Steam reforming of methane over Ni catalysts prepared from hydrotalcite-type precursors:Catalytic activity and reaction kinetics☆

Yang Qi,Zhenmin Cheng,Zhiming Zhou*

State Key Laboratory of Chemical Engineering,East China University of Science and Technology,Shanghai 200237,China

Keywords:Methane steam reforming Hydrogen Hydrotalcite Catalyst Kinetics

ABSTRACT Ni/Mg-Al catalysts derived from hydrotalcite-type precursors were prepared by a co-precipitation technique and applied to steam reforming of methane.By comparison with Ni/γ-Al2O3 and Ni/α-Al2O3 catalysts prepared by incipient wetness impregnation,the Ni/Mg-Al catalyst presented much higher activity as a result of higher specific surface area and better Ni dispersion.The Ni/Mg-Alcatalyst with a Ni/Mg/Al molarratio of0.5:2.5:1 exhibited the highest activity for steam methane reforming and was selected for kinetic investigation.With external and internal diffusion limitations eliminated,kinetic experiments were carried out at atmospheric pressure and over a temperature range of823-973 K.The results demonstrated that the overall conversion of CH4 and the conversion of CH4 to CO2 were strongly influenced by reaction temperature,residence time of reactants as well as molar ratio of steam to methane.A classical Langmuir-H inshelwood kinetic model proposed by Xu and Froment(1989)fitted the experimental data with excellent agreement.The estimated adsorption parameters were consistent thermodynamically.

1.Introduction

Hydrogen has been widely used in chemical industry.For example,hydrogen is an important and indispensable raw material for manufacture of bulk chemicals such as ammonia and methanol[1,2].Large quantities of hydrogen are being consumed in petroleum refining since crude oil becomes sourer and heavier and increasing environmental awareness results in the need for clean fuels[3].In addition,hydrogen is regarded as the next generation energy carrier due to its nonpolluting,inexhaustible,efficient,and potentially cost effective features[4,5].Therefore,hydrogen production has attracted much attention from both academy and industry in recent years.

Among various methods for hydrogen production at present,steam reforming of hydrocarbons,especially of methane,is the most widely used route.For the steam reforming of methane,Ni and noble metals such as Ru,Rh,Pd,and Pt are usually used as active metals in supported catalysts[6-8].In comparison with noble metal based catalysts,Ni-based catalysts are normally less active and subject to more serious deactivation by coke formation and sintering of metallic Ni active phase at high temperatures[2,9,10].However,because of its low cost(only 1/150-1/100 of that of noble metals)[11],Ni is the most widely used metal.Among many Ni-based catalysts,those derived from hydrotalcite-type precursors exhibit excellent performance for steam methane reforming due to their interesting properties such as large surface area,basic property,and memory effect that can reconstruct the original structure of hydrotalcites once contacting with aqueous solutions[12,13].In addition,active metals can distribute uniformly on the surface of these catalysts[14-16].

In the last decade,Ni catalysts obtained from hydrotalcite-type precursors for steam methane reforming have been studied by many researchers.Fonseca and Assaf[16]used different techniques(traditional co-precipitation technique and chelation-based techniques including co-precipitation and anion-exchange methods)to prepare NiMgAlO catalysts,and found that the catalyst obtained through Ni chelates was the most active in steam methane reforming.Ochoa-Fernández et al.[17]compared different Ni/Mg-Al catalysts prepared by co-precipitation and incipient wetness impregnation methods,and concluded that the co-precipitation method led to smaller Nicrystals than incipient wetness impregnation,which is responsible for the excellent stability and high activity in steam reforming of CH4.Basile et al.[18]investigated the effects of the composition of hydrotalcite-type precursors and the preparation method on the physical-chemical properties as well as the catalytic performance of the catalysts.Takehira et al.[15,19,20]developed a solid phase crystallization method to prepare Ni/Mg-Al catalysts,and reported that the Ni/Mg-Alcatalysts prepared by co-precipitation method gave better metal dispersion and thus higher activity for steam reforming of CH4.Very recently,we prepared a Ni/Mg-Al catalyst for hydrogen production by the sorption-enhanced steam methane reforming process(SESMR),showing high activity and stability in multicycle operation[21].However,the structure-activity relationship of this catalyst was not taken into account.Furthermore,it should be stressed that all the literature mentioned above mainly focuses on catalyst preparation and structure characterization.As far as the kinetics of steam methane reforming over this type of catalysts is concerned,no relevant study has been reported despite the significant role of kinetics in simulation and design of the reactor as well as optimization of operation conditions.

In this work,three Ni/Mg-Al catalysts with various metal compositions are prepared by co-precipitation method,and as a reference,Ni/γ-Al2O3and Ni/α-Al2O3catalysts are prepared by traditional incipient wetness impregnation.These catalysts are characterized by a variety of analytical techniques and applied to steam reforming of CH4.The reaction kinetics is studied over the screened Ni/Mg-Alcatalyst.The main objectives of this work are two-fold:(1)elucidate the structure-activity relationship of the Ni/Mg-Al catalysts;and(2)develop a rigorous and reliable kinetic model for steam reforming of CH4over the Ni/Mg-Al catalysts,which can be used for designing and optimizing the SESMR process.

2.Experimental

2.1.Materials

Ni(NO3)2·6H2O(98.5%purity),Mg(NO3)2·6H2O(99.0%purity),and Al(NO3)3·9H2O(99.0%purity)purchased from Sinopharm Chemical Reagent were used as Ni,Mg,and Al sources,respectively.Na2CO3(99.8%purity)and NaOH(96.5%purity)were supplied by Shanghai Lingfeng Chemical Reagent.

2.2.Preparation of Ni/Mg–Al catalyst

The Ni/Mg-Alcatalysts were prepared by the co-precipitation method.In a typical procedure,two aqueous solutions,one being the mixture of Ni(NO3)2,Mg(NO3)2and Al(NO3)3,and the other being the mixture of Na2CO3and NaOH,were added slowly into a four-neck flask under strong stirring.The total concentration of Ni2+,Mg2+and Al3+ions in the former solution was 1 mol·L?1,and the amount of each nitrate was predetermined from the desired ratio of Ni/Mg/Alin the final product.For the latter solution,the molar concentration of OH?ion was 2 mol·L?1and the concentration of CO32?ion was half that of Al3+in the former solution.The pH value of the product solution was adjusted to 10 by controlling the dropping speeds of the above two solutions.Upon completion of the addition,the formed precipitates together with the solution were aged at 333 K for 12 h,and then the precipitates were filtered and washed with distillated water until the filtrate was neutral.The precipitates obtained were dried in air at room temperature for 24 h and then at 383 K for 24 h.Finally,the powder was calcined in air at 1123 K for 5 h.The metal composition of the Ni/Mg-Al catalyst was adjusted by varying the initial nitrate concentrations.For convenience,the Ni/Mg-Al catalysts were denoted by Nix/MgyAl(the atomic ratio of elements Ni:Mg:Al was x:y:1).

As a comparison,Ni/γ-Al2O3and Ni/α-Al2O3catalysts were prepared with the incipient-wetness impregnation method by using γ-Al2O3and α-Al2O3as catalyst supports,respectively.The Ni metal loadings for the catalysts(Nix/MgyAl,Ni/γ-Al2O3and Ni/α-Al2O3)were around 15.5%(by mass)determined by inductively coupled plasma optical emission spectroscopy(ICP,Thermo Elemental IRIS 1000).

2.3.Characterization

The surface area and pore size distribution of the prepared catalysts were determined by N2adsorption at?196 °C using a Micromeritics ASAP 2010 instrument.All samples were degassed under vacuum at 150°C for 6 h prior to measurement.X-ray diffraction(XRD)patterns of the catalysts were obtained with a Rigaku(D/Max 2550 VB/PC)diffractometer employing CuKαradiation.The catalysts were also characterized by field-emission scanning electron microscopy(SEM,Hitachi S-4800)and transmission electron microscopy(TEM,JEOL JEM-2010).For the TEM analysis,a carbon-coated Cu grid was dipped into the sample suspension(dispersed in ethanol under ultrasound)and then dried in air.The temperature-programmed reduction(TPR)experiments were conducted on a Micromeritics AutoChem 2920 apparatus using 10%H2in Ar with a heating rate of 10 K·min?1.The consumption of H2was monitored by a thermal conductivity detector(TCD).H2-pulse chemisorption experiments were also carried out to determine the Ni dispersion of catalyst in this apparatus.The chemisorption stoichiometry of H2/Ni was assumed to be 2.The weighed catalysts were first reduced in a mixture of 10%H2/Ar(100 ml·min?1)at 1073 K for 30 min,and then cooled to 308 K in Ar atmosphere.H2pulses were injected into the quartz reactor and the net volume of hydrogen was monitored with a TCD.The metal composition of each catalyst was determined by ICP.

2.4.Activity test

The activity of the prepared catalysts was tested in a quartz reactor(i.d.5 mm),as shown in Fig.1.The catalyst particles were diluted with inert SiC in the mass ratio of 1:10(unless specified otherwise,the catalyst was always diluted with SiC in this ratio),and placed on a quartz holder positioned at the center of the quartz tube.Prior to each test,the catalyst was reduced on-line in the stream of 20%(by volume)H2-N2according to a temperature program:start at room temperature,ramp at 10 K·min?1to 773 K and then at 2 K·min?1to 1073 K,and finally hold at1073 K for 30 min.After reduction,the reactor temperature was adjusted to the desired value,and CH4and H2O were introduced into the reaction system.Note that another stream of N2was directed into the outlet of the reactor,which was used as the internal standard for analysis of product by gas chromatography(GC).The experimental conditions for the catalyst screening were as follows:temperature 923 K(measured in the catalyst bed),pressure 0.1 MPa,molar ratio H2O/CH4of 4,catalyst mass 50 mg,and space time(W/Ft0)from 450 to 1350 g·s·mol?1.

Fig.1.Schematic diagram of experimental apparatus.1—Methane;2—nitrogen;3—hydrogen;4—nitrogen;5—pressure regulator;6—ball valve;7—mass flow controller;8—check valve;9—distilled water;10—HPLC pump;11—evaporator;12—heater and thermal insulator;13—reactor;14—ice bath;15—dryer;16—gas chromatography.

2.5.Kinetic study

The kinetic experiments for steam reforming of CH4over the promising Ni/Mg-Al catalyst were performed in the same reactor as mentioned above.Preliminary tests were conducted in order to determine the experimental conditions under which the effects of external and internal diffusion can be eliminated.The influence of internal diffusion was checked by varying the particle size of catalyst[with five particle sizes,10-20 mesh(average size dP=1.42 mm),20-40 mesh(0.63 mm),40-60 mesh(0.34 mm),60-80 mesh(0.21 mm),and 80-100 mesh(0.16 mm)].The effect of internal mass transfer can be considered negligible when the conversion does not vary with the particle size.The effect of external diffusion was tested at various space time by changing the flow rate of reactant.Two series of tests were performed,the first with 10 mg of catalyst,and the second with 15 mg.In each series the space time was varied and two curves were obtained for conversion vs.space time.If the two curves overlap at certain space time and the conversion is independent of the linear velocity through the catalyst bed,the effect of external diffusion can be neglected[22].During the kinetic experiments,a stream of H2was introduced along with CH4(molar ratio H2/CH4=1)into the reactor in order to prevent reoxidation of the catalyst by steam and to avoid the formation of carbon deposits in the catalyst[23-25].

2.6.Analytic method

The dry product was analyzed on line by a GC(HP-6890)equipped with a TCD.A TDX packed column(2 m)was used for separation.Ar and N2were used as the carrier gas and the internal standard,respectively.The oven temperature was kept constant at 393 K,and the temperatures at the injector and the detector were set at 423 K and 453 K,respectively.

3.Kinetic Model

Primary reactions for conversion of CH4to H2are steam-methane reforming and water-gas shift:

Steam-methane reforming[Reactions(1)and(3)]is highly endothermic while water-gas shift(Reaction(2))is moderately exothermic.Among various kinetic models for steam reforming of CH4,the reaction rate equation presented by Xu and Froment[23]is most widely used.Although it is originally developed over a Ni/MgAl2O4catalyst,this kinetic model is able to predict experimental results on other catalysts such as Ni/CaAl2O4[10],Ni/Al2O3[24-26],Ni/K2O-Al2O3[27],and Rh/CeαZr1?αO2[28].In this work,this model is extended to describe the kinetic behavior of steam methane reforming over the Ni/Mg-Al catalyst.

With Eqs.(1)-(3),the Langmuir-Hinshelwood-type kinetic model is derived as follows[23],

where Pjis the partial pressure of species j,and Kiis the equilibrium constant of the i th reaction.Pjis related with the overall conversion of CH4(αCH4)and the conversion of CH4to CO2(αCO2),

where Eiis the activation energy of the i th reaction,and ΔHjis the adsorption enthalpy of species j.Reaction rates for the disappearance of CH4and for the formation of CO2are calculated by

For a fixed-bed reactor,the continuity equations for CH4and CO2are given by

Eqs.(17)and(18)relate the operation conditions to the kinetic model[Eqs.(4)-(6)].The kinetic and adsorption parameters involved in Eqs.(13)and(14)can be estimated from the experimental data(total experimental runs N=70).The ordinary differential equations are solved by the fourth-order Runge-Kutta method.The Rosenbrock algorithm[29-31]is applied to minimize the objective function,the residual sum of squares between observed αCH4and αCO2at the outlet of the reactor and the calculated values,i.e.,.The initial values for these parameters are assumed to be those obtained by Xu and Froment[23].

4.Results and Discussion

4.1.Structure-activity relationship of the catalysts

In this work,three types of Nix/MgyAl catalysts with different metal compositions are prepared for steam reforming of methane,i.e.,Ni0.38/Mg1.62Al,Ni0.50/Mg2.50Al and Ni0.62/Mg3.38Al.In addition,catalysts Ni/γ-Al2O3and Ni/α-Al2O3are prepared for comparison.The textural properties and the Ni dispersion of these catalysts are listed in Table 1.It can be seen that the Nix/MgyAl catalysts have much higher specific surface area than catalysts Ni/γ-Al2O3and Ni/α-Al2O3.The Ni0.50/Mg2.50Al catalyst possesses the highest specific surface area(177.6 m2·g?1)and Nimetal dispersion(13.5%)among these catalysts,which seemingly indicates that the Ni0.50/Mg2.50Al catalyst will exhibit the highest catalytic activity for steam reforming of CH4.

The N2adsorption-desorption isotherms of these catalysts as well as corresponding pore-size distribution curves are presented in Fig.2.All the catalysts display type IV isotherms with the H3 hysteresis loop according to the IUPAC classification,indicating that the aggregation of plate-like particles gives rise to slit-shaped pores[32].In addition,the pore-size distribution curves of the catalysts except Ni/α-Al2O3show a unimodal distribution of pores mostly falling within the mesopore range,while for Ni/α-Al2O3there are many macropores with the diameter larger than 50 nm as a result of high temperature sintering.

XRD patterns of Ni0.38/Mg1.62Al,Ni0.50/Mg2.50Al,Ni0.62/Mg3.38Al,Ni/γ-Al2O3and Ni/α-Al2O3catalysts are illustrated in Fig.3.All the Nix/MgyAl catalysts after drying(a)exhibit diffraction peaks assigned to hydrotalcite.After calcination at1123 K(b)the hydrotalcite structure completely disappears and the mixed metal oxide(MgxNi1?xO)appears.The peaks of MgxNi1?xO solid solutions become sharper and more intense with the increase of Mg/Al ratio,which indicates that the crystal size of MgxNi1?xO increases as the Ni content decreases in MgxNi1?xO solid solutions.When the calcined Nix/MgyAl catalysts are reduced at 1073 K(c),Ni metal forms due to the reduction of Ni2+in the MgxNi1?xO solid solutions[17,33].As for the Ni/γ-Al2O3catalyst after calcination,a mixture of NiO and γ-Al2O3phases is observed,and after the reduction Ni metal is detected.A similar result is also obtained for the Ni/α-Al2O3catalyst.

Fig.4 shows the SEM and TEM images of Nix/MgyAl catalysts.The synthesized catalyst after drying(a)shows rosette-like morphology,related to the isoelectric point of hydrotalcite(IEP=10).The electrically neutral surface of the initially formed primary particles(pH=IEP)leads to the growth of hydrotalcite along the 001 plane,which finally results in the formation of the rosette-like morphology[34].The black points in the TEM images[Fig.4(b)-(d)]are Ni particles,displaying a quasi-spherical shape,and the mean Ni particle size can be calculated by Σ(nidi)/Σni,where niis the number of particles with a diameter of di.For Ni0.38/Mg1.62Al,Ni0.50/Mg2.50Al and Ni0.62/Mg3.38Al,the mean Ni particle sizes are determined to be 10.7,8.9 and 11.3 nm,respectively.Compared to catalysts Ni0.38/Mg1.62Al and Ni0.62/Mg3.38Al,the Ni particle size of Ni0.50/Mg2.50Al catalyst is the smallest,and correspondingly,its Ni dispersion is the best(Table 1).This result agrees well with those reported by Takehira et al.[33].

Fig.2.N2 adsorption-desorption isotherms and pore size distribution curves of different catalysts.

H2-TPR profiles of various catalysts are shown in Fig.5.For Ni/γ-Al2O3two reduction peaks are observed.The lower temperature peak centered at806 Kisattributed to the reduction of Ni2+in the NiOphase,and the higher peak at 1076 K corresponds to the reduction of Ni2+in the spinel phase(NiAl2O4)[33,35,36].Ni/α-Al2O3,which also does not contain magnesium,exhibits a much more complex TPR profile than Ni/γ-Al2O3and presents at least five peaks.The α-Al2O3support was prepared by calcination of γ-Al2O3at 1373 K.It is most likely that a minority of γ-Al2O3and other Al2O3phases(not detected by XRD due to the small quantity of these phases which may be below the detection limit of XRD)exist in the calcined α-Al2O3,which could result in different interactions between NiO and different Al2O3[12].The highest peak centered at 1122 K represents the reduction of Ni2+in NiAl2O4.For Nix/MgyAl catalysts,only one reduction peak is observed for each catalyst,which is ascribed to the reduction of Ni2+in MgxNi1?xO(see XRD results),and the reduction temperature increases from 1080 to 1100 K with an increase in the Mg/Al ratio,which is also observed by Takehira et al.[15,33],Schulze et al.[35],and Melo and Morlanés[37].This is because the formation of MgxNi1-xO phases is probably enhanced under MgO-rich conditions,which in turn makes the reduction of Ni2+dissolved in MgO more difficult.

The catalytic activities of the prepared catalysts for steam reforming of CH4are presented in Fig.6.The order of the activity is:Ni0.50/Mg2.50Al＞ Ni0.38/Mg1.62Al＞ Ni0.62/Mg3.38Al＞ Ni/γ-Al2O3＞Ni/α-Al2O3.All the Nix/MgyAl catalysts exhibit much higher activity than catalysts Ni/γ-Al2O3and Ni/α-Al2O3,indicating excellent catalytic activity of the Nix/MgyAl catalyst.The order of activity of the catalysts follows the order of Ni dispersion as well as the order of BET surface areagiven in Table 1,implying significant effects of Ni dispersion and specific surface area on the activity of the catalyst[15].Since the Ni0.50/Mg2.50Al catalyst possesses the highest activity for steam reforming of CH4among the catalysts prepared in this work,it is selected for further kinetic investigation.

Table 1 Textural properties and Ni dispersion of the catalysts

Fig.3.XRD patterns of different catalysts(a—after drying;b—after calcination;c—after reduction).

The stability of Ni0.50/Mg2.50Al was evaluated in a 120 h test under a constant reaction condition.As shown in Fig.7,the conversion of CH4is maintained to be about74%during 120 h,indicating good stability of the catalyst.This result is in accordance with that reported by Takehira et al.[15],who stated that no deterioration in the catalytic activity of the Ni0.50/Mg2.50Al catalyst was observed for as long as 600 h of reaction time.

4.2.Kinetic analysis

As shown in Fig.8,the effect of internal diffusion can be neglected when the catalyst particle size is smaller than 0.34 mm.Fig.9 demonstrates that there is no difference in conversion of CH4under two different catalyst loadings of10 mg and 15 mg when the space time is shorter than about 90 g·s·mol?1,so at this space time or shorter the effect of external mass transfer can be ignored.

Based on preliminary tests,70 runs for the intrinsic kinetic experiments were performed under the following conditions:T=823-973 K,Pt=0.1 MPa,W/Ft0＜ 90 g·s·mol?1,molar ratio of H2O to CH4is 1.85-5.85,and dP=0.16 mm.The service time of the Ni0.50/Mg2.50Al catalyst for the kinetic study was about 100 h.At the end of the kinetic experiments,an extra run was conducted to repeat the first run under the same operation condition,and good agreement was found in the conversion of CH4as well as the product distribution for the two runs,implying no catalyst deactivation during the kinetic experiments.This result further demonstrates the good stability of Ni0.50/Mg2.50Al.

Fig.10 presents the overall conversion of CH4and the conversion of CH4to CO2versus the inlet flow rate of CH4.The space time(W/Ft0)ranges from 28 to 72 g·s·mol?1.As expected,both αCH4and αCO2increase with temperature due to the endothermic character of the reforming reaction.At a fixed temperature,αCH4and αCO2decrease as inlet flow rate increases,which resulted from a reduction of residence time of reactants in the catalyst bed.

Fig.11 shows variations ofαCH4andαCO2with molarratio H2O/CH4at different temperatures.W/Ft0is varied between 42 and 86 g·s·mol?1.High H2O/CH4ratio is generally favorable to increase the overall CH4conversion and the conversion of CH4into CO2,but this effect becomes less pronounced when the molar ratio H2O/CH4is greater than 4.0.

The effects of inlet CH4flow rate and molar ratio H2O/CH4on the selectivity to CO2are shown in Fig.12.The selectivity to CO2in this paper is defined as the flow rate ratio of CO2to CO at the outlet of the reactor.It can be seen that at a given temperature the CH4flow rate has less effect on the selectivity to CO2,while the influence of H2O/CH4ratio is very remarkable.This result is reasonable because high concentration of steam helps to produce more CO2from methane directly[Reaction(3)]and inhibit the reverse water gas shift reaction[Reaction(2)].Another important feature is that an increase in temperature decreases the selectivity to CO2,which is caused by the exothermic nature of the water gas shift reaction.

The estimated kinetic parameters are summarized in Table 2.The activation energies of Reactions(1)and(3)over the Ni0.50/Mg2.50Al catalyst are 219.1 and 221.4 kJ·mol?1,which are very close to those obtained by Oliveira et al.over a Ni/Al2O3catalyst[25](217.01 and 215.84 kJ·mol?1)and over a Ni/K2O-Al2O3catalyst[27](218.55 and 236.85 kJ·mol?1),and are also consistent with those reported by Hoang et al.[26]over a sulfide Ni/γ-Al2O3catalyst(209.5 and 211.5 kJ·mol?1).However,the activation energies of the reforming reactions in this work are smaller than the data reported by Xu and Froment[23]over a Ni/MgAl2O4catalyst(240.1 and 243.9 kJ·mol?1),indicating that the Ni0.50/Mg2.50Al catalyst prepared in the present work is more active than the conventional Ni/MgAl2O4catalyst.

As shown in Figs.10 and 11,the calculated overall conversion of CH4and the conversion of CH4to CO2based on the kinetic model agree excellently with the experimental data under various conditions.These comparisons are further presented in Fig.13.Corresponding to the overall conversion of CH4and the conversion of CH4to CO2,the average relative errors between the observed and predicted values over the range of experimental conditions are 6.3%and 5.6%,respectively,demonstrating that the suggested kinetic model can describe very well the steam reforming of CH4over the Ni/Mg-Al catalyst.In addition,the F-value based on the regression sum of squares and the residual sum of squares is found to be 5.8×103,which is much higher than the tabulated F-value of 2.8,indicating the adequacy of the model.

Fig.4.SEM image of synthesized Ni0.50/Mg2.50Al after drying(a)and TEM images of Ni x/Mg y Al after calcination and reduction(b-d).

Fig.5.H2-TPR of different catalysts(a—Ni/γ-Al2O3;b—Ni/α-Al2O3;c—Ni0.38/Mg1.62Al;d—Ni0.50/Mg2.50Al;e—Ni0.62/Mg3.38Al).

The adsorption constants should satisfy the thermodynamic rules proposed by Boudart et al.[38]:

Fig.6.Comparison of the activity of different catalysts(T=923 K,P t=0.1 MPa,molar ratio of H2O to CH4 is 4,W=50 mg,d P=0.16 mm).

Fig.7.Stability test of the Ni0.50/Mg2.50Al catalyst(T=923 K,P t=0.1 MPa,F0CH4=0.89 mmol·min?1,molar ratio of H2O to CH4 is 4,W=50 mg,d P=0.16 mm).

As listed in Table 3,the estimated adsorption constants satisfy the above three criteria.The fourth criterion is that the absolute values of entropy changes for the non-dissociative adsorption are usually higher than 42 J·mol?1·K?1.Obviously,the absolutevalues for CH4,CO,and H2given in Table 3 are greater than 42 J·mol?1·K?1,and thus this rule is also satisfied.

However,it should be noted that the adsorption of steam does not satisfy these criteria,which is also observed by other researchers[23-25,28].Because steam can readily adsorb on different surfaces,not only on the support but also on the nickel surface,the adsorption constant of steam bH2O,which resulted from steam adsorption on the catalyst support,could not be considered a true equilibrium constant.In fact,it only reflects a steady-state condition reached by more than one adsorption steps,and as a result,bH2Odoes not follow the thermodynamic rules[24,28].

Fig.8.Effect of catalyst particle size on the overall conversion of CH4(P t=0.1 MPa,=2.68 mmol·min-1,molar ratio of H2O to CH4 is 4,W=20 mg).

Fig.9.Effect of residence time on the overall conversion of CH4(T=973 K,P t=0.1 MPa,molar ratio of H2O to CH4 is 4,d P=0.16 mm).

Fig.10.Effect of inlet flow rate of CH4 on the overall conversion of CH4(a)and conversion of CH4 to CO2(b)(P t=0.1 MPa,molar ratio of H2O to CH4 is 4,/=1,W=10 mg,=0.16 mm).

Fig.11.Effect of molar ratio of steam to CH4 on the overall conversion of CH4(a)and conversion of CH4 to CO2(b)(P t=0.1 MPa,=1.80 mmol·min?1,/=1,W=10 mg,d P=0.16 mm).

5.Conclusions

The Ni/Mg-Al catalysts obtained from hydrotalcite-type precursors by co-precipitation method show higher activity for steam reforming of methane than catalysts Ni/γ-Al2O3and Ni/α-Al2O3prepared by incipient wetness impregnation.Detailed characterization of the catalysts by N2adsorption-desorption,XRD,TPR,H2chemisorption,SEM,and TEM indicates that the high activity of the Ni/Mg-Al catalysts is closely associated with the high specific surface area and highly dispersed Ni particles after the reduction.Among these Ni/Mg-Al catalysts,the highest activity is obtained when the Ni/Mg/Al molar ratio is 0.5:2.5:1.

The kinetic study with the Ni/Mg-Al catalyst shows that the overall conversion of CH4and the conversion of CH4to CO2increase with the increase of temperature,residence time and molar ratio H2O/CH4,and that the selectivity to CO2is favored by low temperature and high H2O/CH4ratio.It is found that a Langmuir-Hinshelwood kinetic model is thermodynamically consistent and describes the experimental data very well,with the average relative errors for the overall conversion of CH4and the conversion of CH4to CO2over the range of experimental conditions being 6.3%and 5.6%,respectively.It is anticipated that the model can be used for the simulation and optimization of the SESMR process,which makes use of Ni/Mg-Al catalyst and CO2s or bent for hydrogen production.

Fig.12.Variation of the selectivity ofCO2 with the inlet flow rate ofCH4(a)and molar ratio H2O/CH4(b).

Nomenclature

bjadsorption constant of species j,kPa?1(for CH4,CO and H2,and non-dimensional for H2O)

Eiactivation energy of reaction i,kJ·mol?1

Fj0molar flow rate of species j at the inlet of reactor,mol·s?1

Fjemolar flow rate of species j at the exit of reactor,mol· s?1

Ft0total molar flow rate at the inlet of reactor,mol·s?1

ΔH enthalpy change of reaction or adsorption,kJ·mol?1

Kiequilibrium constant of reaction i,kPa2for Reactions(1)and(3),and non-dimensional for Reaction(2)

kirate constant of reaction i,mol·kPa0.5·kg?1·s?1for Reactions(1)and(3),and mol·kg?1·s?1·kPa?1for Reaction(2)

N number of experimental runs

Pjpartial pressure of species j,kPa

Pttotal pressure,kPa

R gas constant,J·mol?1·K?1

r reaction rate,mol·kg?1·s?1

T temperature,K

W catalyst mass,kg

αCH4overall conversion of CH4

αCO2conversion of CH4to CO2

Table 2 Estimated kinetic and adsorption parameters

Fig.13.Comparison of experimental and calculated values for the overall conversion of CH4(a)and conversion of CH4 to CO2(b).

Table 3 Satisfaction of the thermodynamic rules