- Email: [email protected]
Steam reforming of tars at low temperature and elevated pressure for model tar component naphthalene
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 9 2 0 e1 2 9 2 8
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Steam reforming of tars at low temperature and elevated pressure for model tar component naphthalene Michael Speidel*,1, Holger Fischer German Aerospace Center, Institute of Engineering Thermodynamics, Pfaffenwaldring 38-40, 70569 Stuttgart, Germany
article info
abstract
Article history:
A process of pressurized gasification and power generation in a hybrid system of Solid
Received 26 January 2016
Oxide Fuel Cell (SOFC) and gas turbine enables an efficient use of biomass. This process
Received in revised form
requires tar reforming in order to protect the SOFC from plugging. Tars must be converted
15 April 2016
at 5 bar absolute pressure (bara) while avoiding secondary steam reforming of methane in
Accepted 4 May 2016
order to reduce the required heat input for the tar reformer. This can be realized at low
Available online 2 June 2016
reforming temperatures (<700 C) where methane conversion is reduced due to chemical equilibrium. A laboratory-scale test rig is introduced, which enables an investigation of the
Keywords:
steam reforming of the model tar component naphthalene at up to 5 bara. Deactivation of
Biomass
the nickel catalyst caused by coke formation was detected. Despite the reduced amount of
Pressurized gasification
free active centers on the catalyst surface, stationary naphthalene conversions are possible
SOFC
at temperatures between 600 C and 700 C. The lower the temperature, the more active
Tar reforming
centers are covered. For stationary conditions a hyperbolic approach for the reaction rate
Naphthalene
of steam reforming of naphthalene is developed and parameters for 650 C and 700 C are
Coke formation
determined. © 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
Introduction In biomass gasification, renewable raw materials or residues can be used to produce electrical energy or fuels. The predominantly hydrogen-rich product gas of the absorption enhanced reforming (AER-) process [1e5] can be converted into electricity in gas engines, gas turbines or fuel cells. A further possibility is to convert the product gas catalytically into gaseous fuels like pure hydrogen or synthetic natural gas
(SNG) as well as liquid fuels using such methods as the FischereTropsch process. Advantages of using a pressurized gasifier are its smaller volume, as well as its compatibility with pressurized downstream applications like methanation or conversion in a gas turbine without the need to cool the gas for compression. Product gases of any gasifier contain impurities such as particles, chlorine, sulfur and tars [6], which must be removed before the power generation or fuel synthesis. The present work focuses on tar removal, for which tars are defined as all undesired higher hydrocarbons of the
* Corresponding author. E-mail address: [email protected] (M. Speidel). 1 I did this work at German Aerospace Center, but now I do not work there any more. Yet my former colleague Holger Fischer still works there. http://dx.doi.org/10.1016/j.ijhydene.2016.05.023 0360-3199/© 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 9 2 0 e1 2 9 2 8
12921
Fig. 1 e Concept of pressurized gasification of biomass for power generation in a hybrid solid oxide fuel cell (SOFC) and gas turbine system.
gasification gas [7]. More specifically, the catalytic steam reforming of tars is investigated [8,9]. Since the use of catalytic additives in the gasifier bed material [10,11] or in ceramic filters [12] is not sufficient to achieve the required tar conversions of over 99% for power generation in fuel cells, tar reforming in a fixed bed downstream from the gasifier is considered. As can be seen in Fig. 1, the pressurized product gas of the AER-process is cleaned and then converted in a hybrid power generation system consisting of a Solid Oxide Fuel Cell (SOFC) and a gas turbine [13e19]. The AER-process consists of two reactors. In the first one biomass is gasified with steam and in the second one, combustion of residual coke takes place, which provides the necessary heat for the gasification. CO2 is bound at the bed material CaO and therefore the AER-product gas contains small concentrations of CO2 and high concentrations of H2. The product gas composition can be seen in Table 1.
Table 1 e Simulated AER product gas composition at 5 bara in mole fraction and typical tar yield of AER product gas referred to dry gas [20]. yH2 O
yH2
yCO
yCH4
yC2 H4
yC2 H6
yC3 H8
yCO2
Tar
60%
26%
2.4%
5.6%
1.2%
0.6%
1.2%
3.0%
1e5 g/m3N
Particles, chlorine and sulfur impurities can be removed at temperatures greater than 600 C [6] (see Fig. 1). Afterwards tars are reformed catalytically by using the steam contained in the gas (see Eq. (1) in Fig. 1). The reaction takes place in a temperature range of 600 Ce700 C and at a pressure of 5 bar absolute pressure (bara). The advantage of a low reformer temperature (<700 C) is a reduction in the secondary steam reforming of methane (see Eq. (2) in Fig. 1) due to chemical equilibrium and therefore less heat demand for that endothermic reaction. In the case of the pressurized AER product gas shown in Table 1, at approximately 650 C no methane will be converted in the tar reformer at 5 bara. The challenges of lower temperatures, however, are the slower reaction kinetics of tar reforming and the risk of coke formation at the nickel catalyst, especially at pressures higher than atmospheric. Therefore a detailed understanding of tar reforming reaction kinetics is required for the design of that process unit.
A laboratory-scale test rig with a fixed catalytic bed was built to investigate the reaction kinetics experimentally. Naphthalene (C10H8) is used as the model tar component because it is the main component of tars in AER product gas (mass fraction of 31.5% of tars heavier than toluene [2]), it is often an intermediate product at the tar reforming process and its reaction rate is low in comparison to other tars. Jess [21] found that naphthalene as a model tar is strongly adsorbed on the surface of a nickel catalyst and inhibits the simultaneous conversion of methane and benzene. Due to this inhibition at higher concentrations the reaction order of steam reforming of naphthalene itself is, at 0.2, much lower than 1. Therefore Depner [22] uses a hyperbolic EleyeRideal approach for steam reforming of naphthalene on a nickel catalyst. This approach is based on adsorption of naphthalene and includes inhibition of the reaction due to active centers on the catalyst surface being covered by naphthalene. While Jess [21], Depner [22] and current works of Hamel et al. [23], Kaisalo et al. [24] and Kurkela et al. [25] investigate tar reforming at temperatures higher than 700 C, Fraubaum et al. [26] and Kienberger et al. [27] focus on temperatures even lower than 500 C with combined methanation. The aims of this work are to demonstrate stable steam reforming of naphthalene in a temperature range of 600 Ce700 C, to derivate an approach for the conversion rate of naphthalene, and to determine the kinetic parameters for this approach.
Experimental An average molar gas composition of the technical tar reformer in Fig. 1 of 45%e55% hydrogen and 35%e37% steam is synthesized as feed for the reformer at the test rig (see Fig. 2). Further gas components such as carbon monoxide (CO), carbon dioxide (CO2), methane (CH4) and other
Fig. 2 e Schematic draft of the test rig.
12922
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 9 2 0 e1 2 9 2 8
Fig. 3 e The pressurized tar saturator (left) and the catalytic fixed bed reformer heated by an oven (right).
hydrocarbons (see Table 1) are not considered for the reaction of tar with steam and are replaced by 10%e18% nitrogen. Controlled introduction of naphthalene into the gas is carried out in the saturator via a continuous flow of nitrogen. The mixed gas can be analyzed either before or after the reforming reaction. Hydrocarbons are detected discretely in the wet gas by a gas-phase chromatograph (GC) with a flame ionization detector (FID), whereas CO, CO2 and CH4 are detected continuously in the dry gas by infrared spectroscopy (IR).
Saturator The left side of Fig. 3 shows the pressurized saturator, which is filled with solid naphthalene at room temperature. Naphthalene is then melted by heat input from a heating jacket Hliq, which controls the temperature of the molten naphthalene Tliq. Preheated nitrogen of the same temperature flows through the molten naphthalene and gets saturated. A filter at the end of the inlet tube ensures small bubbles, which have a longer retention time and better mass transport properties. The saturator works at 10 bara, which is controlled with a mechanical pressure controller (PC). A second heating jacket Hgas controls the gas temperature Tgas in the saturator. It was experimentally found that stable, reproducible operation of the saturator is achieved when the naphthalene fill level is less than half and the gas temperature is approximately 4e5 K higher than the liquid temperature. It can be observed that the nitrogen stream is fully saturated with naphthalene. Due to the oscillating gas and liquid temperatures in the saturator, the concentration of naphthalene in the gas also oscillates with a relative standard deviation of 1%e1.5%. To minimize deviation, the measured concentrations of naphthalene are averaged over time.
Reformer The reformer contains a catalytic fixed bed (see right side of Fig. 3) with alumina pellets of 2.3 mm diameter, which are
impregnated with nickel and supplied by Johnson Matthey. The properties of the catalyst pellets and of the fixed bed are given in Table 2. The fixed bed section is housed in an electrical oven, which controls the temperature at the outer wall of the tube at the inlet of the catalytic bed (TW1 in Fig. 3). An inert fixed bed of alumina pellets is included above the catalytic bed and separated by glass wool. Thus, heat transfer from the tube to the gas is improved and a uniform flow field and an equal temperature of gas and pellets can be assumed from the inlet of the catalytic bed downwards. The gas is pre-heated before entering the oven to ensure that the gas temperature in the catalytic bed TR1 (see Fig. 3) is the same as the tube temperature TW1. Maximum temperature deviation in the catalytic bed due to the endothermic reaction (TR1TR2) is lower than 5 K in the experiments presented in this work. The arithmetic mean of TR1 and TR2 is used as the reaction temperature.
Gas-phase chromatograph The flame ionization detector (FID) of the gas-phase chromatograph detects the hydrocarbons in the gas sample.
Table 2 e Properties of the catalyst pellets and of the fixed bed. Manufacturer Active material Carrier material Manufacturing method Pellet diameter Pellet density Pellet porosity BET-surface Active surface Median pore diameter Pore shape Fixed bed length Fixed bed density Fixed bed porosity
Johnson Matthey Nickel (15% mass fraction) Alumina (gamma) Impregnation 2.3 mm 1380 kg/m3 34.5% 64 m2/g 3.5 m2/g 13.4 nm cylindrical 20 mm 587 kg/m3 57.5%
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 9 2 0 e1 2 9 2 8
Reformate gas flows continuously through a sampling loop with a specific volume at a rest temperature of 200 C and a pressure of approximately 1 bara. Every 18.5 min, the sample gas in the loop is carried to the column by the hydrogen carrier gas. The gas-phase chromatograph is calibrated by a naphthalene solution, which is vaporized in the injector and then carried to the column. The calibration is verified by the mass loss of the saturator. More sample gas is carried to the column with higher pressure in the sampling loop. Therefore the measured gas-phase composition is adjusted by the average sample pressure.
Methodology The nickel catalyst is activated by reducing nickel oxide to nickel at 850 C with 50 vol% hydrogen (rest nitrogen) for half an hour after being heated up with pure nitrogen. In order to find out whether half an hour is sufficient for the activation process methane steam reforming is carried out, which reaches stationary conditions and methane conversion within minutes. The activation process and methane reforming is repeated until constant methane conversion indicates complete activation of the catalyst. The whole activation process is done once after which several naphthalene steam reforming tests can be carried out. The reformer is heated up with nitrogen before the test. Shortly before opening the tar saturator, steam and hydrogen are switched on. After the test the saturator is closed, then steam and hydrogen are switched off and the reformer is either cooled down with nitrogen or the temperature is maintained by a small volume flow of nitrogen until the next test begins. Furthermore the weight loss of the naphthalene in the saturator is determined after every test in order to confirm the concentration measurements in the gas-phase chromatograph or to correct them. For naphthalene concentrations below 300 ppmv it is assumed that total mole flow is constant. Therefore relative naphthalene conversion can be calculated based on molar fractions ynaph: Xnaph ¼
out yin naph ynaph
(3) yin naph By measuring the naphthalene concentrations upstream and downstream of the reformer the relative naphthalene conversion can be calculated. All experiments are carried out at 5 bara and a feed volume flow of 545 lN/h, which corresponds to a gas hourly space velocity of 107,000 1/h. With maximum naphthalene concentration of 300 ppmv, 45%e55% hydrogen and 35%e37% steam, as explained at the beginning of Chapter 2, there is a high excess of hydrogen and steam in comparison to naphthalene (minimum ratio of 1500 in the case of hydrogen and 1167 in the case of steam).
Results and discussion In this chapter the experimental results of the steam reforming of naphthalene at the laboratory-scale test rig are introduced. First in Chapter Transient progression of the naphthalene conversion the transient progression of naphthalene conversion is examined before the stationary conversions at temperatures between 600 C and 700 C are
12923
shown in Chapter Stationary experimental naphthalene conversion. A kinetic approach for stationary conversion is subsequently derived in Chapter Derivation of a kinetic approach of naphthalene conversion rate. In Chapter Determination of kinetic parameters the kinetic parameters for this kinetic approach are determined.
Transient progression of the naphthalene conversion Fig. 4 shows several tests at different temperatures which are carried out sequentially. All tests are performed with a naphthalene feed concentration of 130 ppmv. The first experiment (see Fig. 4 (a)) is carried out at 655 C with a fresh catalyst, which has been reduced with hydrogen (see Section Methodology). High naphthalene conversion of more than 80% decreases within several hours and reaches a stable stationary conversion of approximately 50%. Due to the findings of Jess [21] and Depner [22] that strong naphthalene adsorption reduces the amount of free active centers, the deactivation of the nickel catalyst is explained by the active centers on the surface of the catalyst becoming covered. Yet there seem to be some uncovered centers which enable stationary naphthalene conversion. The next test (see Fig. 4 (b)) is carried out with the same now used catalyst at 640 C starting at a conversion of approximately 50% before the catalyst is further deactivated and reaches a stationary conversion of approximately 40%. The test thereafter is carried out at 620 C with the same catalyst. The results in Fig. 4 (c) show the repetition of this test at 620 C. The subsequent test at 600 C (d) shows the same characteristics as the test at 640 C (b). The conversion starts high before the catalyst deactivates further and reaches a stationary conversion of approximately 10%. Repeating the test at 655 C (e, f) after the test at 600 C indicates that the catalyst is partially reactivated after several hours. The conversion of naphthalene starts at approximately 30%. It then increases before reaching the same reproducible stationary conversion of approximately 50% as in the first test (a). It is concluded that the deactivation of the catalyst is caused by active centers on the surface being covered. Furthermore, it is concluded that more active centers are covered at lower temperatures. This can be caused by adsorption of naphthalene as described in Refs. [21] and [22]. Yet the long deactivation times indicate that coke formation can also be a reason for deactivation. The outlet concentrations for a steam reforming test of naphthalene at 5 bara and 650 C can be seen in Fig. 5. The concentrations in Fig. 5 refer to the amount of carbon atoms in order to facilitate the comparison of higher hydrocarbons, in this case naphthalene (C10H8) and benzene (C6H6). The feed concentration of the reformer is approximately 165 ppmv naphthalene respectively 1650 ppmv carbon atoms. Using steam, naphthalene is converted to CO and H2 (see Eq. (1) in Fig. 1). Secondary reactions of CO (CO-Shift and methanation) lead to CO2 and H2 or methane and steam. Due to the high hydrogen feed concentration of 55 vol% and the small difference between feed concentration and outlet concentration, hydrogen is not measured. Benzene (C6H6) is the only detectable by-product of naphthalene, which is subsequently also converted with steam to CO and H2. The appearance of benzene as a by-product indicates that naphthalene cracking also takes
12924
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 9 2 0 e1 2 9 2 8
Fig. 4 e Naphthalene conversions (calculated using Eq. (3)) for several tests at different temperatures for 5 bara and 130 ppmv naphthalene feed concentration.
Fig. 5 e Reformate gas composition during a steam reforming test with 165 ppmv naphthalene feed concentration at 5 bara and 650 C.
place. The concentration of naphthalene rises while the sum of the carbon-containing products (methane, CO, CO2 and benzene) declines. Thus, the conversion of naphthalene falls and does not reach stationary conditions within 7 operating hours. Additionally, conversion rates of the secondary reactions of CO also fall, as evidenced by the decreasing concentrations of the CO2 and methane products. Covered active centers on the surface of the catalyst inhibit all of these reactions. Hydrocracking of naphthalene also leads to naphthalene conversion and formation of methane, but is assumed to play a tangential role compared to steam reforming of naphthalene. At the test rig, methane steam reforming was also investigated. It was found that following naphthalene steam reforming test, the methane conversion at the nickel catalyst is significantly inhibited. This indicates that the active centers are covered not only by naphthalene adsorption but also by coke formation, which has long-term effects on catalyst activity. It was found that
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 9 2 0 e1 2 9 2 8
12925
the coke canbe gasifiedbyaddition of steam at high temperatures and the activity of the catalyst can be restored. The results of hydrogen-based laboratory chemisorption measurements performed on the used catalyst following tests with a 130 ppmv naphthalene feed concentration and 650 C and 700 C confirm a significantly less active catalyst surface compared to the fresh catalyst.
Stationary experimental naphthalene conversion Fig. 6 contains the stationary conversions of naphthalene for several tests at temperatures between 600 C and 700 C. For naphthalene feed concentrations of approximately 130 ppmv, stable stationary naphthalene conversions are reached for temperatures as low as 600 C. A strong temperature dependency is nevertheless observed in the temperature range studied. It is assumed that the low naphthalene conversions at low temperatures are caused by the covered active centers in particular.
Fig. 7 e Schematic diagram of reaction steps on the catalytic surface.
rnaph ¼ knaph $qnaph
where knaph is the reaction rate constant. The adsorption rate is proportional to the partial pressure of naphthalene and the fraction of uncovered active centers. The desorption rate is proportional to the centers covered with naphthalene. Equilibrium of adsorption and desorption is expressed by: kads $ 1 qcoke qnaph $pnaph ¼ kdes $qnaph Kads $ 1 qcoke qnaph $pnaph ¼ qnaph
Fig. 6 e Stationary naphthalene conversions for several tests at temperatures between 600 C and 700 C at 5 bara and 130 ppmv naphthalene feed concentration (compare tests in Fig. 4).
The aim of this work is to describe the stationary part of naphthalene steam reforming. Therefore a kinetic approach is derived that takes into consideration the effects of covered active centers by adsorbed naphthalene and coke.
Derivation of a kinetic approach of naphthalene conversion rate Fig. 7 shows the assumed reaction steps on the surface of the nickel catalyst. Jess [21] and Depner [22] indicate that naphthalene is strongly adsorbed at the surface and therefore covers some of the active centers. Presented results show that coke formation also takes place and leads to covered active centers. As shown previously coke can be gasified with steam. After adsorption, a naphthalene molecule can be desorbed, can react with steam e the desired steam reforming reaction e or can be converted directly or indirectly to coke. The reaction rate of steam reforming of naphthalene is determined for stationary conditions. The reaction rate is proportional to the fraction of the surface covered with naphthalene and is independent of steam concentration as there is an excess of steam in the feed gas:
(4)
(5)
where qcoke is the fraction of surface covered with coke, kads the adsorption rate constant and kdes the desorption rate constant for naphthalene. The adsorption equilibrium constant Kads is the ratio of the adsorption and desorption rate constants. Catalytic coke formation is assumed to be first order in naphthalene [28]: rcoke ¼ kcoke $qnaph The reaction rate of coke gasification with steam is assumed to be proportional to the fraction of the surface covered with coke and independent of steam as there is an excess of steam in the gas: rgas ¼ kgas $qcoke The quasi-stationary state of coke formation and coke gasification can be calculated to kcoke $qnaph ¼ kgas $qcoke qcoke ¼ Kcoke $qnaph
(6)
where Kcoke is the ratio of coke formation and gasification rate constants kcoke and kgas. Substituting qcoke from Eq. (6) into Eq. (5) results in: qnaph ¼
Kads $pnaph 1 þ Kads $pnaph þ Kads $Kcoke $pnaph
Substituting qnaph into Eq. (4) yields for the reaction rate of steam reforming: k0naph
rnaph
zfflfflfflfflfflffl}|fflfflfflfflfflffl{ knaph $Kads $pnaph ¼ 1 þ Kads $ð1 þ Kcoke Þ $pnaph |fflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflffl} K0naph
(7)
12926
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 9 2 0 e1 2 9 2 8
As it is experimentally not possible to differentiate between active centers covered by adsorbed naphthalene or coke, the parameter K0naph is defined including both effects. The reaction rate constant knaph and the adsorption constant Kads are also summarized to parameter k0naph . The parameters K0naph and k0naph are only dependent on temperature. For low partial pressures of naphthalene the reaction order is 1 and for high partial pressures of naphthalene the limit value for the reaction order is 0. If hydrocracking of naphthalene contributes to naphthalene conversion and methane formation, in addition to the described steam reforming of naphthalene to CO and subsequent methanation (compare description of Fig. 5), the influence of the constant hydrogen concentration will be contained in parameter knaph of Eq. (4). Coke formation can be caused by catalytic cracking or hydrocracking. In the case of hydrocracking the impact of the constant hydrogen concentration is contained in parameter kcoke. A slightly positive effect of hydrogen on the naphthalene conversion is detected similar to the dependency described in Ref. [21] with reaction order of 0.3 for hydrogen, but not considered in this work. Strictly speaking the kinetic approach is only valid for a constant hydrogen concentration. Therefore the following experiments are carried out at constant hydrogen concentrations.
Determination of kinetic parameters Assuming constant total mole flow n_tot , the following equation describes the reduction of naphthalene in an infinitesimally small volume element with the cross sectional area AC and the length dx. n_tot $dynaph ¼ rnaph $rbed $Ac $dx
(8)
where rbed is the mass-density of the catalytic fixed bed (see Table 2). Integration from inlet to outlet for a reformer with the length L results in the following one-dimensional balance equation for naphthalene: yout naph
Z
1 rnaph
yin naph
$dynaph ¼
rbed $Ac $ n_tot
ZL dx
(9)
x¼0
Substituting the conversion rate of Eq. (7) into the molar balance of naphthalene (9) and integrating results in the following equation:
Table 3 e Parameters determined for stationary conditions of steam reforming of naphthalene. Temperature [ C] 649 700
yin 1 naph $ln out p ynaph
!
k0naph [mol/kgcat s Pa]
K0naph [1/Pa]
9.42$106 1.20$105
0.042 0.0165
1 0 out þ K0naph $ yin naph ynaph ¼ _ $knaph $rbed $Ac $L ntot
(10)
For two tests at the same temperature and different feed concentrations yin naph , these two Eq. (10) are divided. Since the right side of Eq. (10) is equal for equal temperatures, it can be eliminated. The divided equation results in: ! ! 1 yin yin 1 naph;1 naph;2 out $ln out $ln þ K0naph $ yin y naph;1 naph;1 p p ynaph;1 yout naph;2 0 in out Knaph $ ynaph;2 ynaph;2 ¼ 0
(11)
Measuring the concentrations of naphthalene upstream out and downstream from the reformer (yin naph and ynaph ) for the two tests, parameter K0naph can be determined with Eq. (11). At 650 C and at 700 C several tests are carried out at different feed concentrations. Parameter K0naph is determined with the method of least squares for Eq. (11) by combining in each case two tests at different naphthalene concentrations. Afterwards k0naph is determined with Eq. (10) for the test with the highest feed concentration. Table 3 lists the kinetic parameters determined for tests at 650 C or rather 649 C and 700 C. The higher value for K0naph at 649 C compared to 700 C confirms that at lower temperatures more active centers are covered and the catalyst is more deactivated. Fig. 8 compares the calculated naphthalene conversions to the experimental conversions. The effect of the covered catalyst surface is evident, especially at higher naphthalene partial pressures or concentrations, and can be described well with the introduced kinetic approach. The reaction in these experiments is strongly limited by pore diffusion, especially for lower naphthalene concentrations when the reaction is faster. This can be seen by the high values of the Weisz-Prater parameter [29] even for a feed concentration of 300 ppmv shown in Table 4, according to the definition of the Weisz-Prater parameter (WPP) presented in Ref. [30]. For conditions without influence of pore diffusion K0naph would be higher.
Fig. 8 e Comparison of experimental and calculated conversion rates at 649 C (left) and 700 C (right).
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 9 2 0 e1 2 9 2 8
12927
Table 4 e Weisz-Prater parameter (WPP) for 649 C and 700 C and different reaction orders n. T [ C]
ynaph [ppmv]
649 700
300 300
Dnaph [m2/s] 5
2.72$10 2.99$105
DK [m2/s] 6
1.74$10 1.79$106
The Weisz-Prater parameter is calculated as follows [30]: WPP ¼
bed 0:5$ðn þ 1Þ$rnaph $rεbed $L2c
p
naph Deff $ R$T
(12)
where n is the reaction order, rbed the density of the catalytic bed (see Table 2), εbed the porosity of the bed (see Table 2) and pnaph the partial pressure of naphthalene in the bulk phase. rnaph is the measured reaction velocity calculated using Eq. (7). Lc is the characteristic length of the pellet, in the definition of [30] the ratio of the volume of the pellet and the surface. The effective diffusion coefficient Deff is calculated by means of the molecular diffusion coefficient Dnaph, the Knudsen diffusion coefficient DK, the porosity of the pellet εP (see Table 2) and the tortuosity t as follows: Deff ¼
1 εP $ þ D1K t
1 Dnaph
Deff [m2/s]
WPP (n ¼ 0)
WPP (n ¼ 1)
1.88$107 1.94$107
3.9 10.6
7.8 21.3
Nevertheless, it can be concluded that tar reforming at temperatures as low as 650 C is possible at a pressure of 5 bara and a feed concentration of the model tar naphthalene of 300 ppmv, which corresponds to approximately 1.7 g/m3N based on wet gas. This enables tar conversion without secondary methane conversion because of chemical equilibrium limitation. The heat input required for the tar reformer can be reduced considerably, which has a positive effect on the overall process of pressurized gasification of biomass for power generation in a hybrid Solid Oxide Fuel Cell (SOFC) and gas turbine system. With inhibited methane conversion caused by coke it is also conceivable that methane conversion can be limited due to kinetic reasons even at temperatures higher than 650 C if an appropriate reactor concept is developed.
(13)
Tortuosity of the pellet is not known, but estimated with a typical value of 3. Dnaph is calculated by means of binary diffusion coefficients for a gas composition of 45% hydrogen, 37% steam and 18% nitrogen. The reaction is limited by pore diffusion at values of the Weisz-Prater parameter higher than 0.25 and strongly limited at values higher than 5 [30]. Table 4 shows that the reaction is even limited by pore diffusion at 649 C and a limit value of 0 for the reaction order.
Conclusions A laboratory-scale test rig was introduced, which enables the investigation of steam reforming of tar at pressures of up to 5 bara in a temperature range of 600 Ce700 C, as well as the determination of the reaction kinetics at stationary conditions. The model tar component naphthalene is converted with steam out of the synthesized gasification gas to CO and H2. Although the nickel catalyst deactivates, a reproducible, stationary conversion of naphthalene is reached within several hours or days. Adsorption of naphthalene and naphthalene based coke formation inhibit the conversion of naphthalene itself as well as secondary reactions of the product CO. The lower the temperature, the more the catalyst gets deactivated at stationary conditions. A hyperbolic approach for the conversion rate at stationary conditions was derived. Therefore it is assumed that active centers on the surface of the catalyst are covered by coke and adsorbed naphthalene. Furthermore, a quasi-stationary state of coke formation and coke gasification with steam is assumed. The influence of naphthalene concentration is determined for temperatures of 650 C and 700 C. The higher value of parameter K0naph for 650 C confirms that more active centers are covered at lower temperatures and the activity of the catalyst is lower.
Acknowledgments The authors thank Sandra Adelung and Dr. Ralph-Uwe Dietrich for their assistance presenting the results of this work. The Helmholtz Gemeinschaft grant within the DLR@UniST project is gratefully acknowledged. The nickel catalyst used in this work is kindly provided by Johnson Matthey.
references
[1] Pfeifer C, Puchner B, Hofbauer H. In-situ CO2-absorption in a dual fluidized bed biomass steam gasifier to produce a hydrogen rich syngas. Int J Chem React Eng 2007;5:A9. [2] Pfeifer C, Puchner B, Hofbauer H. Comparison of dual fluidized bed steam gasification of biomass with and without selective transport of CO2. Chem Eng Sci 2009;64:5073e83. http://dx.doi.org/10.1016/j.ces.2009.08.014. [3] Poboß N, Swiecki K, Charitos A, Hawthorne C, Zieba M, Scheffknecht G. Experimental investigation of the absorption enhanced reforming of biomass in a 20 kWth dual fluidized bed system. Int J Thermodyn 2012;1:53e9. [4] Koppatz S, Pfeifer C, Rauch R, Hofbauer H, MarquardMoellenstedt T, Specht M. H2 rich product gas by steam gasification of biomass with in situ CO2 absorption in a dual fluidized bed system of 8 MW fuel input. Fuel Process Technol 2009;90:914e21. http://dx.doi.org/10.1016/ j.fuproc.2009.03.016. [5] Udomsirichakorn J, Basu P, Salam PA, Acharya B. Effect of CaO on tar reforming to hydrogen-enriched gas with inprocess CO2 capture in a bubbling fluidized bed biomass steam gasifier. Int J Hydrogen Energy 2013;38:14495e504. http://dx.doi.org/10.1016/j.ijhydene.2013.09.055. [6] Aravind PV, de Jong W. Evaluation of high temperature gas cleaning options for biomass gasification product gas for solid oxide fuel cells. Prog Energ Combust 2012;38:737e64. http://dx.doi.org/10.1016/j.pecs.2012.03.006.
12928
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 9 2 0 e1 2 9 2 8
[7] Li C, Suzuki K. Tar property, analysis, reforming mechanism and model for biomass gasificationdan overview. Renew Sust Energ Rev 2009;13:594e604. http://dx.doi.org/10.1016/ j.rser.2008.01.009. [8] Ising M. Zur katalytischen Spaltung teerartiger Kohlenwasserstoffe bei der Wirbelschichtvergasung von € t Dortmund, Biomasse [Dissertation]. German: Universita Fraunhofer IRB Verlag Stuttgart; 2002. [9] Pfeifer C, Huettler H, Hofbauer H, Schulzke T, Unger C. Catalytic tar decomposition in a slip stream of an 8 MW dual fluidized bed biomass steam gasifier. In: International Symposium on Gas Cleaning at High Temperatures (GCHT); 2008. Newcastle, UK. [10] Choi YK, Cho MH, Kim JS. Air gasification of dried sewage sludge in a two-stage gasifier. Part 4: application of additives including Ni-impregnated activated carbon for the production of a tar-free and H2-rich producer gas with a low NH3 content. Int J Hydrogen Energy 2016;41:1460e7. http:// dx.doi.org/10.1016/j.ijhydene.2015.11.125. [11] Mun TY, Kim JW, Kim JS. Air gasification of dried sewage sludge in a two-stage gasifier: part 1. The effects and reusability of additives on the removal of tar and hydrogen production. Int J Hydrogen Energy 2013;38:5226e34. http:// dx.doi.org/10.1016/j.ijhydene.2012.10.120. [12] Savuto E, Di Carlo A, Bocci E, D'Orazio A, Villarini M, Carlini M, et al. Development of a CFD model for the simulation of tar and methane steam reforming through a ceramic catalytic filter. Int J Hydrogen Energy 2015;40:7991e8004. http://dx.doi.org/10.1016/ j.ijhydene.2015.04.044. [13] Methling T, Armbrust N, Haitz T, Speidel M, Poboss N, BraunUnkhoff M, et al. Power generation based on biomass by combined fermentation and gasification - a new concept derived from experiments and modelling. Bioresour Technol 2014;169:510e7. http://dx.doi.org/10.1016/ j.biortech.2014.07.036. € rner A. A new process concept for [14] Speidel M, Kraaij G, Wo highly efficient conversion of sewage sludge by combined fermentation and gasification and power generation in a hybrid system consisting of a SOFC and a gas turbine. Energy Convers Manag 2015;98:259e67. http://dx.doi.org/10.1016/ j.enconman.2015.03.101. [15] Dekker N, Rabou L, van Wees H, Rietveld B. Operation of a Staxera SOFC stack - fuelled with cleaned gas from a gasifier. In: 8th European SOFC Forum; 2008. Lucerne, Switzerland. [16] Henke M, Willich C, Kallo J, Friedrich A. Theoretical study on pressurized operation of solid oxide electrolysis cells. Int J Hydrogen Energy 2014;39:12434e9. http://dx.doi.org/10.1016/ j.ijhydene.2014.05.185. [17] Leucht F, Bessler WG, Kallo J, Friedrich KA, Mu¨llerSteinhagen H. Fuel cell system modeling for solid oxide fuel cell/gas turbine hybrid power plants, part i: modeling and simulation framework. J Power Sources 2011;196:1205e15. http://dx.doi.org/10.1016/j.jpowsour.2010.08.081.
[18] Panne T, Widenhorn A, Boyde J, Matha D, Abel V, Aigner M. Thermodynamic process analyses of SOFC/GT hybrid cycles. In: 5th International Energy Conversion Engineering Conference and Exhibit (IECEC); 2007. St. Louis, Missouri (USA). [19] Veyo S, Lundberg W, Vora S, Litzinger K. Tubular SOFC hybrid power system status. In: Proceedings of ASME Turbo Expo; 2003 June 16-19. Atlanta, Georgia, USA. [20] Steiert S. FuE-Plattform “Biomass-to-Gas” e Energetische Nutzung biogener Reststoffe mit AER-Technologie zur Poly-Generation von Strom, Wasserstoff, Erdgassubstitut € rme, Schlussbericht. Technische und Wa € tsbibliothek, Informationsbibliothek u. German: Universita Stuttgart; 2013. [21] Jess A. Catalytic upgrading of tarry fuel gases: a kinetic study with model compounds. Chem Eng Process 1996;35:487e94. http://dx.doi.org/10.1016/S0255-2701(96)04160-8. [22] Depner H. Untersuchungen zur katalytischen Umsetzung flu¨chtiger Schwelprodukte in Rohgasen der Verkokung und Vergasung fester Brennstoffe [Dissertation]. German: € t Karlsruhe; 1998. Universita [23] Hamel C. Options to reach syngas quality requirements by catalytic reforming of tars and methane from FB-gasification of biomass. In: 6th International Freiberg Conference on IGCC & XtL Technologies; 2014. Freiberg, Germany. [24] Kaisalo N, Kihlman J, Hannula I, Simell P. Reforming solutions for biomass-derived gasification gas e experimental results and concept assessment. Fuel 2015;147:208e20. http://dx.doi.org/10.1016/j.fuel.2015.01.056. [25] Kurkela E, Kurkela M, Hiltunen I. Steameoxygen gasification of forest residues and bark followed by hot gas filtration and catalytic reforming of tars: results of an extended time test. Fuel Process Technol 2016;141:148e58. http://dx.doi.org/ 10.1016/j.fuproc.2015.06.005. [26] Fraubaum M, Walter H, Zuber C. Kinetic modeling of a combined tar removal and methanation reactor for biogenous synthesis gas at medium temperature conditions. Fuel Process Technol 2016;141:159e66. http://dx.doi.org/ 10.1016/j.fuproc.2015.06.003. [27] Kienberger T, Zuber C, Novosel K, Baumhakl C, Karl J. Desulfurization and in situ tar reduction within catalytic methanation of biogenous synthesis gas. Fuel 2013;107:102e12. http://dx.doi.org/10.1016/j.fuel.2013.01.061. [28] Goud SK, Whittenberger WA, Chattopadhyay S, Abraham MA. Steam reforming of hexadecane using a catalyst: kinetics of catalyst deactivation. Int J Hydrogen Energy 2007;32:2868e74. http://dx.doi.org/10.1016/ j.ijhydene.2007.03.019. [29] Weisz PB, Prater CD. Interpretation of measurements in experimental catalysis. Adv Catal 1954;6:143e96. http:// dx.doi.org/10.1016/S0360-0564(08)60390-9. [30] Missen RW, Mims CA, Saville BA. Introduction to chemical reaction engineering and kinetics. Wiley VCH; 1999.
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Steam reforming of tars at low temperature and elevated pressure for model tar component naphthalene Michael Speidel*,1, Holger Fischer German Aerospace Center, Institute of Engineering Thermodynamics, Pfaffenwaldring 38-40, 70569 Stuttgart, Germany
article info
abstract
Article history:
A process of pressurized gasification and power generation in a hybrid system of Solid
Received 26 January 2016
Oxide Fuel Cell (SOFC) and gas turbine enables an efficient use of biomass. This process
Received in revised form
requires tar reforming in order to protect the SOFC from plugging. Tars must be converted
15 April 2016
at 5 bar absolute pressure (bara) while avoiding secondary steam reforming of methane in
Accepted 4 May 2016
order to reduce the required heat input for the tar reformer. This can be realized at low
Available online 2 June 2016
reforming temperatures (<700 C) where methane conversion is reduced due to chemical equilibrium. A laboratory-scale test rig is introduced, which enables an investigation of the
Keywords:
steam reforming of the model tar component naphthalene at up to 5 bara. Deactivation of
Biomass
the nickel catalyst caused by coke formation was detected. Despite the reduced amount of
Pressurized gasification
free active centers on the catalyst surface, stationary naphthalene conversions are possible
SOFC
at temperatures between 600 C and 700 C. The lower the temperature, the more active
Tar reforming
centers are covered. For stationary conditions a hyperbolic approach for the reaction rate
Naphthalene
of steam reforming of naphthalene is developed and parameters for 650 C and 700 C are
Coke formation
determined. © 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
Introduction In biomass gasification, renewable raw materials or residues can be used to produce electrical energy or fuels. The predominantly hydrogen-rich product gas of the absorption enhanced reforming (AER-) process [1e5] can be converted into electricity in gas engines, gas turbines or fuel cells. A further possibility is to convert the product gas catalytically into gaseous fuels like pure hydrogen or synthetic natural gas
(SNG) as well as liquid fuels using such methods as the FischereTropsch process. Advantages of using a pressurized gasifier are its smaller volume, as well as its compatibility with pressurized downstream applications like methanation or conversion in a gas turbine without the need to cool the gas for compression. Product gases of any gasifier contain impurities such as particles, chlorine, sulfur and tars [6], which must be removed before the power generation or fuel synthesis. The present work focuses on tar removal, for which tars are defined as all undesired higher hydrocarbons of the
* Corresponding author. E-mail address: [email protected] (M. Speidel). 1 I did this work at German Aerospace Center, but now I do not work there any more. Yet my former colleague Holger Fischer still works there. http://dx.doi.org/10.1016/j.ijhydene.2016.05.023 0360-3199/© 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 9 2 0 e1 2 9 2 8
12921
Fig. 1 e Concept of pressurized gasification of biomass for power generation in a hybrid solid oxide fuel cell (SOFC) and gas turbine system.
gasification gas [7]. More specifically, the catalytic steam reforming of tars is investigated [8,9]. Since the use of catalytic additives in the gasifier bed material [10,11] or in ceramic filters [12] is not sufficient to achieve the required tar conversions of over 99% for power generation in fuel cells, tar reforming in a fixed bed downstream from the gasifier is considered. As can be seen in Fig. 1, the pressurized product gas of the AER-process is cleaned and then converted in a hybrid power generation system consisting of a Solid Oxide Fuel Cell (SOFC) and a gas turbine [13e19]. The AER-process consists of two reactors. In the first one biomass is gasified with steam and in the second one, combustion of residual coke takes place, which provides the necessary heat for the gasification. CO2 is bound at the bed material CaO and therefore the AER-product gas contains small concentrations of CO2 and high concentrations of H2. The product gas composition can be seen in Table 1.
Table 1 e Simulated AER product gas composition at 5 bara in mole fraction and typical tar yield of AER product gas referred to dry gas [20]. yH2 O
yH2
yCO
yCH4
yC2 H4
yC2 H6
yC3 H8
yCO2
Tar
60%
26%
2.4%
5.6%
1.2%
0.6%
1.2%
3.0%
1e5 g/m3N
Particles, chlorine and sulfur impurities can be removed at temperatures greater than 600 C [6] (see Fig. 1). Afterwards tars are reformed catalytically by using the steam contained in the gas (see Eq. (1) in Fig. 1). The reaction takes place in a temperature range of 600 Ce700 C and at a pressure of 5 bar absolute pressure (bara). The advantage of a low reformer temperature (<700 C) is a reduction in the secondary steam reforming of methane (see Eq. (2) in Fig. 1) due to chemical equilibrium and therefore less heat demand for that endothermic reaction. In the case of the pressurized AER product gas shown in Table 1, at approximately 650 C no methane will be converted in the tar reformer at 5 bara. The challenges of lower temperatures, however, are the slower reaction kinetics of tar reforming and the risk of coke formation at the nickel catalyst, especially at pressures higher than atmospheric. Therefore a detailed understanding of tar reforming reaction kinetics is required for the design of that process unit.
A laboratory-scale test rig with a fixed catalytic bed was built to investigate the reaction kinetics experimentally. Naphthalene (C10H8) is used as the model tar component because it is the main component of tars in AER product gas (mass fraction of 31.5% of tars heavier than toluene [2]), it is often an intermediate product at the tar reforming process and its reaction rate is low in comparison to other tars. Jess [21] found that naphthalene as a model tar is strongly adsorbed on the surface of a nickel catalyst and inhibits the simultaneous conversion of methane and benzene. Due to this inhibition at higher concentrations the reaction order of steam reforming of naphthalene itself is, at 0.2, much lower than 1. Therefore Depner [22] uses a hyperbolic EleyeRideal approach for steam reforming of naphthalene on a nickel catalyst. This approach is based on adsorption of naphthalene and includes inhibition of the reaction due to active centers on the catalyst surface being covered by naphthalene. While Jess [21], Depner [22] and current works of Hamel et al. [23], Kaisalo et al. [24] and Kurkela et al. [25] investigate tar reforming at temperatures higher than 700 C, Fraubaum et al. [26] and Kienberger et al. [27] focus on temperatures even lower than 500 C with combined methanation. The aims of this work are to demonstrate stable steam reforming of naphthalene in a temperature range of 600 Ce700 C, to derivate an approach for the conversion rate of naphthalene, and to determine the kinetic parameters for this approach.
Experimental An average molar gas composition of the technical tar reformer in Fig. 1 of 45%e55% hydrogen and 35%e37% steam is synthesized as feed for the reformer at the test rig (see Fig. 2). Further gas components such as carbon monoxide (CO), carbon dioxide (CO2), methane (CH4) and other
Fig. 2 e Schematic draft of the test rig.
12922
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 9 2 0 e1 2 9 2 8
Fig. 3 e The pressurized tar saturator (left) and the catalytic fixed bed reformer heated by an oven (right).
hydrocarbons (see Table 1) are not considered for the reaction of tar with steam and are replaced by 10%e18% nitrogen. Controlled introduction of naphthalene into the gas is carried out in the saturator via a continuous flow of nitrogen. The mixed gas can be analyzed either before or after the reforming reaction. Hydrocarbons are detected discretely in the wet gas by a gas-phase chromatograph (GC) with a flame ionization detector (FID), whereas CO, CO2 and CH4 are detected continuously in the dry gas by infrared spectroscopy (IR).
Saturator The left side of Fig. 3 shows the pressurized saturator, which is filled with solid naphthalene at room temperature. Naphthalene is then melted by heat input from a heating jacket Hliq, which controls the temperature of the molten naphthalene Tliq. Preheated nitrogen of the same temperature flows through the molten naphthalene and gets saturated. A filter at the end of the inlet tube ensures small bubbles, which have a longer retention time and better mass transport properties. The saturator works at 10 bara, which is controlled with a mechanical pressure controller (PC). A second heating jacket Hgas controls the gas temperature Tgas in the saturator. It was experimentally found that stable, reproducible operation of the saturator is achieved when the naphthalene fill level is less than half and the gas temperature is approximately 4e5 K higher than the liquid temperature. It can be observed that the nitrogen stream is fully saturated with naphthalene. Due to the oscillating gas and liquid temperatures in the saturator, the concentration of naphthalene in the gas also oscillates with a relative standard deviation of 1%e1.5%. To minimize deviation, the measured concentrations of naphthalene are averaged over time.
Reformer The reformer contains a catalytic fixed bed (see right side of Fig. 3) with alumina pellets of 2.3 mm diameter, which are
impregnated with nickel and supplied by Johnson Matthey. The properties of the catalyst pellets and of the fixed bed are given in Table 2. The fixed bed section is housed in an electrical oven, which controls the temperature at the outer wall of the tube at the inlet of the catalytic bed (TW1 in Fig. 3). An inert fixed bed of alumina pellets is included above the catalytic bed and separated by glass wool. Thus, heat transfer from the tube to the gas is improved and a uniform flow field and an equal temperature of gas and pellets can be assumed from the inlet of the catalytic bed downwards. The gas is pre-heated before entering the oven to ensure that the gas temperature in the catalytic bed TR1 (see Fig. 3) is the same as the tube temperature TW1. Maximum temperature deviation in the catalytic bed due to the endothermic reaction (TR1TR2) is lower than 5 K in the experiments presented in this work. The arithmetic mean of TR1 and TR2 is used as the reaction temperature.
Gas-phase chromatograph The flame ionization detector (FID) of the gas-phase chromatograph detects the hydrocarbons in the gas sample.
Table 2 e Properties of the catalyst pellets and of the fixed bed. Manufacturer Active material Carrier material Manufacturing method Pellet diameter Pellet density Pellet porosity BET-surface Active surface Median pore diameter Pore shape Fixed bed length Fixed bed density Fixed bed porosity
Johnson Matthey Nickel (15% mass fraction) Alumina (gamma) Impregnation 2.3 mm 1380 kg/m3 34.5% 64 m2/g 3.5 m2/g 13.4 nm cylindrical 20 mm 587 kg/m3 57.5%
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 9 2 0 e1 2 9 2 8
Reformate gas flows continuously through a sampling loop with a specific volume at a rest temperature of 200 C and a pressure of approximately 1 bara. Every 18.5 min, the sample gas in the loop is carried to the column by the hydrogen carrier gas. The gas-phase chromatograph is calibrated by a naphthalene solution, which is vaporized in the injector and then carried to the column. The calibration is verified by the mass loss of the saturator. More sample gas is carried to the column with higher pressure in the sampling loop. Therefore the measured gas-phase composition is adjusted by the average sample pressure.
Methodology The nickel catalyst is activated by reducing nickel oxide to nickel at 850 C with 50 vol% hydrogen (rest nitrogen) for half an hour after being heated up with pure nitrogen. In order to find out whether half an hour is sufficient for the activation process methane steam reforming is carried out, which reaches stationary conditions and methane conversion within minutes. The activation process and methane reforming is repeated until constant methane conversion indicates complete activation of the catalyst. The whole activation process is done once after which several naphthalene steam reforming tests can be carried out. The reformer is heated up with nitrogen before the test. Shortly before opening the tar saturator, steam and hydrogen are switched on. After the test the saturator is closed, then steam and hydrogen are switched off and the reformer is either cooled down with nitrogen or the temperature is maintained by a small volume flow of nitrogen until the next test begins. Furthermore the weight loss of the naphthalene in the saturator is determined after every test in order to confirm the concentration measurements in the gas-phase chromatograph or to correct them. For naphthalene concentrations below 300 ppmv it is assumed that total mole flow is constant. Therefore relative naphthalene conversion can be calculated based on molar fractions ynaph: Xnaph ¼
out yin naph ynaph
(3) yin naph By measuring the naphthalene concentrations upstream and downstream of the reformer the relative naphthalene conversion can be calculated. All experiments are carried out at 5 bara and a feed volume flow of 545 lN/h, which corresponds to a gas hourly space velocity of 107,000 1/h. With maximum naphthalene concentration of 300 ppmv, 45%e55% hydrogen and 35%e37% steam, as explained at the beginning of Chapter 2, there is a high excess of hydrogen and steam in comparison to naphthalene (minimum ratio of 1500 in the case of hydrogen and 1167 in the case of steam).
Results and discussion In this chapter the experimental results of the steam reforming of naphthalene at the laboratory-scale test rig are introduced. First in Chapter Transient progression of the naphthalene conversion the transient progression of naphthalene conversion is examined before the stationary conversions at temperatures between 600 C and 700 C are
12923
shown in Chapter Stationary experimental naphthalene conversion. A kinetic approach for stationary conversion is subsequently derived in Chapter Derivation of a kinetic approach of naphthalene conversion rate. In Chapter Determination of kinetic parameters the kinetic parameters for this kinetic approach are determined.
Transient progression of the naphthalene conversion Fig. 4 shows several tests at different temperatures which are carried out sequentially. All tests are performed with a naphthalene feed concentration of 130 ppmv. The first experiment (see Fig. 4 (a)) is carried out at 655 C with a fresh catalyst, which has been reduced with hydrogen (see Section Methodology). High naphthalene conversion of more than 80% decreases within several hours and reaches a stable stationary conversion of approximately 50%. Due to the findings of Jess [21] and Depner [22] that strong naphthalene adsorption reduces the amount of free active centers, the deactivation of the nickel catalyst is explained by the active centers on the surface of the catalyst becoming covered. Yet there seem to be some uncovered centers which enable stationary naphthalene conversion. The next test (see Fig. 4 (b)) is carried out with the same now used catalyst at 640 C starting at a conversion of approximately 50% before the catalyst is further deactivated and reaches a stationary conversion of approximately 40%. The test thereafter is carried out at 620 C with the same catalyst. The results in Fig. 4 (c) show the repetition of this test at 620 C. The subsequent test at 600 C (d) shows the same characteristics as the test at 640 C (b). The conversion starts high before the catalyst deactivates further and reaches a stationary conversion of approximately 10%. Repeating the test at 655 C (e, f) after the test at 600 C indicates that the catalyst is partially reactivated after several hours. The conversion of naphthalene starts at approximately 30%. It then increases before reaching the same reproducible stationary conversion of approximately 50% as in the first test (a). It is concluded that the deactivation of the catalyst is caused by active centers on the surface being covered. Furthermore, it is concluded that more active centers are covered at lower temperatures. This can be caused by adsorption of naphthalene as described in Refs. [21] and [22]. Yet the long deactivation times indicate that coke formation can also be a reason for deactivation. The outlet concentrations for a steam reforming test of naphthalene at 5 bara and 650 C can be seen in Fig. 5. The concentrations in Fig. 5 refer to the amount of carbon atoms in order to facilitate the comparison of higher hydrocarbons, in this case naphthalene (C10H8) and benzene (C6H6). The feed concentration of the reformer is approximately 165 ppmv naphthalene respectively 1650 ppmv carbon atoms. Using steam, naphthalene is converted to CO and H2 (see Eq. (1) in Fig. 1). Secondary reactions of CO (CO-Shift and methanation) lead to CO2 and H2 or methane and steam. Due to the high hydrogen feed concentration of 55 vol% and the small difference between feed concentration and outlet concentration, hydrogen is not measured. Benzene (C6H6) is the only detectable by-product of naphthalene, which is subsequently also converted with steam to CO and H2. The appearance of benzene as a by-product indicates that naphthalene cracking also takes
12924
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 9 2 0 e1 2 9 2 8
Fig. 4 e Naphthalene conversions (calculated using Eq. (3)) for several tests at different temperatures for 5 bara and 130 ppmv naphthalene feed concentration.
Fig. 5 e Reformate gas composition during a steam reforming test with 165 ppmv naphthalene feed concentration at 5 bara and 650 C.
place. The concentration of naphthalene rises while the sum of the carbon-containing products (methane, CO, CO2 and benzene) declines. Thus, the conversion of naphthalene falls and does not reach stationary conditions within 7 operating hours. Additionally, conversion rates of the secondary reactions of CO also fall, as evidenced by the decreasing concentrations of the CO2 and methane products. Covered active centers on the surface of the catalyst inhibit all of these reactions. Hydrocracking of naphthalene also leads to naphthalene conversion and formation of methane, but is assumed to play a tangential role compared to steam reforming of naphthalene. At the test rig, methane steam reforming was also investigated. It was found that following naphthalene steam reforming test, the methane conversion at the nickel catalyst is significantly inhibited. This indicates that the active centers are covered not only by naphthalene adsorption but also by coke formation, which has long-term effects on catalyst activity. It was found that
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 9 2 0 e1 2 9 2 8
12925
the coke canbe gasifiedbyaddition of steam at high temperatures and the activity of the catalyst can be restored. The results of hydrogen-based laboratory chemisorption measurements performed on the used catalyst following tests with a 130 ppmv naphthalene feed concentration and 650 C and 700 C confirm a significantly less active catalyst surface compared to the fresh catalyst.
Stationary experimental naphthalene conversion Fig. 6 contains the stationary conversions of naphthalene for several tests at temperatures between 600 C and 700 C. For naphthalene feed concentrations of approximately 130 ppmv, stable stationary naphthalene conversions are reached for temperatures as low as 600 C. A strong temperature dependency is nevertheless observed in the temperature range studied. It is assumed that the low naphthalene conversions at low temperatures are caused by the covered active centers in particular.
Fig. 7 e Schematic diagram of reaction steps on the catalytic surface.
rnaph ¼ knaph $qnaph
where knaph is the reaction rate constant. The adsorption rate is proportional to the partial pressure of naphthalene and the fraction of uncovered active centers. The desorption rate is proportional to the centers covered with naphthalene. Equilibrium of adsorption and desorption is expressed by: kads $ 1 qcoke qnaph $pnaph ¼ kdes $qnaph Kads $ 1 qcoke qnaph $pnaph ¼ qnaph
Fig. 6 e Stationary naphthalene conversions for several tests at temperatures between 600 C and 700 C at 5 bara and 130 ppmv naphthalene feed concentration (compare tests in Fig. 4).
The aim of this work is to describe the stationary part of naphthalene steam reforming. Therefore a kinetic approach is derived that takes into consideration the effects of covered active centers by adsorbed naphthalene and coke.
Derivation of a kinetic approach of naphthalene conversion rate Fig. 7 shows the assumed reaction steps on the surface of the nickel catalyst. Jess [21] and Depner [22] indicate that naphthalene is strongly adsorbed at the surface and therefore covers some of the active centers. Presented results show that coke formation also takes place and leads to covered active centers. As shown previously coke can be gasified with steam. After adsorption, a naphthalene molecule can be desorbed, can react with steam e the desired steam reforming reaction e or can be converted directly or indirectly to coke. The reaction rate of steam reforming of naphthalene is determined for stationary conditions. The reaction rate is proportional to the fraction of the surface covered with naphthalene and is independent of steam concentration as there is an excess of steam in the feed gas:
(4)
(5)
where qcoke is the fraction of surface covered with coke, kads the adsorption rate constant and kdes the desorption rate constant for naphthalene. The adsorption equilibrium constant Kads is the ratio of the adsorption and desorption rate constants. Catalytic coke formation is assumed to be first order in naphthalene [28]: rcoke ¼ kcoke $qnaph The reaction rate of coke gasification with steam is assumed to be proportional to the fraction of the surface covered with coke and independent of steam as there is an excess of steam in the gas: rgas ¼ kgas $qcoke The quasi-stationary state of coke formation and coke gasification can be calculated to kcoke $qnaph ¼ kgas $qcoke qcoke ¼ Kcoke $qnaph
(6)
where Kcoke is the ratio of coke formation and gasification rate constants kcoke and kgas. Substituting qcoke from Eq. (6) into Eq. (5) results in: qnaph ¼
Kads $pnaph 1 þ Kads $pnaph þ Kads $Kcoke $pnaph
Substituting qnaph into Eq. (4) yields for the reaction rate of steam reforming: k0naph
rnaph
zfflfflfflfflfflffl}|fflfflfflfflfflffl{ knaph $Kads $pnaph ¼ 1 þ Kads $ð1 þ Kcoke Þ $pnaph |fflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflffl} K0naph
(7)
12926
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 9 2 0 e1 2 9 2 8
As it is experimentally not possible to differentiate between active centers covered by adsorbed naphthalene or coke, the parameter K0naph is defined including both effects. The reaction rate constant knaph and the adsorption constant Kads are also summarized to parameter k0naph . The parameters K0naph and k0naph are only dependent on temperature. For low partial pressures of naphthalene the reaction order is 1 and for high partial pressures of naphthalene the limit value for the reaction order is 0. If hydrocracking of naphthalene contributes to naphthalene conversion and methane formation, in addition to the described steam reforming of naphthalene to CO and subsequent methanation (compare description of Fig. 5), the influence of the constant hydrogen concentration will be contained in parameter knaph of Eq. (4). Coke formation can be caused by catalytic cracking or hydrocracking. In the case of hydrocracking the impact of the constant hydrogen concentration is contained in parameter kcoke. A slightly positive effect of hydrogen on the naphthalene conversion is detected similar to the dependency described in Ref. [21] with reaction order of 0.3 for hydrogen, but not considered in this work. Strictly speaking the kinetic approach is only valid for a constant hydrogen concentration. Therefore the following experiments are carried out at constant hydrogen concentrations.
Determination of kinetic parameters Assuming constant total mole flow n_tot , the following equation describes the reduction of naphthalene in an infinitesimally small volume element with the cross sectional area AC and the length dx. n_tot $dynaph ¼ rnaph $rbed $Ac $dx
(8)
where rbed is the mass-density of the catalytic fixed bed (see Table 2). Integration from inlet to outlet for a reformer with the length L results in the following one-dimensional balance equation for naphthalene: yout naph
Z
1 rnaph
yin naph
$dynaph ¼
rbed $Ac $ n_tot
ZL dx
(9)
x¼0
Substituting the conversion rate of Eq. (7) into the molar balance of naphthalene (9) and integrating results in the following equation:
Table 3 e Parameters determined for stationary conditions of steam reforming of naphthalene. Temperature [ C] 649 700
yin 1 naph $ln out p ynaph
!
k0naph [mol/kgcat s Pa]
K0naph [1/Pa]
9.42$106 1.20$105
0.042 0.0165
1 0 out þ K0naph $ yin naph ynaph ¼ _ $knaph $rbed $Ac $L ntot
(10)
For two tests at the same temperature and different feed concentrations yin naph , these two Eq. (10) are divided. Since the right side of Eq. (10) is equal for equal temperatures, it can be eliminated. The divided equation results in: ! ! 1 yin yin 1 naph;1 naph;2 out $ln out $ln þ K0naph $ yin y naph;1 naph;1 p p ynaph;1 yout naph;2 0 in out Knaph $ ynaph;2 ynaph;2 ¼ 0
(11)
Measuring the concentrations of naphthalene upstream out and downstream from the reformer (yin naph and ynaph ) for the two tests, parameter K0naph can be determined with Eq. (11). At 650 C and at 700 C several tests are carried out at different feed concentrations. Parameter K0naph is determined with the method of least squares for Eq. (11) by combining in each case two tests at different naphthalene concentrations. Afterwards k0naph is determined with Eq. (10) for the test with the highest feed concentration. Table 3 lists the kinetic parameters determined for tests at 650 C or rather 649 C and 700 C. The higher value for K0naph at 649 C compared to 700 C confirms that at lower temperatures more active centers are covered and the catalyst is more deactivated. Fig. 8 compares the calculated naphthalene conversions to the experimental conversions. The effect of the covered catalyst surface is evident, especially at higher naphthalene partial pressures or concentrations, and can be described well with the introduced kinetic approach. The reaction in these experiments is strongly limited by pore diffusion, especially for lower naphthalene concentrations when the reaction is faster. This can be seen by the high values of the Weisz-Prater parameter [29] even for a feed concentration of 300 ppmv shown in Table 4, according to the definition of the Weisz-Prater parameter (WPP) presented in Ref. [30]. For conditions without influence of pore diffusion K0naph would be higher.
Fig. 8 e Comparison of experimental and calculated conversion rates at 649 C (left) and 700 C (right).
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 9 2 0 e1 2 9 2 8
12927
Table 4 e Weisz-Prater parameter (WPP) for 649 C and 700 C and different reaction orders n. T [ C]
ynaph [ppmv]
649 700
300 300
Dnaph [m2/s] 5
2.72$10 2.99$105
DK [m2/s] 6
1.74$10 1.79$106
The Weisz-Prater parameter is calculated as follows [30]: WPP ¼
bed 0:5$ðn þ 1Þ$rnaph $rεbed $L2c
p
naph Deff $ R$T
(12)
where n is the reaction order, rbed the density of the catalytic bed (see Table 2), εbed the porosity of the bed (see Table 2) and pnaph the partial pressure of naphthalene in the bulk phase. rnaph is the measured reaction velocity calculated using Eq. (7). Lc is the characteristic length of the pellet, in the definition of [30] the ratio of the volume of the pellet and the surface. The effective diffusion coefficient Deff is calculated by means of the molecular diffusion coefficient Dnaph, the Knudsen diffusion coefficient DK, the porosity of the pellet εP (see Table 2) and the tortuosity t as follows: Deff ¼
1 εP $ þ D1K t
1 Dnaph
Deff [m2/s]
WPP (n ¼ 0)
WPP (n ¼ 1)
1.88$107 1.94$107
3.9 10.6
7.8 21.3
Nevertheless, it can be concluded that tar reforming at temperatures as low as 650 C is possible at a pressure of 5 bara and a feed concentration of the model tar naphthalene of 300 ppmv, which corresponds to approximately 1.7 g/m3N based on wet gas. This enables tar conversion without secondary methane conversion because of chemical equilibrium limitation. The heat input required for the tar reformer can be reduced considerably, which has a positive effect on the overall process of pressurized gasification of biomass for power generation in a hybrid Solid Oxide Fuel Cell (SOFC) and gas turbine system. With inhibited methane conversion caused by coke it is also conceivable that methane conversion can be limited due to kinetic reasons even at temperatures higher than 650 C if an appropriate reactor concept is developed.
(13)
Tortuosity of the pellet is not known, but estimated with a typical value of 3. Dnaph is calculated by means of binary diffusion coefficients for a gas composition of 45% hydrogen, 37% steam and 18% nitrogen. The reaction is limited by pore diffusion at values of the Weisz-Prater parameter higher than 0.25 and strongly limited at values higher than 5 [30]. Table 4 shows that the reaction is even limited by pore diffusion at 649 C and a limit value of 0 for the reaction order.
Conclusions A laboratory-scale test rig was introduced, which enables the investigation of steam reforming of tar at pressures of up to 5 bara in a temperature range of 600 Ce700 C, as well as the determination of the reaction kinetics at stationary conditions. The model tar component naphthalene is converted with steam out of the synthesized gasification gas to CO and H2. Although the nickel catalyst deactivates, a reproducible, stationary conversion of naphthalene is reached within several hours or days. Adsorption of naphthalene and naphthalene based coke formation inhibit the conversion of naphthalene itself as well as secondary reactions of the product CO. The lower the temperature, the more the catalyst gets deactivated at stationary conditions. A hyperbolic approach for the conversion rate at stationary conditions was derived. Therefore it is assumed that active centers on the surface of the catalyst are covered by coke and adsorbed naphthalene. Furthermore, a quasi-stationary state of coke formation and coke gasification with steam is assumed. The influence of naphthalene concentration is determined for temperatures of 650 C and 700 C. The higher value of parameter K0naph for 650 C confirms that more active centers are covered at lower temperatures and the activity of the catalyst is lower.
Acknowledgments The authors thank Sandra Adelung and Dr. Ralph-Uwe Dietrich for their assistance presenting the results of this work. The Helmholtz Gemeinschaft grant within the DLR@UniST project is gratefully acknowledged. The nickel catalyst used in this work is kindly provided by Johnson Matthey.
references
[1] Pfeifer C, Puchner B, Hofbauer H. In-situ CO2-absorption in a dual fluidized bed biomass steam gasifier to produce a hydrogen rich syngas. Int J Chem React Eng 2007;5:A9. [2] Pfeifer C, Puchner B, Hofbauer H. Comparison of dual fluidized bed steam gasification of biomass with and without selective transport of CO2. Chem Eng Sci 2009;64:5073e83. http://dx.doi.org/10.1016/j.ces.2009.08.014. [3] Poboß N, Swiecki K, Charitos A, Hawthorne C, Zieba M, Scheffknecht G. Experimental investigation of the absorption enhanced reforming of biomass in a 20 kWth dual fluidized bed system. Int J Thermodyn 2012;1:53e9. [4] Koppatz S, Pfeifer C, Rauch R, Hofbauer H, MarquardMoellenstedt T, Specht M. H2 rich product gas by steam gasification of biomass with in situ CO2 absorption in a dual fluidized bed system of 8 MW fuel input. Fuel Process Technol 2009;90:914e21. http://dx.doi.org/10.1016/ j.fuproc.2009.03.016. [5] Udomsirichakorn J, Basu P, Salam PA, Acharya B. Effect of CaO on tar reforming to hydrogen-enriched gas with inprocess CO2 capture in a bubbling fluidized bed biomass steam gasifier. Int J Hydrogen Energy 2013;38:14495e504. http://dx.doi.org/10.1016/j.ijhydene.2013.09.055. [6] Aravind PV, de Jong W. Evaluation of high temperature gas cleaning options for biomass gasification product gas for solid oxide fuel cells. Prog Energ Combust 2012;38:737e64. http://dx.doi.org/10.1016/j.pecs.2012.03.006.
12928
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 9 2 0 e1 2 9 2 8
[7] Li C, Suzuki K. Tar property, analysis, reforming mechanism and model for biomass gasificationdan overview. Renew Sust Energ Rev 2009;13:594e604. http://dx.doi.org/10.1016/ j.rser.2008.01.009. [8] Ising M. Zur katalytischen Spaltung teerartiger Kohlenwasserstoffe bei der Wirbelschichtvergasung von € t Dortmund, Biomasse [Dissertation]. German: Universita Fraunhofer IRB Verlag Stuttgart; 2002. [9] Pfeifer C, Huettler H, Hofbauer H, Schulzke T, Unger C. Catalytic tar decomposition in a slip stream of an 8 MW dual fluidized bed biomass steam gasifier. In: International Symposium on Gas Cleaning at High Temperatures (GCHT); 2008. Newcastle, UK. [10] Choi YK, Cho MH, Kim JS. Air gasification of dried sewage sludge in a two-stage gasifier. Part 4: application of additives including Ni-impregnated activated carbon for the production of a tar-free and H2-rich producer gas with a low NH3 content. Int J Hydrogen Energy 2016;41:1460e7. http:// dx.doi.org/10.1016/j.ijhydene.2015.11.125. [11] Mun TY, Kim JW, Kim JS. Air gasification of dried sewage sludge in a two-stage gasifier: part 1. The effects and reusability of additives on the removal of tar and hydrogen production. Int J Hydrogen Energy 2013;38:5226e34. http:// dx.doi.org/10.1016/j.ijhydene.2012.10.120. [12] Savuto E, Di Carlo A, Bocci E, D'Orazio A, Villarini M, Carlini M, et al. Development of a CFD model for the simulation of tar and methane steam reforming through a ceramic catalytic filter. Int J Hydrogen Energy 2015;40:7991e8004. http://dx.doi.org/10.1016/ j.ijhydene.2015.04.044. [13] Methling T, Armbrust N, Haitz T, Speidel M, Poboss N, BraunUnkhoff M, et al. Power generation based on biomass by combined fermentation and gasification - a new concept derived from experiments and modelling. Bioresour Technol 2014;169:510e7. http://dx.doi.org/10.1016/ j.biortech.2014.07.036. € rner A. A new process concept for [14] Speidel M, Kraaij G, Wo highly efficient conversion of sewage sludge by combined fermentation and gasification and power generation in a hybrid system consisting of a SOFC and a gas turbine. Energy Convers Manag 2015;98:259e67. http://dx.doi.org/10.1016/ j.enconman.2015.03.101. [15] Dekker N, Rabou L, van Wees H, Rietveld B. Operation of a Staxera SOFC stack - fuelled with cleaned gas from a gasifier. In: 8th European SOFC Forum; 2008. Lucerne, Switzerland. [16] Henke M, Willich C, Kallo J, Friedrich A. Theoretical study on pressurized operation of solid oxide electrolysis cells. Int J Hydrogen Energy 2014;39:12434e9. http://dx.doi.org/10.1016/ j.ijhydene.2014.05.185. [17] Leucht F, Bessler WG, Kallo J, Friedrich KA, Mu¨llerSteinhagen H. Fuel cell system modeling for solid oxide fuel cell/gas turbine hybrid power plants, part i: modeling and simulation framework. J Power Sources 2011;196:1205e15. http://dx.doi.org/10.1016/j.jpowsour.2010.08.081.
[18] Panne T, Widenhorn A, Boyde J, Matha D, Abel V, Aigner M. Thermodynamic process analyses of SOFC/GT hybrid cycles. In: 5th International Energy Conversion Engineering Conference and Exhibit (IECEC); 2007. St. Louis, Missouri (USA). [19] Veyo S, Lundberg W, Vora S, Litzinger K. Tubular SOFC hybrid power system status. In: Proceedings of ASME Turbo Expo; 2003 June 16-19. Atlanta, Georgia, USA. [20] Steiert S. FuE-Plattform “Biomass-to-Gas” e Energetische Nutzung biogener Reststoffe mit AER-Technologie zur Poly-Generation von Strom, Wasserstoff, Erdgassubstitut € rme, Schlussbericht. Technische und Wa € tsbibliothek, Informationsbibliothek u. German: Universita Stuttgart; 2013. [21] Jess A. Catalytic upgrading of tarry fuel gases: a kinetic study with model compounds. Chem Eng Process 1996;35:487e94. http://dx.doi.org/10.1016/S0255-2701(96)04160-8. [22] Depner H. Untersuchungen zur katalytischen Umsetzung flu¨chtiger Schwelprodukte in Rohgasen der Verkokung und Vergasung fester Brennstoffe [Dissertation]. German: € t Karlsruhe; 1998. Universita [23] Hamel C. Options to reach syngas quality requirements by catalytic reforming of tars and methane from FB-gasification of biomass. In: 6th International Freiberg Conference on IGCC & XtL Technologies; 2014. Freiberg, Germany. [24] Kaisalo N, Kihlman J, Hannula I, Simell P. Reforming solutions for biomass-derived gasification gas e experimental results and concept assessment. Fuel 2015;147:208e20. http://dx.doi.org/10.1016/j.fuel.2015.01.056. [25] Kurkela E, Kurkela M, Hiltunen I. Steameoxygen gasification of forest residues and bark followed by hot gas filtration and catalytic reforming of tars: results of an extended time test. Fuel Process Technol 2016;141:148e58. http://dx.doi.org/ 10.1016/j.fuproc.2015.06.005. [26] Fraubaum M, Walter H, Zuber C. Kinetic modeling of a combined tar removal and methanation reactor for biogenous synthesis gas at medium temperature conditions. Fuel Process Technol 2016;141:159e66. http://dx.doi.org/ 10.1016/j.fuproc.2015.06.003. [27] Kienberger T, Zuber C, Novosel K, Baumhakl C, Karl J. Desulfurization and in situ tar reduction within catalytic methanation of biogenous synthesis gas. Fuel 2013;107:102e12. http://dx.doi.org/10.1016/j.fuel.2013.01.061. [28] Goud SK, Whittenberger WA, Chattopadhyay S, Abraham MA. Steam reforming of hexadecane using a catalyst: kinetics of catalyst deactivation. Int J Hydrogen Energy 2007;32:2868e74. http://dx.doi.org/10.1016/ j.ijhydene.2007.03.019. [29] Weisz PB, Prater CD. Interpretation of measurements in experimental catalysis. Adv Catal 1954;6:143e96. http:// dx.doi.org/10.1016/S0360-0564(08)60390-9. [30] Missen RW, Mims CA, Saville BA. Introduction to chemical reaction engineering and kinetics. Wiley VCH; 1999.