Experiments and Modeling of Methane Autothermal Reforming over

Subscriber access provided by UNIVERSITY OF THE SUNSHINE COAST

Article

Experiments and modeling of methane autothermal reforming over structured Ni-Rh-based Si-SiC foam catalysts Mathilde Luneau, Elia Gianotti, Nolven Guilhaume, Emmanuel Landrivon, Frederic C. Meunier, Claude Mirodatos, and Yves Schuurman Ind. Eng. Chem. Res., Just Accepted Manuscript • DOI: 10.1021/acs.iecr.7b01559 • Publication Date (Web): 21 Jul 2017 Downloaded from http://pubs.acs.org on July 29, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Industrial & Engineering Chemistry Research is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

Experiments and modeling of methane autothermal reforming over structured Ni-Rh-based Si-SiC foam catalysts

Mathilde Luneau, Elia Gianotti, Nolven Guilhaume, Emmanuel Landrivon, Frédéric C. Meunier, Claude Mirodatos, Yves Schuurman* Université de Lyon, Université Claude Bernard Lyon 1, CNRS, IRCELYON - UMR 5256, 2 Avenue Albert Einstein, 69626 Villeurbanne Cedex, France *Corresponding author: [email protected]; Tel: (33) 472 445 482; Fax: (33) 472 445 399

Abstract Novel silicon infiltrated silicon carbide (Si-SiC) foams coated with a 10-0.3 wt.% Ni-Rh/MgAl2O4 catalyst were studied for the autothermal reforming of model biogas for the production of fuel cell hydrogen. Kinetic studies were performed by varying the inlet concentration of methane, water, carbon dioxide as well as the temperature. Despite the very good heat conductivity of the SiC structured support, it was not possible to operate the reactor at isothermal conditions due to the succession of fast reactions that were, firstly, strongly exothermic and, secondly, strongly endothermic. The experimental data were compared to a reactor model taking explicitly the heat balance in the gas and solid phase into account. The reforming kinetics were based on the methane steam reforming model by Xu and Froment [Xu, J.; Froment, G.F. Methane steam reforming, methanation and water-gas shift: 1. Intrinsic kinetics. AIChE J., 1989, 35, 88-96] complemented by a simple equation for methane oxidation. The kinetics were adapted by introducing an oxygen adsorption term to describe the consecutive complete methane oxidation and reforming reactions, in line with previous observations of two separate zones in the reactor. The model gives an adequate description of the data, giving the correct trends of the reaction orders of the reactants.

Keywords: Biogas; ceramic sponges; tri-reforming; modeling; kinetics; hydrogen 1 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 28

1. Introduction Producing hydrogen from renewable resources such as biogas to supply fuel cells is an important way for generating electricity with high energy efficiency and reduced CO2 emissions. Biogas is produced by anaerobic digestion of biomass and is primarily composed of methane and carbon dioxide. Biogas can be reformed into hydrogen-rich syngas for fuel-cell applications. Rather than removing the carbon dioxide before the reforming step, the biogas mixture can be used directly, provided its impurities have been removed (H2S and organosulfur compounds, siloxanes, halogenated compounds [1]). Dry reforming, the reaction between methane and carbon dioxide, is slow compared to steam reforming and the main effect of the presence of carbon dioxide in the feed is on the Water-Gas Shift equilibrium. In this study, autothermal reforming (ATR) of model biogas was performed. Previous screening of catalysts for this reaction led to the selection of a Ni-Rh/MgAl2O4 catalyst, which showed a stable performance over more than 300 hours [2]. Ceramic foams are cellular materials made of interconnected struts, resulting in irregular structures presenting high external surface areas and high porosities. These properties makes those attractive as catalyst supports since these structures present a lower pressure drop than traditional beds made of packed catalyst powders or pellets. In contrast to refractory ceramic foams such as cordierite or mullite, metallic and SiC foams also exhibit a high thermal conductivity, which improves heat transfer. They are thus expected to minimize hotspots and prevent mechanical-strength and thermal-shock limitations [3]. It is a critical factor for the ATR reaction, which combines both highly exothermic and endothermic reactions. One drawback of SiC materials is their sensitivity to oxidation at high temperature. For this reason, liquid siliconinfiltrated SiC (Si-SiC) foams were used as catalyst supports in the present study [4]. Compared to standard SiC, they exhibit improved oxidation resistance at high temperature and are particularly suitable as porous burners. Ceramic open-cell foams are promising catalyst supports for small size ATR units. They have excellent thermal characteristics and allow high gas flows with low pressure drops, permitting a very compact reformer. For a better reformer design a kinetic model of the biogas ATR reaction is highly desirable. Several kinetic models for methane reforming are available in the literature but only a few models exist for methane ATR [5, 6, 7, 8, 9]. In the present study, an existing kinetic model [6] was further developed for the ATR of model biogas over silicon infiltrated silicon carbide (Si-SiC) foams coated with a 10-0.3 2 ACS Paragon Plus Environment

Page 3 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


wt.% Ni-Rh/MgAl2O4 catalyst, to better comply with the experimental observations. Operating parameters such as temperature, gas hourly space velocity as well as steam/CH4, CO2/CH4 and O2/CH4 ratios were varied, in order to gain information on the kinetics of the reaction.

2. Experimental part 2.1. Catalysts preparation The magnesium aluminum spinel powder (MgAl2O4) was prepared by co-precipitation. An aqueous solution containing Mg nitrate (0,33 M) and Al nitrate (0,66 M) was added dropwise to an aqueous solution of (NH4)2CO3 containing an excess of CO32- ions of 40% with regard to the total amount of NO3- ions. The slurry was aged at 60°C under stirring for 4 h. After centrifugation and washing with de-ionized water, the powders were dried overnight at 120°C, then calcined at 800°C for 4 h in air. This powder was suspended in water and the slurry was dip-coated on the Si-SiC foams, followed by calcination in air for 4 h at 800°C. The washcoated foams were sequentially impregnated with Ni nitrate and Rh nitrate, each impregnation followed by a calcination at 550°C (4 h in air) in order to decompose the nitrates. Two silicon infiltrated silicon carbide, foam monoliths [4] (Erbicol, SA Switzerland) with a common diameter of 25 mm and a height of 25 mm or 14 mm ("entire" or "half" monolith foam, respectively) were used for the catalytic tests. The geometry of these cellular materials can be characterized by their porosity, by the average pore and cell diameters (dp and dc) and by the strut thickness (ts) [3]. Table 1 lists the geometrical properties of the Si-SiC foams used as catalyst supports in the present study.

Table 1: Properties and characteristics of the Si-SiC foams Density (ρ)

2.8 g/cm3

Pores per inch (PPI)

10

Cellular porosity (ɛ)

85 vol.%

3 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Pore diameter (dp)

3 mm

Cell diameter (dc)

6 mm

Strut thickness (ts)

1.5 mm

Page 4 of 28

The catalyst mass deposited on the foams varied from 0.23 ± 0.034 g for entire foams and 0.147 ± 0.016 g for half foams.

2.2. Catalytic tests ATR catalytic tests were performed at a pressure of 1.8 bar in a quartz fixed bed reactor (i.d. = 25.4 mm). Pre-reduction of catalysts was carried out in-situ at 700°C in a mixture of 20% H2 in Ar (total flow 300 mL/min) for 1 h. In standard ATR tests, the reactant mixture was composed of 42% H2O, 14% CH4, 9% CO2, 7% O2 and 28% Ar. This composition corresponded to the reaction of a model biogas (composed of 60% CH4 and 40% CO2) with O2 and steam in accordance with the ratios O2/CH4 = 0.5 and H2O/CH4 = 3. The Ar concentration in the inlet stream corresponded to the amount of N2 that would be present if oxygen was supplied as air. The reaction mixture composition falls outside the flammability range of a methane/oxygen mixture diluted with inert Ar and steam. However, this mixture combined with high reaction temperatures might become flammable in case of failure of the mass flow controllers leading to maximum opening for CH4 or O2 flows, or to a stop of the Ar or H2O flows. Therefore, temperature and pressure sensors were connected to automated valves and controllers in order to stop the methane and oxygen flows and all heating devices in case of alarm set-points overrunning or electrical dysfunction. Online analysis of the effluent gases was performed with a mass spectrometer (MS) using the Ar signal as internal standard for quantification. The use of Ar as internal standard also allowed compensating for the gas expansion resulting from the reactions stoichiometry. The ion current intensities were converted to molar flows knowing the molar flow rate of Ar and the sensitivity coefficient of each m/z fragment relative to Ar, which is constant. The MS response was calibrated using different partial pressures of all relevant compounds diluted in argon. The liquid water flow rate was precisely controlled by a calibrated HPLC pump (Shimadzu LC-20AD) and dispensed in an evaporator kept at 200°C. All gas lines and


Page 5 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


valves were heated at 140°C to prevent steam condensation. The GHSV was calculated using the external volume of the cylindrical foams (including empty spaces) as the catalyst volume.

2.3. Reactor modeling The experiments for the reactor modelling were run over the smaller foam of 14x25 mm and at a much higher GHSV of approximately 28,000 h-1, in order to reach lower methane conversion levels. Observing the effect of the space time on the conversion and selectivity is interesting when studying the kinetics of a reaction. However, in the case of methane ATR, the variation of the GHSV led to important variations of the temperature profile, due to the strong exothermicity of the combustion of methane at the reactor inlet. Therefore, instead of studying the effect of the variation of one parameter on the conversion, we could actually only monitor the combined effect of the simultaneous variation of two parameters. The interpretation was thus less straightforward and could be misleading. For this reason, the GHSV was kept constant for kinetic studies. The operating conditions are listed in Table 2. The set of reaction conditions involved the variation of H2O/CH4, CO2/CH4 and O2/CH4 molar ratios as well as the temperature. To assess the catalyst stability and ensure meaningful comparisons, the catalyst was periodically re-tested under standard conditions and reduced under H2 at 700°C before each new test series. These tests showed that the catalyst was stable during the whole testing period. Table 2: Operating conditions of the kinetic study H2O/CH4 O2/CH4 CO2/CH4 Temperature GHSV (h-1)

Range 1.1 – 3 0.32 -0.53 0.14 - 0.77 525 – 620°C 28,000

Standard 3 0.5 0.64 28,000

A set of 62 reaction conditions was used. The methane conversion ranged from 49 to 81%. The H2/CO outlet ratio was always close to thermodynamic equilibrium. The approach to equilibrium for the combination of reactions (1)-(3), β, ranged from 0.77 to 1.22. The expected



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 28

values should be below 1 and deviations might be due to the accuracy of the product analysis as well as the outlet temperature measurement.

2.3.1. Reactions The ATR of methane involves many reactions. Only the most prevailing reactions have been taken into account in the model. Xu and Froment studied the kinetics of methane steam reforming over Ni supported on MgAl2O4 [10], and Hou and Hugues performed a similar study over Ni/α-Al2O3 [11]. Both studies proposed that the steam reforming process could be described using three reactions only: two reactions of steam reforming and the water gas shift (WGS): CH4 + H2O CO + 3H2

(1)

CH4 + 2H2O CO2 + 4H2

(2)

CO + H2O CO2 + H2

(3)

In the case of ATR the complete oxidation of methane into carbon dioxide and water also needs to be included: CH4 + 2O2 → CO2 + 2H2O

(4)

Two different sets of rate equations were used. The first set was adapted from the study by Halabi et al. [6] who combined the methane steam reforming kinetic model of Xu and Froment [10] (reactions 1-3) with the methane total oxidation kinetic model of Trimm and Lam [12], modified to comply with a nickel catalyst rather than a platinum one. A very small amount of hydrogen (0.01%) was assumed to be present in the feed in the model to avoid that the rate of reactions (1)-(3) becomes infinite. The kinetic rate equation for the oxidation step was simplified by considering only the first term of the original equation, after some initial simulations validating this simplification [6]. This model assumes that reforming and total oxidation take place in parallel, but over different surface sites. The equations are given in Table 3. The rate constants and equilibrium constants are given in Table 4.


Page 7 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The second set of rate equations, (5)-(7) & (11) based on Ω2, was based on the observation in a recent study [24] that oxygen competes with methane for nickel sites, but transforms these sites into nickel oxide, which is inactive for reforming and WGS reactions. These results were in line with the observations of Dissanayake et al. [13]. This phenomenon results in two distinct reaction zones inside the reactor. The first will be a zone where only total combustion takes place over nickel oxide until the oxygen is depleted, the second zone consisting of reduced nickel sites catalysing steam reforming and WGS. This separation of oxidation and reforming reactions has consequences for the reactor operation. One can expect a hot spot followed by a cold spot. Hydrogen production will start later in the reactor, inside the second zone. De Groote and Froment [14] addressed this issue for the partial oxidation of methane by multiplying rate equations (5) – (7) by the fractional oxygen conversion raised to the power of 12. In contrast, we have modelled this effect by including a competition between the adsorption of oxygen and methane into the rate equations. To accomplish this an adsorption term for oxygen in the denominator of all rate equations was included and the rate equation for the oxidation step, r’4 (equation (11)), was changed accordingly. For this model the denominator, Ω2, is used as reported in Table 3. From the simulation it was found that the rate of methane oxidation was completely limited by oxygen transfer to the catalyst surface. This was also observed by Maestri et al. [REF] for methane oxidation over rhodium. They used kinetic rate equation based only on the concentration of methane and not oxygen. Considering that the methane partial pressure is not impacted significantly by the oxidation reaction due to the low amount of oxygen, the rate equation for the methane oxidation can be reduced to constant value. No differences between this simple rate equation and the Langmuir-Hinshelwood equation was observed on the overall model performance, indicating that indeed oxygen mass transfer dominates completely the oxidation rate. The consumption rate of each component was calculated by multiplying the rates (5) – (8) by the corresponding stoichiometric coefficient. Halabi et al. [6] used fixed efficiency factors for all reaction steps. Calculating the Thiele modulus showed that the efficiency factor for reactions (5) – (7) was close to 1, while it was very low for the oxidation reactions (4) or (11). The high efficiency factor for the reforming reactions is due to the thin catalyst layer, which is more than two orders of magnitude smaller than that in the study by Xu and Froment [7], who found efficiency factors around 0.05. Therefore no diffusion limitation was considered for reaction (5) – (7). The efficiency factor for reaction (4) was calculated by first calculating the Thiele modulus 7 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 28

assuming a pseudo first order reaction with respect to oxygen and then estimating the efficiency factor from =

.

Table 3. Sets of rate equations used for the two tested models. Rate equations (1)-(3) with ω1 were originally proposed in [7]. Reaction rate equations =

. −

Ω

Eq. (5)

/!"#$ /%

( −

) = /!"#$ /% Ω = ( =

( ( − ) .

)1

Ω

(

+

+

,²

(6)

(7)

/!"#$ /%

/!"#$ /%

(8)

. = 1 + + + + ( ⁄ )

(9)

(0 = (0 /!"#$ /%

(11)

. = 1 + + + + ( ⁄ ) +

(10)

Table 4. Reaction equilibrium constants and Arrhenius kinetic parameters (taken from reference [6], except k’4 and KO2 determined by this study). Equilibrium constant

parameter value

Kp2

exp(-26830/T + 30.114) bar2

Rate constant

pre-exponential factor

activation energy

1.17 1015 mol bar0.5 kgcat-1 s-1

240.1 kJ/mol

Kp1 Kp3 k1 k2

exp(4400/T – 4.036) KP1*KP2 bar2

5

5.43 10 mol bar

-1

kgcat-1 s-1

67.1 kJ/mol 8

ACS Paragon Plus Environment

Page 9 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


k3 k4

k'4

Sorption equilibrium constant KCH4 KCO

KH2

KH2O

KCCH4 KCO2 KO2

2.83 1014 mol bar0.5 kgcat-1 s-1

243.9 kJ/mol

8.11 105 mol bar-2 kgcat-1 s-1

86.0 kJ/mol

1.0 102 mol kgcat-1 s-1

0.0 kJ/mol

pre-exponential factor

heat of adsorption

6.65 10-4 bar-1

-38.3 kJ/mol

8.23 10-5 bar-1

-70.7 kJ/mol

6.12 10-9 bar-1

-82.9 kJ/mol

1.77 105

+88.7 kJ/mol

1.26 10-1 bar-1

-27.3 kJ/mol

7.78 10-7 bar-1

-92.8 kJ/mol

-7

-400 kJ/mol

1.0 10 bar

-1

2.3.2. Mass, energy and momentum balance equations Experiments for kinetic modelling are ideally performed in the absence of heat and mass transfer limitations [15]. The presence of gradients can a priori be taken into account in a reactor model, but will induce uncertainties as the transport phenomena are for most part described by empirical correlations. In this case, however, it was not possible to reduce the size of the foam, especially the diameter, to avoid temperature gradients in the reactor. Therefore the temperature profile over the catalyst bed was calculated by integrating the heat balance and using appropriate correlations, developed for foam catalysts, for the heat transfer coefficients. Similarly, mass transfer limitations occurred and were taken into account by the reactor model. A 1-D heterogeneous model was used to simulate the laboratory reactor operation and to compare it to the experimental data. The model allowed to verify various kinetic schemes and to evaluate the influence of the foam properties on the reactor performance. The model was rather similar as that used by Halabi et al. [6] for the analysis of a fixed bed autothermal reformer. The major model assumptions are the following: 1. ideal gas behaviour; 2. plug flow conditions;



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 28

3. due to the good heat conductivity of the foam, axial heat conduction was taken into account; 4. no radial concentration and temperature gradients; 5. no temperature gradient inside the catalyst layer; 6. no homogeneous reactions; 7. negligible pressure drop over the foam; Governing equations The mass and energy balance equations for the gas and solid phase are presented in Table 5. The mass balance in the gas phase is based on the convection diffusion equation. It is based on the mass fraction of the components to incorporate directly the change in moles due to reaction. Internal mass transfer is taken into account via efficiency factors, as discussed above. The heat balance in the gas phase takes into account external heat transfer and heat transfer at the wall. External heat transfer is very often encountered in fixed bed reactors and therefore the energy balance for the solid phase is explicitly taken into account. This balance also contains a term for the axial heat conductivity because of the very good heat conductivity of the foam.

Table 5. Mass and energy balances Mass and energy balances in gas phase ; ;

Experiments and Modeling of Methane Autothermal Reforming over

Recommend Documents