 50, 2585–2599

Industrial & Engineering Chemistry Research

ARTICLE

gaseous species (bar); k2, k5, k6, k7, k8, and k11 are the rate constants of reactions defined by the subscripts; KCH4comb; KO2, KCH4, and KH2O are the constants of adsorption for the species also defined by the subscripts; and τ (kg cat 3 s/m3) is the fluidsolid contact spatial time. Mathematical Modeling. The mathematical model representing the behavior of the steady process of the ATR—eqs 43-48—had the solutions of the models adjusted and compared to changes in

concentrations of the components involved. For these purposes, the adsorption parameters were introduced following the specialized literature.51 Estimation of Adsorption Parameters. To determine the adsorption constants, it was taken into consideration that the adsorption of methane during combustion has different constants than in the steps of steam reforming and cracking of methane, represented by KCH4comb and KCH4, respectively. The constant of adsorption of oxygen in the combustion reaction, as a matter of similarity, also begins to be represented by KO2comb. The values of the constants of adsorption were determined as proposed by Hoang and Chan,51 expressed as follows:   12:448 comb 5 ð49Þ KCH4 ¼ 4:02  10 exp bar - 1 T   7:962 comb 4 KO2 ¼ 5:08  10 exp ð50Þ bar - 0 :5 T   4:604 -4 KCH4 ¼ 6:65  10 exp ð51Þ bar - 1 T   10:666 KH2 O ¼ 1:77  105 exp ð52Þ T

Table 5. Parameters of the Constants of Adsorption51 species (i)

Ko,i

Hi (kJ/mol)

CH4 (combustion) O2 (combustion)

4.02  105 bar-1 5.08  104 bar-0.5

103.50 66.20

CH4

6.65  10-4 bar-1

-38.28

H2O

1.77  105

88.68

Table 6. Values of the Constants of Adsorption T (K)

KCH4comb (bar-1)

KO2comb (bar-0,5)

KCH4 (bar-1)

KH2O

1023 1073

2.089 3.683

21.28 30.44

0.0598 0.0486

5.255 8.542

1123

6.255

42.63

0.0406

13.295

Table 7. Kinetic Constants (mol 3 kg-1 cat 3 s-1) k3 (105; bar-2)

k4 (10)

k5 (108; bar-0.5)

k6 (105; bar-1)

8.25

4.22

4.00

0.95

2.03

31.37

6.10

6.48

2.15

2.91

8.53

10.07

4.54

4.04

k2 (102; bar0.5)

T (K)

k1 (10)

1023

1.31

1073

3.30

1123

7.63

105.95

Table 8. Parameters of Kinetic Constants constants (i) partial oxidation

k2

5

steam reforming

k5

13

reverse of WGSR cracking

k6 k7

29 3

gasification of C by CO2

k8

27

k11

24

gasification of C by O2

ko,i

eq

4.94  107 mol/(s 3 kg cat)

2.36  1011 mol 3 bar0.5/(s 3 kg cat) 1.07  10-1 mol/(bar2 3 s 3 kg cat) 1.28  104 mol/(s 3 kg cat)

4.08  10-1 mol/(bar0.5 3 s 3 kg cat) 4.74  10-2 mol/(bar 3 s 3 kg cat)

Ea,i (kJ/mol) 167.97 243.90 67.13 88.25 149.26 66.22

Figure 16. Model predictions and experimental results of ATR at 1023, 1073, and 1123 K. Partial pressures of effluents as a function of contact time (τ = mcat/Qtotal,reac). Operating conditions: feeding molar ratio 5:1:9 CH4/O2/H2O; 1.00 bar; Ni(5.75 wt %)/γ-Al2O3; 2 h of operation. 2596

dx.doi.org/10.1021/ie100637b |Ind. Eng. Chem. Res. 2011, 50, 2585–2599

Industrial & Engineering Chemistry Research The values of the pre-exponential factors and energies of adsorption are, therefore, presented in Table 5. The calculated values for the operating temperatures of this work are presented in Table 6. Kinetic Constants. The kinetic constants (kj, j = 2, 5, 6, 7, 8, and 11) resulted from the application of a methodology to adjust the solutions of the equations of the model to experimental results. Initially, preliminary values were obtained via a differential method applied to the balance equations. These values were introduced in the solutions of the equations, which could be compared, in terms of partial pressures of components, with the corresponding experimental results. Smaller changes in the values of the constants followed by an optimization method allowed for a final adjustment. The solutions of the equations were obtained using the Runge-Kutta fourthorder method (Matlab), then upgraded to a curve fitting method performed by the Runge-Kutta-Fehlberg method,52 developed in Maple. The estimated kinetic constants are presented in Table 7. The values of the pre-exponential factors and activation energies estimated are presented in Table 8. Validation of the Model Proposed. The evolutions of the partial pressures of the components present in the operations of the ATR process are presented in Figure 16. Calculations based on the solutions of the equations of the model and its corresponding experimental results are presented in the figure for the purpose of demonstrating the capability of representing the behavior of the process. The application of the model representative of the kinetic behavior of the operating-process ATR in the presence of Ni(5,75 wt %)/γ-Al2O3 allowed the description of the evolution kinetics of reactants and products under the operating conditions set forth in this work. In terms of partial component pressures, there were deviations that averaged from 7.5% to 14.0% between the calculated partial pressures (simulation) and the experimental pressures. The overall average deviation representative of the adequacy of the proposed model with respect to the experimental results was 10.5 ( 0.82%.

’ CONCLUSION A proposed mechanism to the ATR, based on six reaction steps, employing a nickel catalyst supported on γ-alumina, under thermodynamic and stoichiometric conditions, allowed reaching yields of up to 81% hydrogen, along with selectivity of 81% hydrogen, with methane conversions around 63%.

’ AUTHOR INFORMATION Corresponding Author

*Tel.: þ55 81 2126 7289. E-mail: [email protected].

’ ACKNOWLEDGMENT We would like to thank to the Department of Chemical Engineering of UFPE (Universidade Federal de Pernambuco), to FINEP, and to Petrobras, for their financial support. CTGAS-RN and UFRN are also acknowledged, for enabling use of facilities to perform some of the analyses. L.J.L.M. gratefully acknowledges CAPES for the fellowship that was awarded. ’ NOMENCLATURE aik = number of atoms in the kth element in the molecule of species i (dimensionless)

ARTICLE

Ak = total number of atomic masses of the kth element (dimensionless) Ci = molar concentration of species i (mol/m3) dr = inner diameter of the reactor (cm) dio = inner diameter of inlet and outlet of the gas flow in the reactor (mm) dp = diameter of particle ds = pore diameter of sintered plate in the bed (μm) dt = inner diameter of the pit to the thermocouple (mm) Ea,i = activation energy of species i (kJ/mol) Hi = energies of adsorption of species i (kJ/mol) ko,i = pre-exponential factors of species i for kinetic constants (variable dimension) k2 = kinetic constant (mol 3 kg-1 cat 3 s-1) k5 = kinetic constant (mol 3 kg-1 cat 3 s-1 3 bar0.5) k6, k7, k11 = kinetic constants (mol 3 kg-1 cat 3 s-1 3 bar-1) k8 = kinetic constant (mol 3 kg-1 cat 3 s-1 3 bar-0.5) Ko,i = pre-exponential factors of species i for equilibrium constants of adsorption (variable dimension) KCH4 = equilibrium constant of adsorption for methane (atm-1) KCH4comb = equilibrium constant of adsorption for methane combustion (atm-1) KH2O = equilibrium constant of adsorption for steam (dimensionless) KO2, KO2comb = adsorption equilibrium constant for oxygen (atm-0.5) Lb = length of bed useful (cm) Lr = overall length (cm) mcat = mass of the catalyst (kg) N = number of moles (mol); number of equations for equilibrium ni = number of moles of species i (dimensionless) pi = partial pressure of species i (bar) P = pressure (bar) P = standard pressure, chemical species formation (bar) Q_ inlet = volumetric flow rate of the reactants (m3/h) rj = rate of reaction j (mol 3 kg-1 cat 3 s-1) R = universal gas constant (kJ/kmol K) Ri = rate of reaction with respect to component i (mol 3 kg-1 cat 3 s-1) T = absolute temperature (K) w = number of equations for mass balances yi = mole fraction of species i (dimensionless) Greek symbols

ΔG,ΔGfi = Gibbs energy, of formation of the species i (kJ/mol) ΔH = enthalpy (kJ/mol) λk = Lagrange multiplier for the atomic species k (kJ/mol) Fcat = density of the catalyst (kg/m3) τ = contact spatial time (h 3 kg cat 3 m-3) Subscripts

a = activation b = bed cat = catalyst fi = formation of species i i = chemical species io = inlet and outlet j = chemical reaction k = chemical element o = initial condition r = reactor s = sintered plate t = thermocouple 2597

dx.doi.org/10.1021/ie100637b |Ind. Eng. Chem. Res. 2011, 50, 2585–2599

Industrial & Engineering Chemistry Research Superscripts

comb = combustion  = standard Abbreviations and acronyms

AAS = atomic absorption spectroscopy A/F = air/fuel ATR = autothermal reforming of methane BET = Brunauer-Emmett-Teller isotherm CAPES = Coordenac-~ao de Aperfeic- oamento de Pessoal de Nível Superior CRM = combined reforming of methane CTGAS-RN = Centro de Tecnologias do Gas and Energias Renovaveis do Rio Grande do Norte DRM = dry reforming of methane DTA = differential thermal analysis DTG = derivate thermogravimetry FINEP = Financiadora de Estudos e Projetos FTIR = Fourier transform infrared spectroscopy GC = gas chromatography GTL = gas-to-liquids GHSV = gas hours space velocity IR = infrared NG = natural gas NC-POM = noncatalytic partial oxidation of methane POM = catalytic partial oxidation of methane POX = partial oxidation SEM = scanning electron microscope SRM = steam reforming of methane STP = standard temperature and pressure (25 C, 1 atm) syngas = synthesis gas TCD = thermal conductivity detection TG = thermogravimetry UFPE = Universidade Federal de Pernambuco UFRN = Universidade Federal do Rio Grande do Norte W/F = water/fuel WGSR = water-gas shift reaction wt % = % by weight XRD = X-ray diffraction

’ REFERENCES (1) Pe~na, M. A.; Gomez, J. P.; Fierro, J. L. G. New catalytic routes for syngas and hydrogen production. Appl. Catal., A 1996, 144, 7–57. (2) H.A/S Topsøe, Societe Belge de L’Azote. Hydrocarbon Process. Int. Ed. 1998, 67 (4), 77. (3) Wang, S.; Lu, G. Q. Reforming of methane with carbon dioxide over Ni/Al2O3 catalysts: effect of nickel precursor. Appl. Catal., A 1998, 169, 271–280. (4) Nichio, N.; Casella, M.; Ferreti, O.; Gonzalez, M.; Nicot, C.; Moraweck, B.; Frety, R. Partial oxidation of methane to synthesis gas: behaviour of different Ni supported catalysts. Catal. Lett. 1996, 42, 65– 72. (5) Chen, I.; Lin, S.-Y.; Shiue, D.-W. Calcination of nickel/alumina catalysts. Ind. Eng. Chem. Res. 1988, 27, 926–929. (6) Smith, J. M.; Van Ness, H. C.; Abbott, M. M. Introduc-~ao a Termodin^amica da Engenharia Química, 5th ed.; LTC Editoria: Rio de Janeiro, 2000. (7) Smith, J. M.; Van Ness, H. C.; Abbott, M. M. Introduction to Chemical Engineering Thermodynamics, 5th ed.; McGraw-Hill: New York, 1996. (8) Perry, R. H.; Green, D. W. Perry’s chemical engineers’ handbook, 7th ed.;McGraw-Hill: New York, 1997. (9) Kelley, K. K. Contributions to the data on theoretical metallurgy. XIII. High-temperature heat-content, heat-capacity, and entropy data for

ARTICLE

the elements and inorganic compounds. U.S. Bur. Mines Bull. 1960, 584, 232. (10) Takahashi, Y.; Westrum, E. F., Jr. Glassy carbon. Low-temperature thermodynamic properties. J. Chem. Thermodyn. 1970, 2, 847–854. (11) Spencer, H. M. Empirical heat capacity equations of gases and graphite. Ind. Eng. Chem. 1948, 40, 2152–2154. (12) Pankratz, L. B. Thermodynamic properties of elements and oxides. U. S. Bur. Mines Bull. 1982, 672, 509 p. (13) Wagman, D. D.; Evans, W. H.; Parker, V. B.; Schumm, R. H.; Halow, I.; Bailey, S. M.; Churney, K. L.; Nuttall, R. L. The NBS tables of chemical thermodynamic properties. J. Phys. Chem. Ref. Data 1982, 11 (2), 2–12. (14) Chase, M. W., Jr. NIST-JANAF Themochemical Tables, Fourth Edition. J. Phys. Chem. Ref. Data 1998, Monograph 9, 1-1951. (15) Ersoz, A.; Olgun, H.; Ozdogan, S.; Gungor, C.; Akgun, F.; Tiris, M. Autothermal reforming as a hydrocarbon fuel processing option for PEM fuel cell. J. Power Sources 2003, 118, 384–392. (16) Chan, S. H.; Wang, H. M. Thermodynamic analysis of naturalgas fuel processing for fuel cell applications. Int. J. Hydrogen Energy 2000, 25, 441–449. (17) Takehira, K.; Shishido, T.; Wang, P.; Kosaka, T.; Takaki, K. Autothermal reforming of CH4 over supported Ni catalysts prepared from Mg-Al hydrotalcite-like anionic clay. J. Catal. 2004, 221, 43–54. (18) Ayabe, S.; Omoto, H.; Utaka, T.; Kikuchi, R.; Sasaki, K.; Teraoka, Y.; Eguchi, K. Catalytic autothermal reforming of methane and propane over supported metal catalysts. Appl. Catal., A 2003, 241, 261–269. (19) Hagh, B. F. Optimization of autothermal reactor for maximum hydrogen production. Int. J. Hydrogen Energy 2003, 28, 1369–1377. (20) Lee, S. H. D.; Applegate, D. V.; Ahmed, S.; Calderone, S. G.; Harvey, T. L. Hydrogen from natural gas: part I - autothermal reforming in an integrated fuel processor. Int. J. Hydrogen Energy 2005, 30, 829– 842. (21) Ma, L.; Trimm, D. L.; Jiang, C. The design and testing of an autothermal reactor for the conversion of light hydrocarbons to hydrogen, I. The kinetics of the catalytic oxidation of light hydrocarbons. Appl. Catal., A 1996, 138, 275–283. (22) Maluf, S. S.; Assaf, E. M.; Assaf, J. M. Catalisadores Ni/Al2O3 promovidos com molibd^enio para a reac-~ao de reforma a vapor de Metano. Quím. Nova 2003, 26 (2), 181–187. (23) Bitter, J. H.; Seshan, K.; Lercher, J. A. Deactivation and coke accumulation during CO2/CH4 reforming over Pt catalysts. J. Catal. 1999, 183, 336–343. (24) Abreu, C. A. M.; Santos, D. A.; Pacífico, J. A.; Lima Filho, N. M. Kinetic evaluation of methane-carbon dioxide reforming process based on the reaction steps. Ind. Eng. Chem. Res. 2008, 47, 4617–4622. (25) Chesnokov, V. V.; Zaikovskii, V. I.; Buyanov, R. A.; Molchanov, V. V.; Plyasova, L. M. Morphology of carbon from methane on nickelcontaining catalysts. Catal. Today 1995, 24, 265–267. (26) Freni, S.; Calogero, G.; Cavallaro, S. Hydrogen production from methane through catalytic partial oxidation reactions. J. Power Sources 2000, 87, 28–38. (27) Aasberg-Petersen, K.; Bak Hansen, J.-H.; Christensen, T. S.; Dybkjaer, I.; Christensen, P. S.; Nielsen, C. S.; MADSEN, S. E. L. W.; Rostrup-Nielsen, J. R. Technologies for large-scale gas conversion. Appl. Catal., A 2001, 221, 379–387. (28) Li, B.; Maruyama, K.; Nurunnabi, M.; Kunimori, K.; Tomishige, K. Temperature profiles of alumina-supported noble metal catalysts in autothermal reforming of methane. Appl. Catal., A 2004, 275, 157– 172. (29) Jin, W.; Gu, X.; Li, S.; Huang, P.; Xu, N.; Shi, J. Experimental and simulation study on a catalyst packed tubular dense membrane reactor for partial oxidation of methane to syngas. Chem. Eng. Sci. 2000, 55, 2617–2625. (30) Dissanayake, D.; Rosynek, M. P.; Lunsford, J. H. Are the equilibrium concentrations of carbon monoxide and hydrogen exceeded during the oxidation of methane over a nickel/ytterbium oxide catalyst. J. Phys. Chem. 1993, 97, 3644–3646. 2598

dx.doi.org/10.1021/ie100637b |Ind. Eng. Chem. Res. 2011, 50, 2585–2599

Industrial & Engineering Chemistry Research

ARTICLE

(31) Mallens, E. P. J.; Hoebink, J. H. B. J.; Marin, G. B. An investigation on the reaction mechanism for the partial oxidation of methane to synthesis gas over platinum. Catal. Lett. 1995, 33, 291–304. (32) Tomishige, K.; Kanazawa, S.; Suzuki, K.; Asadullah, M.; Sato, M.; Ikushima, K.; Kunimori, K. Effective heat supply from combustion to reforming in methane reforming with CO2 and O2: comparison between Ni and Pt catalysts. Appl. Catal., A 2002, 233, 35–44. (33) Tomishige, K.; Kanazawa, S.; Ito, S.; Kunimori, K. Catalyst development for direct heat supply from combustion to reforming in methane reforming with CO2 and O2. Appl. Catal., A 2003, 244, 71–82. (34) Abashar, M. E. E. Coupling of steam and dry reforming of methane in catalytic fluidized bed membrane reactors. Int. J. Hydrogen Energy 2004, 29, 799–808. (35) Chen, Z.; Elnashaie, S. S. E. H. Bifurcation and its implications for a novel autothermal circulating fluidized bed membrane reformer for the efficient pure hydrogen production. Chem. Eng. Sci. 2005, 60, 4287– 4389. (36) Chen, X.; Honda, K.; Zhang, Z.-G. A comprehensive comparison of CH4-CO2 reforming activities of NiO/Al2O3 catalysts under fixed- and fluidized-bed operations. Appl. Catal., A 2005, 288, 86–97. (37) Jing, Q.; Lou, H.; Mo, L.; Zheng, X. Comparative study between fluidized bed and fixed bed reactors in methane reforming with CO2 and O2 to produce syngas. Energy Conver. Manage. 2005, 47, 459–469. (38) Maciel, L. J. L.; Souza, A. E. A. M.; Vasconcelos, S. M.; Knoechelmann, A.; Abreu, C. A. M. Dry reforming and partial oxidation of natural gas to syngas production. Stud. Surf. Sci. Catal. 2007, 167, 469– 474. (39) Vernon, P. D. F.; Green, M. L. H.; Cheetham, A. K.; Ashcroft, A. T. Partial oxidation of methane to synthesis gas, and carbon dioxide as an oxidizing agent for methane conversion. Catal. Today 1992, 13, 417–426. (40) Choudhary, V. R.; Rajput, A. M.; Prabhakar, B. Energy efficient methane-to-syngas conversion with low H2/CO ratio by simultaneous catalytic reactions of methane with carbon dioxide and oxygen. Catal. Lett. 1995, 32, 391–396. (41) O’Connor, A. M.; Ross, J. R. H. The effect of O2 addition on the carbon dioxide reforming of methane over Pt/ZrO2 catalysts. Catal. Today 1998, 46, 203–210. (42) Liu, S.; Xiong, G.; Dong, H.; Yang, W. Effect of carbon dioxide on the reaction performance of partial oxidation of methane over a LiLaNiO/γ-Al2O3 catalyst. Appl. Catal., A 2000, 202, 141–146. (43) Poirier, M. G.; Sapundzhiev, C. Catalytic decomposition of natural gas to hydrogen for fuel cell applications. Int. J. Hydrogen Energy 1997, 22, 429–433. (44) Gosiewski, K.; Bartmann, U.; Moszczynski, M.; Mleczko, L. Effect of the intraparticle mass transport limitations on temperature profiles and catalytic performance of the reverse-flow reactor for the partial oxidation of methane to synthesis gas. Chem. Eng. Sci. 1999, 54, 4589–4602. (45) Xu, J.; Froment, G. F. Methane steam reforming, methanation and water-gas shift: I. Intrinsic kinetics. AIChE J. 1989a, 35, 88–96. (46) Xu, J.; Froment, G. F. Methane steam reforming: II. Diffusional limitations and reactor simulation. AIChE J. 1989b, 35, 97–103. (47) Yang, K. H.; Hougen, O. A. Determination of mechanism of catalyzed gaseous reactions. Chem. Eng. Prog. 1950, 46, 146–157. (48) Trimm, D. L.; Lam, C.-W. The combustion of methane on platinum-alumina fibre catalysts-I Kinetics and mechanism. Chem. Eng. Sci. 1980, 35, 1405–1413. (49) Snoeck, J. W.; Froment, G. F.; Fowles, M. Kinetic study of the carbon filament formation by methane cracking on a nickel catalyst. J. Catal. 1997, 169, 250–262. (50) Chen, C.; Horio, M.; Kojima, T. Numerical simulation of entrained flow coal gasifiers. Part I: modeling of coal gasification in an entrained flow gasifier. Chem. Eng. Sci. 2000, 55, 3861–3874. (51) Hoang, D. L.; Chan, S. H. Modeling of a catalytic autothermal methane reformer for fuel cell applications. Appl. Catal., A 2004, 268, 207–216. (52) Burden, R. L.; Faires, J. D. Numerical Analysis, 6th ed.; Thomson Brooks/Cole: Florence, KY, 1997; 811 p. 2599

dx.doi.org/10.1021/ie100637b |Ind. Eng. Chem. Res. 2011, 50, 2585–2599