Ca ion-exchanged coal char as H2S sorbent

Ca ion-exchanged coal char as H 2S sorbentBenjamin Garcia *,Takayuki TakaradaDepartment of Biological and Chemical Engineering,Gunma University,1-5-1Tenjin-cho,Kiryu 376,Gunma,JapanAccepted 29September 1998AbstractH 2S removal using Ca ion-exchanged coal char was studied in a fixed bed reactor at a temperature of 900ЊC.Yallourn,an Australian brown coal was used for ion exchange.The CaO obtained from the pyrolysis of Ca ion-exchanged coal was finely dispersed in the char and much more reactive toward H 2S than that from limestone.Ca ion-exchanged coal char showed high H 2S capture and almost no emission of H 2S was observed in the earlier reaction stage.At a reaction time of 100min,the S/Ca atomic ratio in Ca ion-exchanged coal char increased to 1.2.This was due to H 2S capture capacity of both the coal char itself and the CaO distributed in the char.The sorption capacity of char was determined as 1.5mmol-H 2S/g-char.The desulfurization characteristics of calcined limestone strongly depended on the particle size,whereas those of Ca ion-exchanged coal char was almost independent of the char particle size.᭧1999Elsevier Science Ltd.All rights reserved.Keywords:Ca ion-exchange;Ultra-fine CaO particles;H 2S removal1.IntroductionCoal gasification is a versatile method of coal utilization where the coal gas generated can subsequently be used to produce energy.Clean coal technology has become a stan-dard phrase in recent days.It has generally been taken as referring to methods of producing power and heat in which low emissions of sulfur and nitrogen oxides are included into the design [1].The topping cycle (or advanced-PFBC,as it is named in Japan)is one of the advanced coal-firing technologies for power generation which combines features of systems based on gasification and combustion.The main advantage of the topping cycles over the other systems is that topping cycles can exploit the potential of high inlet temperature gas turbines,while minimizing the energy losses incurred in the production of fuel gas.It is well known that H 2S is the most abundant sulfur-containing compound emitted from coal gasification.Therefore,it is necessary to remove the H 2S as much as possible from the coal gasification stream to protect the turbines and related equipment from the corrosive action of this compound.Hot coal desulfurization,which uses solid sorbents such as oxides of metals (e.g.of iron,manganese,zinc,and vana-dium)and limestones,possesses several technical and economic advantages over low temperature desulfurization processes.Limestones and dolomites are the most common sources of calcium oxide used as sulfur sorbent.However,its low conversion makes it necessary to oversupply the amount of sorbent to obtain an acceptable sulfur capture [2].The ther-mal decomposition of calcium carbonate and the sulfidation of the calcium oxide produced can be described by the over-all reactions:CaCO 33CaO ϩCO 2 1 CaO ϩH 2S 3CaS ϩH 2O2The reaction of CaO with H 2S is of technical interest in both gasification [3,4]and fuel rich combustion systems [4,5]when CaO is used as a sulfur sorbent.Experiments performed in TGA showed a weak influence of temperature on the initial sulfidation rates,and low activation energies were obtained (5.16kcal/mol [6]and 3.6kcal/mol [7]).The sulfidation experiments realized by Westmoreland et al.[6]gave a linear relation between the initial reaction rate of reagent grade CaO and the H 2S concentration in the reactive mixture at temperatures in the range of 300ЊC–900ЊC.Evangelos and Stratis [8]found that highly porous sorbent particles are produced from limestone or dolomite calcina-tion and,as a result,the overall rate of sulfidation reaction is in general influenced by both the transport in the intraparti-cle space and surface reaction.Moreover,since the solid product (CaS)in the reaction Eq.(2)has larger molar volume than the solid reactant (CaO),various phenomena associated with pore structure (such as pore plugging andFuel 78(1999)573–5810016-2361/99/$-see front matter ᭧1999Elsevier Science Ltd.All rights reserved.PII:S0016-2361(98)00183-5*Corresponding author.Tel.:ϩ81-277-30-1450;fax:ϩ81-277-30-1454.formation of inaccessible pore space)may play an important role in determining the reactivity of the solid with the progress of its sulfidation.Therefore,the sorbent utilization strongly depends on its particle size[9],and high utilization will be expected by using sorbents with small particle size. Low rank coals such as brown coal and lignite have a large amount of oxygen-containing functional groups and the proton on carboxyl groups is easily exchanged with metal cations.When Ca-exchanged low rank coal is pyro-lyzed,the carboxylate salt dissociates,releasing carbon dioxide and leaving highly dispersed calcium compounds such as CaO and CaCO3.The chemical form of the calcium compound depends on the pyrolysis conditions.It is well known thatfinely dispersed calcium compounds have a quite large catalytic effect for char gasification in O2,CO2 and steam.In the present study,a new desulfurization technique using Ca ion-exchanged coal as an H2S sorbent in the topping cycle system was proposed.That is to say,the Ca ion-exchanged coal with raw coal is fed into the gasifier, where the coal gasification and the desulfurization take place.A highly effective desulfurization is expected by using ultrafine CaO particles produced from the Ca ion-exchanged char.The remaining chars and the produced CaS are transferred into thefluidized bed combustor, where the chars are burnt and the CaS is oxidized to CaSO4for safe disposal.The scope of this study does not cover the aspect of the CaS oxidation.Fine limestone parti-cles of sub-micron meter size are able to be prepared by mechanical grinding,but an enormous energy for grinding will be needed.Moreover,sincefine particles easily make aggregates,it is quite difficult to handlefine particles in the dispersion state.In this technique,there is no problem in handlingfine CaO particles,becausefine CaO particles are incorporated into large coal char particles.In the topping cycle system,the chemical form of calcium compounds in a gasifier depends on the gasification conditions.Both CaO and CaCO3can react with H2S according to reaction Eq.(2) and the following reaction,respectively.CaCO3ϩH2S3CaSϩH2OϩCO2 3 Equilibrium calculations revealed that CaCO3was calcined to CaO under conditions where there was a gasifi-cation temperature of over900ЊC and0.12CO2partial pres-sure[10].In this paper,the reactivities of CaO prepared from Ca-exchanged brown coal and limestone toward H2S were investigated.2.Experimental2.1.MaterialAs the ion-exchange capacity of coal depends on the amount of oxygen-containing functional groups present in the brown coal matrix[11],low rank coals are the most appropriate coals to be used as ion-carriers.Yallourn coal, an Australian brown coal,was selected in this study as an ion-exchanged coal.Carboxyl and hydroxyl groups in the coal were determined as1.7and6.0mequiv/g(daf)of coal [12].The limestone used was supplied by Yoshizawa Lime Industry Co.Ltd.It was from Kuzu,Tochigi prefecture in Japan.The analyses of coal and limestone are shown in Tables1and2,respectively.2.2.Ca ion exchangeThe calcium addition to the coal matrix was achieved by ion-exchange with calcium hydroxide.Yallourn coal(20g) was soaked in a milk of lime which was prepared by mixing 10g of Ca(OH)2with1000ml of deionized water.The Ca-ion exchange was carried out by stirring the slurry with a magnetic stirrer at room temperature for24h.The pH of the mixture was kept in the range of11.8–12.0.The coal was separated from the solution byfiltration,then washed with deionized water to remove the excess of Ca(OH)2and dried under vacuum at107ЊC.The calcium from Ca ion-exchanged coal(YL–Ca coal)was extracted using a diluted solution of hydrochloric acid(HCl:10%).The amount of calcium incorporated in the Yallourn coal was determined by atomic absorption analysis(Shimadzu AA-6400F atomic absorptionflame emission spectrometer).The calcium content in the sample was9.6wt.%.2.3.Reaction with H2SThe pyrolysis of YL–Ca coal,as well as the limestone calcination,were carried out in afixed bed reactor at900ЊC for10min to decompose the carboxylic sites,and provide CaO crystallites in the plete pyrolysis of YL–Ca coal or limestone calcination was achieved in10min.The pyrolyzed YL–Ca coal(YL–Ca char)was stored in a sealed flask and kept in a desiccator until it was used.Limestone calcination was done just before each experiment.B.Garcia,T.Takarada/Fuel78(1999)573–581 574Table1Analysis of coal.The particle size of the sample used was0.25–0.5mm,unless otherwise statedSample Ultimate analysis(wt.%,daf)C H N S OYL a coal65.1 4.60.60.329.4YL(Ca)b char87.4 1.50.80.210.1a YL Yallourn.b YL(Ca)char Calcium ion-exchanged coal char,15.6%calcium.Table2Analysis of limestone(wt.%).The particle size of the sample used was0.25–0.5mm,unless otherwise statedIg.loss SiO2Fe2O3Al2O3CaO MgO42.11 2.160.100.0654.240.23The desulfurization was carried out in a reactor made of quartz,with an inner diameter of 2.5cm and length of 75cm.A plate of sintered quartz filter in the reactor was used as a gas distributor and sample support.The schematic diagram is shown in Fig.1.The sample of YL–Ca char (0.75g)or limestone (0.3g)put into the reactor was heatedelectrically under nitrogen atmosphere at 20ЊC/min to the desired temperature.The temperature was kept for 10min before introducing the reactant gas.The reactant gas consisted of a mixture of 1500ppm hydrogen sulfide and nitrogen.It was fed into the upper part of the reactor.The experiment was allowed to continue for 100min.Since the SO 2analyzer cannot realize direct measurements of hydro-gen sulfide,oxygen was fed just under the plate of sintered quartz filter with the purpose of converting the unreacted H 2S into SO 2.The exhaust gas from the reactor was continu-ously analyzed with an SO 2analyzer (Yanako,EIR-500S),and with an infrared gas analyzer (Shimadzu CGT-7000).After 100min,the flows of H 2S,as well as oxygen,were suspended,and then the reactor was cooled down under nitrogen stream to prevent undesired reactions of the product with the atmosphere.Calcium conversion was defined as the moles of H 2S captured,divided by the moles of calcium present in the sample.XC in ϪC outC Ca4where X is the apparent calcium conversion,C in is the moles of H 2S at the inlet of the reactor,C out is the moles of H 2S at the outlet of the reactor,and C Ca is the content of calcium (moles)in the sample.The characterization of the sample was carried out by X-ray diffraction (XRD)(Rigaku Geigerflex),scanning elec-tron microscopy (SEM)(JEOL JSM-5300LV)and nitrogen adsorption (Quantachrome Autosorb 1).3.Results and discussion3.1.Characterization of Ca ion-exchanged coal char 3.1.1.Crystallinity of Ca compoundsYL–Ca coal was pyrolyzed in the temperature range of 700ЊC–1000ЊC and the crystallinity of the calcium compounds in the char was determined by XRD analysis.Fig.2shows the XRD patterns for YL–Ca chars prepared at different temperatures in a N 2atmosphere.No diffraction peaks were observed for YL–Ca char prepared at 700ЊC,in spite of a high calcium content of about 16wt.%.This result suggests that the calcium compounds in this stage were finely dispersed in the char.The dispersion mode of calcium compounds observed in this study was quite similar to those reported by some researchers [13,14].Radovic et al.[15]carried out XRD measurements of calcium-exchanged North Dakota lignite char prepared at 702ЊC and reported that no diffraction peaks attributable to calcium compounds were observed.Ohtsuka and Tomita [13]observed a similar dispersion mode for calcium-loaded coal chars prepared using Ca(CH 3OO)2,Ca(NO 3)2,Ca(OH)2and CaCl 2.They also reported that,in a nitrogenB.Garcia,T.Takarada /Fuel 78(1999)573–581575Fig.1.Schematic diagram of fixed bedreactor.Fig.2.XRD patterns of YL–Ca char at different pyrolysis temperatures in an N 2atmosphere.576B.Garcia,T.Takarada/Fuel78(1999)573–581Fig.3.Pore distribution of Ca ion-exchanged Yallourn coal after heat treatment.Fig.4.SEM images of Yallourn coal char(a,b)and Ca ion-exchanged Yallourn coal char(c,d).Pyrolysis:900ЊC,10min,in N2.atmosphere,CaCO3is decomposed to CaO in the temperature range650ЊC–700ЊC and the decomposition is promoted by carbon.When YL–Ca coal was pyrolyzed at higher temperatures,the broad diffraction peaks assigned to CaO appeared as shown in Fig.2.The average crystallite size of CaO prepared at800ЊC and900ЊC was evaluated by the Debye–Scherrer method as20and28nm,respectively. These results mean that ultra-fine CaO particles are easily prepared by the pyrolysis of calcium ion-exchanged brown coal.The intensity of the CaO peak was increased with an increase in the temperature,due to the sintering of the CaO crystallites[16].3.2.Specific surface areaPhysical properties of samples were determined by nitro-gen adsorption.The BET surface area of YL–Ca char and calcined limestone prepared at900ЊC was214and19m2/g, respectively.The pore distribution profile of YL–Ca char is shown in Fig.3.It was found that the micro and meso pores were developed in the char.The large specific surface area of YL–Ca char observed is mainly due to the surface area attributed to micro pores.From these results,it can be said that YL–Ca char was a highly porous material and well suited to uniformly disperse CaO in the char matrix.3.3.SEM observationThe SEM images of Yallourn coal char(YL char)and YL–Ca char prepared at900ЊC are shown in Fig.4.The outer surfaces for both samples were relatively rough,and no definite difference in the morphology between two chars was observed[Fig.4(a)and(c)].A careful observation, however,revealed that small particles of several micro-meters in diameter were often observed on the YL–Ca char[Fig.4(d)].The particles may be attributable to calcium compounds.In this study,calcium was loaded from a milk of lime and the calcium loading was9.1wt.%.It is possible that a part of the calcium adheres on the char surface and the pore surface in the char and,on heating,the unexchanged calcium compound forms small particles,as observed with SEM.However,judging from the results of the XRD measurements shown in Fig.2,it seems that the content of unexchanged calcium compounds in YL–Ca char is quite low.3.4.Desulfurization3.4.1.Effect of sorbent typeFig.5shows typical profiles of H2S removal for both of the sorbents studied.Limestone showed a considerable reduction of H2S emissions in thefirst10min,although the calcium conversion achieved was only12wt.%.After this period,the emissions gradually increased.The high H2S capture observed for limestone at the early reaction stage can be due to the nascent CaO particles formed by the thermal decomposition of CaCO3.Evangelos and Stratis [8]showed that the CaO produced by the thermal decom-position of CaCO3is highly porous and gives a large surface area for the reaction.As the reaction progresses,it can be supposed that the CaS produced was plugging the pores limiting the gas diffusion and,as an outcome,the H2S capture decreased.On the other hand,YL–Ca char showed a high capacity for H2S capture,and almost no H2S emission was detected for about45min.The desulfurization effi-ciency of YL–Ca char was higher than that of limestone. The H2S removals achieved at100min of the experiment were47%and80%for limestone and YL–Ca char,respec-tively.These results agree with those obtained from the experiments of fuel-rich combustion of coals containing calcium realized by Freund and Lyon[5].The CaO produced by the pyrolysis of Ca ion-exchanged coal wasB.Garcia,T.Takarada/Fuel78(1999)573–581577Fig.5.H2S removal profile for limestone and YL–Ca char at900ЊC.finely dispersed on the char and the char was a highly porous material,as mentioned previously.The ultrafine CaO particles may considerably reduce the influence of product layer diffusion resistance and the reactant gas may easily diffuse through the pores in the char.These properties of YL–Ca char are the main reasons for the high performance of YL–Ca char as a desulfurizing agent.3.5.XRD patterns of calcium speciesThe XRD analyses of reacted limestone and YL–Ca char at different reaction times give information on the progress of CaS formation.Fig.6shows the XRD patterns of calcined limestone during the reaction with H 2S.The diffraction peaks assigned to CaO and Ca(OH)2were observed in the sample before the reaction.It is considered that Ca(OH)2is formed by the reaction between CaO and water vapor in an air atmosphere.At a reaction time of 30min,strong peaks attributable to CaS were clearly seen.After that,the peak intensity from CaS gradually increased and that from CaO gradually decreased with an increase in the reaction time.The XRD pattern at 100min was very similar to that at 60min.From these results,it is considered that,in the early stage of the reaction immedi-ately after the calcination of limestone,the CaO produced is highly porous and the capture of H 2S can be carried out efficiently before pore plugging occurs and,in the latter reaction stage,the CaO from limestone is deactivated due to the product layer formed on the surface of the CaO particle.The XRD patterns of YL–Ca char under the same condi-tions exhibited different behaviors during the reaction of CaO with H 2S,as shown in Fig.7.The gradual increase in the CaS peak intensity and the decrease in the CaO peak intensity with an increase in the reaction time,indicate that the reaction was uninterrupted during the experiment.The ultra-fine CaO particles produced from YL–Ca char seems to diminish the reaction limitations observed for the CaO from limestone.B.Garcia,T.Takarada /Fuel 78(1999)573–581578Fig.6.XRD patterns for limestone during the H 2S removal at 900ЊC.Fig.7.XRD patterns for YL–Ca char during the H 2S removal at 900ЊC.3.6.Calcium conversionThe conversion of CaO to CaS was calculated by means of gas analysis,as mentioned previously.The apparent calcium conversion achieved for calcined limestone and YL–Ca char after 100min was 70%and 120%,respec-tively.The conversion of over 100%observed for the YL–Ca char is due to the fact that the char can capture a part of the H 2S.In order to determine the sulfur retention for the char,the Yallourn coal (0.75g)was pyrolyzed under a nitrogen atmosphere at 900ЊC for 10min,and the H 2S capture was done under the same experimental conditions used for the other sorbents.Fig.8shows the profiles obtained for YL char and YL–Ca char.It is clear that the char is able to retain considerable amounts of H 2S.The amount of sulfur retained in the YL char and YL–Ca charat a reaction time of 100min were 1.5×10Ϫ3and 4.7×10Ϫ3mol/g-char,respectively.Assuming that the H 2S capture by YL char and CaO contained in YL–Ca char independently progresses in paral-lel,the calcium conversion at 100min was estimated as 88%.However,in order to obtain precise calcium conver-sions in YL–Ca char,further investigations on the carrier effects of carbonaceous material of char on the reaction between CaO and H 2S will be needed.3.7.Effect of particle sizeIn the case of limestone,it is well known that the sorbent utilization strongly depends on the particle size.The use of sorbent with a small particle size gives a large surface area and small resistance to the product layer diffusion.ToB.Garcia,T.Takarada /Fuel 78(1999)573–581579Fig.8.YL char and YL–Ca char profiles for the capture of H 2S at 900ЊC.Fig.9.Effect of particle size of limestone during the capture of H 2S.observe the effect of particle size of YL–Ca char on the desulfurization behavior,YL–Ca chars with mean particle sizes of 0.38,0.20and 0.11mm were prepared.Limestone was also used for comparison.The profiles of the H 2S capture for different particle sizes of calcined limestone are shown in Fig.9.The reaction profile of CaO with H 2S was strongly influenced by particle size.The H 2S removal increased with the decrease in the particle size.This obser-vation agrees with that reported by Evangelos and Stratis [3].High H 2S removal was observed for all the samples used at the beginning of the reaction.The smallest particle (0.11mm)showed high activity during 40min,keeping the emissions at about 100ppm.However,comparing with YL–Ca char,it is noteworthy that in the early stage of the reaction,the H 2S concentration at the outlet gas was much higher than that observed from YL–Ca char.The smallest particles (0.11mm)maintained a high reaction rate until they achieved almost 100%conversion.The large particles showed a high initial reaction rate,but the reaction rate rapidly declined,as is shown in Fig.9.This high initial rate can be interpreted as evidence that the reac-tion is carried out on the nascent CaO surface and large H 2S capture can be done before the product layer forms.This interpretation is consistent with the behavior of the reaction between CaO and SO 2,which has been shown to occur throughout the pore structure [17].The calcium conversion achieved after 100min of reaction for the largest particle size used (0.75mm)was 43%.The product layer seems toB.Garcia,T.Takarada /Fuel 78(1999)573–581580Fig.10.Effect of particle size of YL–Ca char during the capture of H 2S at 900ЊC.Fig.11.Dispersion of CaO in YL–Ca char at different particle sizes.be an important resistance to the gas diffusion through the pore when sorbent with large particle size is used.As is illustrated by Fig.10,YL–Ca char is much more effective for retaining H2S by all the particle sizes used than calcined limestone.All samples with different sizes achieved similar high H2S capture.In order to investigate the relation between the particle size of YL–Ca char and the distribution of CaO particles in the char,measurement of the CaO dispersion were realized.The analysis of CaO disper-sion was carried out according to the method developed by Linares-Solano et al.[18],where CO2is reacted with CaO at 300ЊC.At this temperature,CO2reacts only on the surface of CaO,therefore,knowing the amount of reacted CO2,it is possible to calculate the dispersion of CaO.Fig.11shows the CaO dispersion of three different particle sizes of YL–Ca char.Almost no variation in the results is observed, indicating that the particle size of YL–Ca char does not affect the dispersion of CaO.This result agrees with that of Fig.9,corroborating that the particle size and dispersion of CaO is independent of the particle size of the YL–Ca char.4.ConclusionsThe CaO from the calcium ion-exchanged coal was shown to be a highly efficient H2S removal agent.The reac-tivity of CaO prepared from Ca ion-exchanged coal is much higher than that prepared from limestone.The use of Ca ion-exchanged coal char provides high sorbent utilization.The H2S removal is not a function of the particle size of the Ca ion-exchanged coal char,because ultra-fine,highly reactive CaO particles are produced from Ca ion-exchanged coal independently of its particle size.This characteristic can make it a useful sorbent to be used in topping cycles systems.References[1]Topper JM,Cross PJI,Goldthorpe SH.Fuel1994;73:1056.[2]Krishnan SV,Stratis VS.Ind.Engng Chem.Res.1994;33:1444.[3]Squires AM.Science1970;169:821.[4]Moss G,Craig JWT,Tisdall D.AIChE Symp.Ser.1972;68(126):277.[5]Freund H,Lyon bustion Flame1982;45:191.[6]Westmoreland PR,Gibson JB,Harrison DP.Environ.Sci.Technol.1977;11:488.[7]Kamath VS,Petrie TW.Environ.Sci.Technol.1981;115:966.[8]Evangelos EA,Stratis VS.Ind.Engng Chem.Res.1992;31:2311.[9]Borgardt RH,Roache NF,Bruce KR.Ind.Engng Chem.Fundam.1986;25:411.[10]Yrjas P,Lisa K,Hupa M.Fuel1996;75:89.[11]Lafferty C,Hobday M.Fuel1990;64:78.[12]Takarada T,Nabatame T,Ohtsuka Y,Tomita A.I&EC Res.1989;28:505.[13]Ohtsuka Y,Tomita A.Fuel1986;65:1653.[14]Yamashita H,Nomura M,Tomita A.Energy Fuels1992;6:656.[15]Radovic LR,Walker Jr.PL,Jenkins RG.J.Catal.1983;82:382.[16]Takarada T,Mao K,Hashimoto Y,Nakagawa N,Kato K.JSME-ASME International Conference on Power Engineering.1993,p.97.[17]Borgardt RH.Environ.Sci.Technol.1970;4:59.[18]Linares-Solano A,Almela-Alarco´n M,Salinas-Martine´z de Lecea C.J.Catal.1990;125:401.B.Garcia,T.Takarada/Fuel78(1999)573–581581。