Performance characteristics of fluidized bed syngas methanation over Ni-Mg/Al2O3 catalyst☆

Jiao Liu ,Dianmiao Cui,2,Jian Yu ,Fabing Su ,Guangwen Xu ,*

1 State Key Laboratory of Multi-phase Complex Systems,Institute of Process Engineering,Chinese Academy of Sciences,Beijing 100190,China

2 Graduate University of Chinese Academy of Sciences,Beijing 100864,China

Keywords:Syngas methanation Ni catalyst Fluidized bed Substitute natural gas Performance

ABSTRACT The performance characteristics of isothermal fluidized bed syngas methanation for substitute natural gas are investigated over a self-made Ni-Mg/Al2O3 catalyst.Via atmospheric methanation in a laboratory fluidized bed reactor it was clarified that the CO conversion varied in 5%when changing the space velocity in 40-120 L·g?1·h?1 but the conversion increased obviously by raising the superficial gas velocity from 4 to 12.4 cm·s?1.The temperature at 823 K is suitable for syngas methanation while obvious deposition of uneasyoxidizing Cγoccurs on the catalyst at temperatures around 873 K.From a kinetic aspect,the lowest reaction temperature is suggested to be 750 K when the space velocity is 60 L·g?1·h?1.Raising the H2/CO ratio of the syngas increased proportionally the CO conversion and CH4 selectivity,showing that atenough high H2/COratios the active sites on the catalyst are sufficient for CO adsorption and in turn the reaction with H2 for forming CH4.Introducing CO2 into the syngas feed suppresses the water gas shift and Boudouard reactions and thus increased H2 consumption.The ratio of CO2/COin syngas should be better below 0.52 because varying the ratio from 0.52 to 0.92 resulted in negligible increases in the H2 conversion and CH4 selectivity but decreased the CH4 yield.Introducing steam into the feed gas affected little the CO conversion but decreased the selectivity to CH4.The tested Ni-Mg/Al2O3 catalyst manifested good stability in structure and activity even in syngas containing water vapor.

1.Introduction

Complete methanation of syngas for the production of substitute natural gas(SNG)is attracting great attention especially in China because of its anticipated effectiveness for using the coal resource in remote areas and its possible role for the national supply security of natural gas fuel.From coal to SNG is also a relatively simple process,and realizes high energy efficiency and low water consumption so that it is indeed competitive with many other coal conversion technologies such as coal to liquid and coal to DME[1].Since the 1970s a number of methanation processes including fixed and fluidized bed methanation reactors[2]have been developed.Among them,the adiabatic fixed bed methanation process has been well developed,which consists of several fixed bed reactors in series and several intermediate gas coolers,usually with the product gas recycling and the reactant gas feed splitting.The fluidized bed reactor is inherently isothermal and suitable for the highly exothermic reactions[3-5].Therefore,the syngas methanation to SNG is possibly implemented in a single fluidized bed reactor or with a tail-end clean-up fixed bed reactor[6].Thus,a few studies on fluidized bed methanation have been well done in the laboratory and pilot scales[7-9].

The studies conducted for laboratory syngas methanation can be classified into three categories:optimization of the catalyst composition and formation of catalyst structure[10-17],design of methanation reactor and process[9,18-20],and study on kinetics and hydrodynamics[7,21,22].The syngas methanation in fixed and fluidized bed reactors requires completely different catalysts.While the catalyst for fixed bed methanation is commercially available,there is no catalyst ready for fluidized bed methanation.Thus,the Institute of Process Engineering(IPE),Chinese Academy of Sciences(CAS)has worked extensively on developing a fluidized bed methanation catalyst.A new method integrating the acid-base pairing and hydrothermal treatment,instead of impregnation,was successfully developed to prepare the highly attrition-resistant and also high-temperature tolerant powder-type Ni-Mg/Al2O3catalyst[23].Now it is in the stage of applying this catalyst in industrial pilot test for fluidized bed methanation.There is thus an urgent need in understanding the performance characteristics of syngas methanation over the developed Ni-Mg/Al2O3catalyst in the fluidized bed reactor.On the other hand,there is almost no literature on the performance characteristics of isothermal fluidized bed syngas methanation.

The operation of two pilot plants,which guaranteed the flexibility and operability of the Lurgi methanation process[24],demonstrated that the methanation process should be operated in small variations of the H2/CO ratio and with low residual CO2in the feed gas.Meanwhile,from the thermostability aspect of the applied BASF catalyst,the outlet temperature is usually limited to 723-743 K and the pressure in the methanation reactor is governed by the upstream gasification pressure.

The present work is devoted to investigating the atmospheric methanation performance in a range of parameters including reaction temperature,gas velocity and feed gas composition in terms of the CO conversion,selectivity to CH4and carbon deposition on the spent catalysts.The above-mentioned Ni-Mg/Al2O3catalyst was tested by varying the reaction temperature in 673-900 K,space velocity from 40 to 120 NL·g?1·h?1(N indexes the standard conditions),superficial gas velocity from 4 to 12.4 cm·s?1(under experimental conditions),H2/CO ratio in 1.8-3.8,and amount of CO2or steam into the feed gas.The spent catalysts were characterized by the temperature programmed oxidation(TPO)and N2physisorption to understand their carbon deposition features and physical properties,respectively.These characterizations provided also justification for the clarified performance characteristics of fluidized bed methanation over the catalyst.

2.Experimental

2.1.Catalyst preparation

The Ni-Mg/Al2O3catalyst used in the tests was prepared by a method integrating the acid-base pairing and hydrothermal treatment.First,15.8 g of nickel nitrate[Ni(NO3)2·6H2O],12.9 g of magnesium nitrate[Mg(NO3)2·6H2O]and 13.0 g of alumina nitrate[Al(NO3)3·9H2O]were dissolved in 200 ml of distilled water to get an acidic solution with a total metal ion concentration of 0.7 mol·L?1.Meanwhile,29.9 g of NaAlO2was separately dissolved in 200 ml of distilled water to form an alkaline solution.These two solutions were then simultaneously added dropwise into a vessel kept at 313 K under continuous mechanical stirring.The addition rates of solutions were regulated to maintain the reaction mixture at pH=11.The precursor was further transferred to a Teflon-lined stainless steel autoclave that was in turn heated to 473 K and maintained at this temperature for the hydrothermal treatment for 10 h.Afterwards,the reactor was cooled down to room temperature in air,and the formed precipitate was collected by filtration and washed thoroughly using distilled water.The catalyst was finally obtained by drying at 353 K for 10 h and calcined at 773 K for 4 h.As shown in Table 1,the prepared Ni-Mg/Al2O3catalyst exhibits relatively low NiO mass content(19%)and is rich in meso-pores(average pore size:14.3 nm).The dispersion of NiO on the Al2O3support for the tested Ni-Mg/Al2O3catalyst was confirmed by TEM image shown in Fig.1(a).One can figure out from the image that the NiO particles had roughly sizes of 15 nm and are highly dispersed in the catalyst.

Table 1 Major properties of the self-made Ni-Mg/Al2O3 catalyst

The obtained catalyst was crushed and sieved into particles of 180-230 μm(65-80 mesh),and further fluidized in a cold-model fluidized bed for about 60 h to improve the sphericity in order to reduce the catalyst lose during fluidized bed methanation.The final size distribution of the granular catalyst characterized with a particle size analyzer(Mastersizer 2000,UK)is presented in Fig.1(b),showing a relatively narrow size distribution of 100-250 μm.

2.2.Methanation test

Fig.2 presents a schematic diagram of the experimental apparatus for atmospheric fluidized bed methanation tests.A 1.5 g catalyst was loaded into a quartz reactor(16 mm i.d.)with a porous quartz plate distributor.The reactant gas mixture,a simulated syngas,was prepared via online mixing the gases of H2,CO,N2and CO2controlled with mass flow meters.The test procedure started with reducing the catalyst in a 0.040 L·min?1N2-base gas flow containing 10%(by volume)H2at 923 K for 4 h,adjusting the temperature of the catalyst bed to the designated reaction-starting temperature,then switching the gas stream to the simulated syngas for methanation.The reaction system usually reached steady state in a few minutes,and in order to ensure the validity of data the measurement for the product gas started at 30 min using a micro GC(Agilent 3000,USA).Before being sent to the GC,the effluent gas from the reactor was cooled in an ice-water condenser and then dewatered in a column filled with silica gel.

The CO conversion and selectivity to CH4are defined respectively as

where X is the conversion of CO,S is the selectivity to CH4and f is the volumetric flow rate of the gas species(either CO or CH4)that was determined from the GC-analyzed composition and the total gas flow rate estimated by taking the well-metered flux of N2as the internal standard.

2.3.Characterization of the catalyst

At the end of each experiment,the heating furnace was turned off to cool the catalyst bed naturally to room temperature by keeping the reactant gas feed.TPO for the spent catalyst was carried out using an automated chemisorption analyzer(Chem-BET Pulsar TPR/TPD,Quantachrome,USA).A sample of 50 mg was first placed in a quartz U-tube reactor and dried in a He stream at 373 K for 30 min.After cooling the pretreated catalyst sample to ambient temperature,the TPO test was performed by heating the sample from 373 K to 923 K at a rate of 5 K·min?1in a N2-base gas containing 10%(by volume)O2and then holding the sample at 923 K for 60 min to fully oxidize the carbon species on the catalyst.The emitted CO2in the TPO test was monitored on-line with a mass spectrometry(Proline,AMETEK,USA).The surface area and pore size of the catalysts were estimated from the N2physisorption curve measured at 77 K via Autosorb-1(Quantachrome,USA).For this measurement the catalyst sample was thoroughly degassed at 573 K in advance.

3.Results and Discussion

3.1.Conditional determination

3.1.1.Varying reaction temperature

Fig.1.(a)TEM image and(b)particle size distribution of the Ni-Mg/Al2O3 catalyst after fluidization in a cold-model bed for about 60 h.

Fig.2.A schematic diagram of atmospheric fluidized bed methanation apparatus.1—H2;2—CO;3—N2;4—H2/N2;5—mass flow meter;6—rotor flow meter;7—check valve;8—mixer;9—furnace,10—quartz reactor;11—condenser;12—silica-gel dryer;13—GC.

Fig.3.Performance of atmospheric fluidized bed methanation at different reaction temperatures(SV=60 L·g?1·h?1,syngas:H2/CO/N2=3/1/1).

Fig.3 shows the performances of atmospheric methanation influidized bed reactor at different reaction temperatures.The space velocity was 60 L·g?1·h?1and syngas composition was subject to a molar H2/CO/N2ratio of 3/1/1.In Fig.3(a),the experimental CO conversion and selectivity to CH4are compared with their thermodynamic equilibrium values calculated using the Gibbs free energy minimization method[25]considering the reactions of methanation[reactions(3)and(6)],water gas shift[reaction(4)]and carbon deposition[reactions(5)and(7)]:

Limited by the thermodynamics,the experimental CO conversion and selectivity to CH4were lower than their equilibrium values and decreased with rising temperature except that at 700 K.At this low reaction temperature the kinetic rate of the reaction should be low and thus cause the low CO conversion.Above 850 K,the resulting CO conversion was almost equal to the equilibrium value.

All that time, I talked to the sad animal, telling her about my life and her mother and anything I could think to say. When my throat was dry from it, I simply stroked her white muzzle22, or brushed her chestnut23 brown coat until we both slept.

Fig.3(b)compares the time traces of experimental conversion of CO and selectivity to CH4at different reaction temperatures.The performances were evidently stable in the 5-h test at all the temperatures,showing that the catalyst has good stability at temperatures up to 873 K.The data in Fig.3(b)also verified further the effect of reaction temperature revealed in Fig.3(a)and the CO conversion was lowest at 673 K.

Fig.4.TPO profiles of spent Ni-Mg/Al2O3 catalysts from atmospheric fluidized bed methanation at different temperatures in Fig.3(b).

The TPO profiles of the spent catalysts from the tests in Fig.3(b)are shown in Fig.4.It reveals obviously different species and amounts of carbon deposited on the spent catalysts after methanation at different temperatures.From the CO2-release peaks one can see that there are at least two types of carbon deposited on the catalysts.The lower temperature peak could be attributed to the oxidation of reactive carbon Cβ(releasing CO2at 523-750 K),whereas the higher-temperature one ascribed to the oxidation of less-reactive carbon Cγ[19,26,27].The integrated peak area of CO2release increased with the rising reaction temperature,suggesting that higher methanation temperature caused higher carbon deposition and transformation of reactive Cβto hardoxidizing Cγ.In fact,Fig.4 shows that there was little carbon Cγbelow 823 K.The less-reactive carbon Cγwould encapsulate the metalparticles and plug micropores and mesopores to decrease the active sites on the catalyst.On the other hand,the formed Cγat the rear side of metal crystallites would cause the active metal to break away from the catalyst support to deactivate the catalyst as well[28].Although the catalytic performance is stable at 873 K in Fig.3(b),the syngas methanation over Ni-Mg/Al2O3should be at temperatures below 823 K in order to depress the formation of uneasy-oxidizing carbon Cγ.

3.1.2.Varying gas velocities

The evaluation under different space velocities(SV)was performed by adjusting the syngas flow from 1 to 3 L·min?1while fixing the catalyst amount at 1.5 g as shown in Fig.5.These tests were carried out continuously at SV varying from 40 to 120 L·g?1·h?1and maintained 30 min at each point.The conversion of CO only decreased from 86.7%to 82%for the inefficient contact of catalyst surface with reactant gas as elevating the SV from 40 to 120 L·g?1·h?1,indicating that the fluidized bed reactor allows a large gas treatment capacity.As expected,the selectivity to CH4exhibited a stable value but it was lower than the equilibrium-allowed values.

Fig.5.Performance of atmospheric fluidized bed methanation at different space velocities(T=823 K,syngas:H2/CO/N2=3/1/1).

The influence of superficial gas velocity on the fluidized bed methanation at 823 K and an SV of 60 L·g?1·h?1is shown in Fig.6.For the 100-250 μm catalyst particles,the calculated terminal velocity is 38 cm·s?1and minimum fluidization velocity is 0.6 cm·s?1(under experimental conditions).The superficial gas velocity in the tests was varied in 4 to 12.4 cm·s?1to ensure the reactions in the expected thermostatic zone of the furnace and reactor.The result shows that the conversion of CO increased from 77%to 86%while the selectivity to CH4was not obviously changed.The higher superficial gas velocity led to higher bed voidage and released more active surface of the catalyst particles by loosing the particle contact in the fluidized bed.As a result,the influence of external diffusion was weakened and the gas-catalyst interaction,refreshment of particle surface and mass transfer between gas and solid phases were all enhanced to elevate the CO conversion,as is shown in Fig.6.

Fig.6.Performance of atmospheric fluidized bed methanation over Ni-Mg/Al2O3 at various superficial gas velocities(T=823 K,SV=60 L·g?1·h?1,H2/CO/N2 ratio:3/1/1).

The preceding results show that the fluidized bed methanation exhibits a good tolerance to capacity or space velocity fluctuation.The space velocities adopted here are much higher than that used in actual industry.With deeper catalystbed the conversion and selectivity would be close to their equilibrium values.In order to raise the utilization efficiency of the exothermic heat,high methanation temperature is favored to generate high quality steam.For comprehensive consideration of the activity and stability of the tested catalyst,the methanation of syngas should be performed at temperatures of 750-823 K for avoiding the formation of uneasy-oxidizing Cγas temperature increased up to 873 K.The impaction of feedstock on methanation was thus tested herein at 823 K and a space velocity of 40 L·g?1·h?1.

3.2.Feedstock impaction

3.2.1.H2/CO ratio in feed gas

For methanation,the produced syngas from gasifying coal and biomass has usually unfavorable H2/CO ratios of 0.3 to 1.8.A water gas shift unit is thus used to adjust the H2/CO ratio at the upstream of methanation.Increasing this ratio would promote the methanation for high CH4yield and inhibit the WGS reaction for low CO2formation.The excessive H2in the feed gas would be also effective to suppress the carbon deposition due to the promoted reaction of carbon with H2[25].Fig.7 shows the methanation performance varying with the H2/CO ratio of the feed gas based on a N2flow of 0.625 L·min?1.There is little effect on the calculated equilibrium conversion of CO(82.6%to 81.2%)but the equilibrium selectivity to CH4increases indeed from 71.2%to 80.5%with the rise of the H2/CO ratio from 1.7 to 3.7.The experimental selectivity to CH4showed the same variation tendency.

Fig.7.Performance of atmospheric fluidized bed methanation over Ni-Mg/Al2O3 at different H2/CO ratios(T=823 K,SV=40 L·g?1·h?1,f N2=0.625 L·min?1).

3.2.2.Introducing CO2 into feed gas

Introduction of CO2into feed gas would make the water gas shift(WGS)and Boudouard reactions shift to their left sides and suppress the conversion of CO to CO2and carbon deposition,while CO2perhaps reacts with H2to form CH4.The influence of CO2concentration in the feed gas(by replacing a part of its N2)was studied at a bed temperature of 823 K,with CO of 0.076 L·min?1,H2of 0.228 L·min?1and SV=40 L·g?1·h?1.As shown in Fig.8,the conversion of CO(●)decreased while the selectivity to CH4relative to CO(○)was improved with the increasing CO2/CO ratio.Nonetheless the consumption ofH2(?)increased with the rising CO2/CO ratio.

Fig.8.Performance of atmospheric fluidized bed methanation over Ni-Mg/Al2O3 at different CO2/CO ratios(T=823 K,SV=40 L·g?1·h?1,syngas:f CO=0.076 L·min?1,H2/CO:3/1).●CO conversion,▲ equilibrium CO conversion,? H2 conversion,○ selectivity to CH4,△equilibrium selectivity to CH4,□CH4 yield from CO.

Table 2 compares the difference of CO2flux between outlet and inlet at different CO2/CO ratios,here ΔFeqand ΔFexpare for the equilibrium and experimental values,respectively.A positive ΔF means that the CO2flux at the outlet is higher than at the inlet.Thus,the outlet CO2flux at the outlet was higher than that at the inlet when CO2/CO ratiowas low,but this increased CO2flux at the outlet gradually decreased with the raised CO2/CO ratio.When the experimental CO2/CO ratio was 0.92,the CO2flux at the outlet became lower than that at the inlet(negative ΔFexp).This shows that the consumed CO2by CO2methanation exceeded the produced CO2via WGS.The transition based on the calculated equilibrium CO2flux between the outlet and inlet occurred at a lower CO2/CO ratio,about0.5(i.e.,ΔFeqbecoming negative).This shows the fact that the experimental performance did not reach the equilibrium state.

Table 2 Difference of CO2 flux between outlet and inlet of reactor at different CO2/CO ratios

Considering the CH4yield from CO,the yield was stable at56%-58%in varying the CO2/CO ratio in 0 to 0.52 but decreased to 47%at the CO2/CO ratio of 0.92.Consequently,the ratio of CO2/CO in the feed gas is better to be lower than 0.52 for increasing the H2conversion and selectivity to CH4.At rather higher CO2/CO ratio,it has negligible effect on the selectivity to CH4.

3.2.3.Introducing steam into feed gas

Steam was introduced by passing the feed gas(H2/CO/N2)through the hot water in a bottle packed with Raschig rings and the resulted concentration of steam is about3.4%(by volume).In the cases with and without steam,the flow of H2and CO was fixed at 0.228 L·min?1and 0.076 L·min?1,respectively.Fig.9 shows the results of atmospheric fluidized bed methanation at 823 K.The presence of steam in the feed gas affected slightly the CO conversion from 85.6%to 81.7%(Fig.9(a))but the selectivity to CH4decreased from 76%to 64%(Fig.9(b)).The equilibrium calculation revealed the same influence of steam,although the absolute equilibrium CO conversion and selectivity to CH4are higher than the experimental values.Consequently,the introduction of steam into the syngas suppressed the methanation butpromoted the WGS reaction,which lowered in turn the CH4production and the selectivity to CH4.

Water vapor is thought to be conducive in accelerating the sintering rate of active metals and support of catalyst by forming some mobile surface hydroxyl groups which would be subsequently volatilized at high temperatures[28].Thus,the spent catalysts after methanation with and without steam in the feed gas were characterized with N2physisorption.Compared with the fresh catalyst(Table 1),the surface areas of the spent catalysts(Table 3)obviously decreased.This may be attributed to the agglomeration of metallic nickel during reduction and the deposition of carbon which blocks some pores of the catalyst[32].Similar results were obtained in the previous studies by Loc et al.[12]and Luna and Iriarte[33].The introduction of water vapor to the reaction system affected little the physical properties and the stability of the spent Ni-Mg/Al2O3catalyst.Analyzing the total amounts of deposited carbon(Table 3)on the spent catalysts clarified that adding a small amount of steam into the CO methanation system could obviously reduce the carbon deposition.This agrees with the HICOM methanation process[34]in which excessive steam is applied to the first methanation reactor in order to avoid carbon deposition.

Table 3 Major properties of Ni-Mg/Al2O3 catalyst after methanation with and without steam in the feed gas

4.Conclusions

Various reaction conditions were systematically investigated to clarify the limit of major operating parameters for fluidized bed methanation over a self-made Ni-Mg/Al2O3catalyst containing 19%(by weight)NiO and rich in meso pores(average pore size:14 nm).The fluidized bed methanation exhibited a good tolerance to variation in treatment capacity so that varying the space velocity in 40 to 120 L·g?1·h?1did not affect much the methanation performance.Controlled by the thermodynamics,the experimental CO conversion and selectivity to CH4decreased with the raised reaction temperature.The fluidized bed methanation was shown to be better to be at temperatures of 750 to 823 K considering both reaction kinetics and stability(less carbon deposition)of the catalyst.

Investigating the influence of reactant gas composition on atmospheric fluidized bed methanation at 823 K demonstrated that with increasing the H2/CO molar ratio the CO conversion increased and tended to approach the equilibriums.This indicates essentially that the CO dissociation on the catalyst surface is easy to occur and the reaction of the dissociated CO with H2is the rate determining step.When CO2is introduced into the feed gas,the CO conversion decreased while the H2conversion and selectivity to CH4became higher,suggesting that the molar ratio of CO2/CO in the feed gas is better to be lower than 0.52 to gain higher H2conversion and CH4yield.When steam(about 3%-4%)was introduced into the feed gas,there was little effect on the CO conversion but the selectivity to CH4decreased from 77%to 67%.The tested self-made Ni-Mg/Al2O3catalyst showed also good stability in structure and activity in the atmosphere containing water vapor.

Fig.9.Performance of atmospheric fluidized bed methanation over Ni-Mg/Al2O3 with steam in feed gas(T=823 K;SV=40 L·g?1·h?1;syngas:f CO=0.076 L·min?1,H2/CO:3/1).