Jingchun Yan,Weidong Liu,Rong Sun,Shouxi Jiang,Shen Wang,Laihong Shen,*

1 Key Laboratory of Energy Thermal Conversion and Control of Ministry of Education,School of Energy and Environment,Southeast University,Nanjing 210096,China

2 College of Physics and New Energy,Xuzhou University of Technology,Xuzhou 221018,China

Keywords:Biomass Chemical looping gasification Oxygen carrier La-Ni-Fe perovskite Catalysis

ABSTRACT Oxygen carriers(OCs)with perovskite structure are attracting increasing interests due to their redox tunability by introducing various dopants in the structure.In this study,LaNixFe1-xO3(x=0,0.1,0.3,0.5,0.7,1.0) perovskite OCs have been prepared by a citric acid–nitrate sol–gel method,characterized by means of X-ray diffraction(XRD)analysis and tested for algae chemical looping gasification in a fixed bed reactor.The effects of perovskite types,OC/biomass mass ratio (O/B),gasification temperature and water injection rate on the gasification performance were investigated.Lower Ni-doped (0 ≤x ≤0.5) perovskites crystalized in the rhombohedra system which was isostructural with LaNiO3,while those with composition 0.5 ≤x ≤1 crystalized in the orthorhombic system.Despite the high reactivity for LaNiO3,LaNi0.5Fe0.5O3(LN5F5)was found to be more stable at a high temperature and give almost as good results as LaNiO3 in the formation of syngas.The relatively higher syngas yield of 0.833 m3.kg-1 biomass was obtained under the O/B of 0.4,water injection rate of 0.3 ml.min-1 and gasification temperature at 850 °C.Continuous high yield of syngas was achieved during the first 5 redox cycles,while a slight decrease in the reactivity for LN5F5 after 5 cycles was observed due to the adhesion of small grains occurring on the surface of OCs.However,an obvious improvement in the gasification performance was attained for LN5F5 compared to raw biomass direct gasification,indicating that LN5F5 is a promising functional OC for chemical looping catalytic gasification of biomass.

1.Introduction

Global warming induced by doubling atmospheric greenhouse gas(GHG)concentrations since the beginning of the industrial revolution,has been becoming one of the most serious problems that humans are facing with in the 21st century.To deal with GHG emissions,both clean and renewable alternative energy and environmentally friendly fuel conversion technologies are highly required [1].Biomass energy with inexhaustible supply,abundant reserves and low GHG emissions has been proven to be an attractive substitution of fossil fuels to fulfill the growing demand of clean and everlasting energy source [2].Among different biomass conversion routes,the production of synthesis gas from biomass chemical looping gasification (BCLG) is a novel thermochemical conversion technique for upgrading energy density of biomass with lower pollutant emission and exergy loss [2–4].

In the BCLG process,the selection of an appropriate oxygen carrier(OC)is one of the key issues that are attracting a rising concern of many researchers[5].Fe is the most widely used active element in OCs which mainly include monometallic oxide(e.g.Fe2O3[6,7]),natural mineral(e.g.iron ore[8,9])and polymetallic oxide(e.g.ferrites with spinel structure [10–13]),because Fe-based OCs are demonstrated a low cost,high melting point,large reserve and environmental friendliness.Additionally,Ni-based oxide was once considered as the most promising OC in the chemical looping process due to its outstanding redox and high tar elimination ability[14,15].However,some disadvantages limit the widespread use of Ni-based OCs to some extent,including the low strength,toxicity to the environment and health and tendency to deposit carbon[16].In order to combine the strengths of Fe-based and Ni-based OCs,some researchers investigated the NiO-decorated Fe-based OC and the results showed a synergistic effect on enhancing fuel conversion [17,18].Nevertheless,direct impregnation or coprecipitation method is mainly applied for the preparation of modified OCs.One of the drawbacks of these methods is that NiO mainly loaded on the surface of Fe-based OC rather than the interior environment [19].The loss of active ingredient and the build-up of surface carbon may occur during the CLG process leading to the decreased reactivity.For these reasons,it is required to develop new kinds of OC to eliminate the inconveniences encountered with the present OCs in BCLG process.

Recently,perovskite-type oxides with a general formula of ABO3,where A and B are both cations with the lowest limits for cationic radii of rA>0.09 nm and rB>0.051 nm,have been reported to be highly effective for chemical looping applications[20–24].As shown in Fig.1,larger cations located at A-site can be a rare-earth,alkaline earth or alkali metal,which rarely shows catalytic ability but mainly plays a role in supporting the skeleton and regulating the valence of B-site cations.The catalytic performance of perovskite-type oxide is mainly determined by B-site cations which are usually transition metals [25].Perovskite-structured oxides are gaining extensive attention due to their unique tunable bulk and surface properties.By partial substitution of cations in positions A and B,different compounds of the formula AxA’1-xBy-B’1-yO3(0

In our previous work [26],we have studied the performance of LaFeO3as oxygen carrier in the BCLG process.A relatively high syngas yield and stable reactivity of LaFeO3was observed.Moreover,the DFT calculation results showed that CO desorption was identified to be the rate-limiting step.In order to leverage the advantage of Ni in a high catalytic performance and avoid nickel sintering and coke formation simultaneously,it seems to be attractive to develop a tri-metallic La-Ni-Fe perovskite-structured OC by partial substitution of Fe for Ni in position B for the BCLG process.

Fig.1.Ideal cubic structure of perovskite oxides with ABO3-δ formula.

This work aims to give a description of preparation method and characterization of LaNixFe1-xO3perovskite OCs.And BCLG experiments were performed to obtain the optimum reaction conditions including suitable Ni doping content,the mass ratio of OC and biomass (O/B),gasification temperature and water injection rate.The tests help obtain the optimal Ni doping amount for the maximization of syngas yield and the procurement of prolonged reaction activity during the BCLG process.

2.Experimental

2.1.Preparation and characterization of materials

The citric acid–nitrate sol–gel method was applied to synthesize perovskite-type oxides with the general formula LaNixFe1-xO3(0 ≤x ≤1)OCs[27].The stoichiometric amounts of La(NO3)36H2O(AR No.10277-43-7),Ni(NO3)36H2O (AR No.13478-00-7) and Fe(NO3)39H2O (AR No.7782-61-8) were firstly added in a certain volume of distilled water to obtain the desired value of x and heated up to 45 °C for complete dissolution.Then the predissolved citric acid (C6H8O7H2O,AR No.5949-29-1) solution and pure liquid ethylene glycol (C2H6O2,AR No.107-21-1) were added with molar ratio of 1.5:1:1.5 (moles citric acid:moles La+Ni+Fe nitrates:moles glycol).After homogeneous mixing,the aqueous ammonia was added into the mixture drop by drop to precisely regulate the pH of the solution.As can be seen from Fig.2,in the process of adding aqueous ammonia,the precipitation was first produced and then disappeared with the increase of pH value.The clear and transparent solution was obtained when the pH was adjusted to 8.The resulting increase in the pH of the solution containing metal nitrates and citric acid was not only beneficial to the complex formation,but also contributed to the establishment of porous structure in the xerogel that forms upon drying [28,29].This led to a higher combustion rate of the precursor,which is conducive to produce highly uniform oxide particles in size and crystallites size [30].

The resulting solution of the precursor mixture was heated at 80 °C in water bath equipment with magnetic stirring to form a viscous wet gel,which was then dried in a drying oven at 120 °C to produce an xerogel after evaporating the moisture.After that,the xerogel was first pre-combusted at 300°C for 25 min to decompose the polymer,then calcined under the increasing temperature conditions (10oC min-1) and maintained at 750 °C for 6 h to form OC powder.

The phases of as-prepared LaNixFe1-xO3perovskite OCs were detected by X-ray diffraction (XRD) analyses,which were performed in a Rigaku SmartLab diffractometer using Cu Kα radiation(λ=0.15046 nm)with a target current of 30 mA and a target voltage of 40 kV.The scanned diffraction angles (2 theta) varied from 10° to 90°.The surface morphology characteristics of OCs were analyzed by Scanning electron microscope (SEM),which was conducted using an FEI Quanta 400 FEG 450 Scanning Electron Microscope.

Fig.2.Solution appearance varied with different pH values.(a) pH=0.1,(b) pH=2.4,(c) pH=4.6 and (d) pH=8.

The biomass used in this work is algae obtained from Jiangsu Province (China).The proximate and ultimate analyses results are listed in Table 1.The samples were sieved to particles with a size less than 100 μm before experiment.

Table1 Proximate and ultimate analyses of biomass (ad,% (mass))

2.2.Fixed-bed apparatus and gasification tests

The gasification experiments were performed in a fixed bed reactor system,as schematically shown in Fig.3.The reactor system consisted of a straight quartz tube with an inner diameter of 32 mm and a height of 800 mm within an electric heating furnace.

Before each test,0.500±0.005 g of algae was evenly mixed with different amount of OCs according to different O/B (0,0.2,0.4,0.6,0.8,1.0),where OCs were substituted by silica sands when O/B=0.For all runs,the system was purged by N2with a flow rate of 1 L.min-1until the reactor was heated up to the preset value (700°C,750 °C,800 °C,850 °C,900 °C) at a heating rate of 25°C.min-1.When the specified temperature was reached and remained stable,water set at the desired flow rate (0.1 ml.min-1,0.3 ml.min-1,0.5 ml.min-1,or 0.8 ml.min-1) was introduced into the reactor via a syringe pump.The mixture of biomass and OCs was added from the top of the reactor after the temperature reached stability again.The gas productions were collected by gas bags per 2 min and analyzed by a gas analyzer (Emerson,NGA2000)after tar eliminating,dust filtration and water removing,which was detailed reported elsewhere [12].During the oxidation stage,100 ml per min O2mixed with 1 L.min-1N2was introduced into the rector to combust the remaining carbon and complete the regeneration of OCs.

For multiple cycle experiments,the reduction period lasted for at least 20 min and oxygen carrier particles were oxidized for 15 min each cycle.Before biomass was introduced,the system was purged with nitrogen for O2replacement.

2.3.Data evaluation

During the reduction stage,the molar flow rates of outlet gas component(ni,out,i=CO,CO2,CH4,H2,O2)can be calculated based on the nitrogen balance method [31]:

The cumulative molar flow rates of outlet gases are defined as:

The volume fractions of outlet gases can be calculated as:

Fig.3.Scheme of the fixed bed experimental apparatus.1—N2 cylinder;2—O2 cylinder;3—mass flow meter;4—mass flow controller;5—syringe pump,LD-P2020;6—heating tape;7—heating tape temperature controller;8—charging chute;9—furnace;10—thermocouple;11—quartz tube reactor with an inner diameter of 32 mm and a height of 800 mm;12—porous plate with quartz wool;13—ice bath;14—allochroic silicagel;15—gas bags;16—gas analyzer,Emerson,NGA2000.

where ni/mol (i=CO,CO2,CH4,H2) are the cumulative molar flow rates of outlet gases derived from the Eq.(2).

The instantaneous carbon conversion rate can be calculated as:

The carbon conversion efficiency at time t/min of the reduction stage is the fraction of total carbon contained in the fuel converted to carbonaceous gases,which is defined as below:

with mbio/g the mass of biomass initially added in the reactor,φC,bio/% (mass) the carbon content of the biomass and MC/g.mol-1the molar weight of C.

The syngas yield which is generally used to evaluate the gasification performance,can be calculated as:

The lower heating value of the syngas Qs/kJ.m-3and the gasification efficiency ηscan thus be defined as follows [3,12]:

with Ci/% (i=CO,H2,CH4) the relative concentration of syngas and QLHV/kJ.kg-1biomass)the lower heating value of algae,which can be obtained from the proximate and ultimate analyses data shown in Table 1.

For a better understand of the crystal structure of prepared OCs,crystal structure parameters were calculated.The lattice parameters a,b and c can be obtained according to the relationship between interplanar crystal spacing and lattice parameters by jade 5.0 software.The grain size was calculated by Scherrer equation which is illustrated as:

where D/nm is the grain size,θ/rad is the diffraction angle,λ/nm is the wavelength of an X-ray(λ=0.154178 nm),K is a function of the crystallite shape but is generally taken as constant being about 0.89 and B/rad is peak width at a particular value of 2θ.

3.Results and Discussion

3.1.Characterization of as-prepared perovskites

The compositions of as-prepared OC were analyzed repeatedly and consistent results were obtained,indicating the good repeatability of the synthesis method.The XRD results of LaNixFe1-xO3(0 ≤x ≤1) series were shown in Fig.4.It can be seen that a perovskite structure with a well-developed diffractogram was obtained.The corresponding LaNixFe1-xO3(0 ≤x ≤1) phases were detected,indicating the perovskite structure with the expected Ni/Fe ratio.For LaNiO3OC,a trace of La2NiO4which might come from the decomposition of LaNiO3,was detected [34].However,as seen from a zoom view on the 2θ area between 31.8°and 32.9°(Fig.5),a regular and progressive widening of the peaks,decrease of peak intensity and shift of the diffraction peak to a higher angle with increasing x were observed,indicating the formation of a solid solution.

Fig.4.XRD patterns of LaNixFe1-xO3 structures for (a) x=1;(b) x=0.7;(c) x=0.5;(d) x=0.3;(e) x=0.1;(f) x=0.

Fig.5.Zoom XRD patterns of LaNixFe1-xO3 structures for (a) x=1;(b) x=0.7;(c)x=0.5;(d) x=0.3;(e) x=0.1;(f) x=0 in the 2θ area between 31.8° and 32.9° in Fig.4.

Table 2 gives the results of lattice parameters for each x from the six most intensive diffraction peaks.While for x=0,i.e.LaFeO3,the structure is orthorhombic with a space group of Pnma(62).The end member of the series,LaNiO3belongs to the rhombohedral system with a space group of(167).It appears clear that a change from orthorhombic to rhombohedral symmetry is observed with the increasing of x.The cell parameters a,b and c,which are the lengths of each side of parallelepiped unit cell,determines the shape and the size of the unit cell[35].As seen in Table 2,there is an overall decrease of cell parameter a and grain size D as the nickel content increases.And this result could be easily understood considering the ironic radii of Ni3+(low spin) and Fe3+(high spin)which are 0.56 ? and 0.645 ?,respectively [32].The symmetries difference observed in the perovskite series with different Ni/Fe ratios can be illustrated by the so-called Jahn-Teller effects,described as the Goldschmidt’s tolerance factor (t):

where rA,rB,and rOare the radii of the respective ions[33].An ideal cubic perovskite has a tolerance factor of 1.For t values slightly less than 1,a rhombohedral structure is observed(e.g.LaNiO3),whereas an orthorhombic structure can be anticipated for smaller t (e.g.LaFeO3) [21,25].The last column of Table 2 lists the results of thevalues t for different compositional ranges of La-Ni-Fe-O perovskite series.The 12-coordinated effective ionic radius of La3+(1.36 ?),the 6-coordinated radii for Ni3+(LS) (0.56 ?),Fe3+(HS) (0.645 ?) and 1.40 ? for O2-are applied in the calculation [32].It is not hard to find that an orthorhombic structure forms for values of the tolerance factor between 0.954 and 0.974 whereas the rhombohedral phase is more stable for t>0.974.For t=0.974,i.e.the LaNi0.5Fe0.5-O3compound,both orthorhombic and rhombohedral phases are found.

Table2 Lattice parameters for the LaNixFe1-xO3 series perovskites prepared at 700 °C (1?=0.1 nm)

3.2.Screening of perovskites for BCLG

The performances of six kinds of perovskites for algae chemical looping gasification were studied with silica sands as bed material for comparison.The gasification products were obtained under the N2flow rate of 1 L.min-1,O/B of 0.4,water injection rate of 0.3 ml.min-1and gasification temperature at 850°C.As seen from Fig.6a,compared to the biomass direct gasification with silica sands as bed material,the perovskite series played a positive role in promoting gas production and carbon conversion.Furthermore,an obvious increase of syngas yield and gasification efficiency was observed in Fig.6b.Taking six different nickel-containing perovskite OCs into consideration,a small amount substitution of iron ions by nickel irons in position B(0LaNi0.5-Fe0.5O3>LaFeO3>LaNi0.7Fe0.3O3>LaNi0.1Fe0.9O3>LaNi0.3Fe0.7O3>silica sands (pure biomass).These results suggest that partially substituting Fe by Ni can increase its oxygen mobility and availability,thus leading to the higher reactivity.In an effort to understand the mechanism and evolution of the system during the gasification experiments,the perovskites were characterized by XRD after reduction.Furthermore,the OCs were exposed to a flux of oxygen(100 ml.min-1)for at least 15 min after each run to burn off any coke that may have deposited on the particle surface and regenerate the OCs.The XRD patterns of used and re-oxidized OCs were shown in Fig.7a and b.According to the results presented in Fig.7a,three less active perovskites LaFeO3,LaNi0.1Fe0.9O3and LaNi0.3Fe0.7O3remained unchanged after reaction.For used perovskites with 0.5 ≤x ≤1,the XRD patterns showed that NiO was detected,confirming the migration of nickel from bulk phase to the surface.For LaNiO3,La2NiO4as well as NiO was observed.This may be attributed to the LaNiO3phase decomposition at about 850 °C and the reduction of LaNiO3according to[34,36]:

After oxidization,LaFeO3,LaNi0.1Fe0.9O3and LaNi0.3Fe0.7O3kept the crystal structure and LaNi0.5Fe0.5O3restored the perovskite as the NiO phase decreased.However,the perovskite structure of LaNiO3had been totally transformed into La2O3,La2NiO4and NiO.It was noteworthy that the free nickel metal was not detected by XRD,either because the amount was too low,or the nickel was well dispersed or amorphous.

Fig.6.Comparison of (a) gas production and carbon conversion and (b) syngas yield and gasification efficiency of different perovskites.(N2 flow rate:1 L.min-1,OC and biomass mass ratio:0.4,water injection rate:0.3 ml.min-1,gasification temperature:850 °C).

Fig.7.XRD analyses of (a) used and (b) oxidized LaNixFe1-xO3 OCs for (1) x=1;(2) x=0.7;(3) x=0.5;(4) x=0.3;(5) x=0.1;(6) x=0.

Fig.8.Effect of OC and biomass mass ratio(O/B=0,0.2,0.4,0.6,0.8,1.0)on(a)gas products fraction and carbon conversion,(b)syngas yield and gasification efficiency,(c)cumulative CO production and (d) cumulative H2 production.(N2 flow rate:1 L.min-1,water injection rate:0.3 ml.min-1,gasification temperature:850 °C).

According to the XRD results of reduced OCs,it was not surprising to see that the reaction activity of La-Ni-Fe perovskites had a direct correlation with the amount of nickel left the structure during BCLG process.The gasification behaviors of the perovskite series with 0 ≤x ≤0.3 were comparable.The LaNi0.1Fe0.9O3and LaNi0.3Fe0.7O3systems gave a lower carbon conversion and syngas yield,because of the stabilising effect of iron addition on the structure under reductive conditions [37].Relatively high activities to syngas yield were obtained with an increasing x possibly due to the mitigation of nickel from the bulk of the mixed perovskite to the surface.This suggested that the nickel-rich perovskites were less stable during gasification tests than the nickel-poor ones.As mentioned in the introduction section,Ni-based OC was widely concerned owing to its high reaction activity and oxygentransport ability.The formation of NiO from the perovskite structure promoted the gasification process.Although the highest carbon conversion and gasification efficiency were obtained for LaNiO3,the stability of LaNiO3was not very high under the reducing atmosphere according to (R1) and (R2).The regeneration ability is one of the key criteria for appropriate OCs.The perovskite structure of LaNiO3was totally destroyed and unrecoverable during gasification process while the structure of LaNi0.7Fe0.3O3cannot be fully recovered.It should be noted that LaNi0.5Fe0.5O3OC gave almost as good results as LaNiO3in the formation of syngas.Besides,the crystalline structure basically remained unchanged and the perovskite regenerated after oxidization.It must be pointed out that there was a slight carbon deposition observed for LaNiO3during oxidation.However,in other cases,no carbon production was measured even after 10 redox cycles,indicating that the alloying of Ni with Fe improved the coke resistance [38].In light of the high reactivity and stability properties,LaNi0.5Fe0.5-O3which was denoted as LN5F5,would be used as OC in order to study the optimal gasification conditions.

3.3.Effect of OC and biomass mass ratio (O/B)

To investigate the effect of O/B on the gasification performance,samples with different LN5F5 loading ratios were tested,as shown in Fig.8a–d.It can be found from Fig.8a that the addition of LN5F5 enhanced the production of H2,which may be ascribed to the promotion effect of LN5F5 in the bio-char gasification(R3),water–gas shift(WGS)reaction(R4)and tar reforming(R5)and(R6).With the increase of O/B,a slight change of gas products fraction was observed.However,the total syngas yield notably increased as seen in Fig.8b.The highest carbon conversion,syngas yield and gasification efficiency were obtained at an O/B of 0.4.A less syngas yield was produced at a lower O/B ration due to the insufficient oxygen supply for volatiles.Nevertheless,the excessive addition of OC would block the contact of atmosphere and volatiles [39].The cumulative CO production was in the order of O/B=0.4 >O/B=0.2>O/B=0.8>O/B=1.0>O/B=0.6>O/B=0,while the cumulative H2production was in the order of O/B=0.4 >O/B=0.8 >O/B=0.6>O/B=0.2>O/B=1.0>O/B=0,as shown in Fig.8c and 8d.The difference in CO and H2accumulation productions were mainly related to the total syngas yield,since slight difference in the concentrations of CO and H2were seen in Fig.8a.For O/B=0.4,both CO and H2production were greatly enhanced compared with raw biomass,leading to a high-LHV syngas yield.

Fig.9.Effect of gasification temperature (700 °C,750 °C,800 °C,850 °C,900 °C) and water injection rate (0,0.1 ml.min-1,0.3 ml.min-1,0.5 ml.min-1,0.8 ml.min-1) on (a)carbon conversion,(b) gasification efficiency,(c) syngas yield and (d) H2/CO ratio.(N2 flow rate:1 L.min-1,O/B=0.4).

3.4.Effect of gasification temperature and water injection rate

Experimental trials were performed under different gasification temperature and water injection rates in the fixed-bed reactor,in order to evaluate the catalytic gasification performance of LN5F5.Fig.9 showed carbon conversion,gasification efficiency,syngas yield and H2/CO ratio as functions of temperature and water injection rate for the reactivity tests with LN5F5.From Fig.9a and b,it could be seen that when the water injection rate was low,carbon conversion and gasification efficiency increased with the rising temperature monotonically.This could be explained by the fact that higher temperature was advantageous to enhance endothermic reactions,such as biochar gasification (R3),Boudouard reaction (R7),methane steam reforming (R8),as well as tar catalytic reforming on the active sites of perovskite surface (R5) and (R6),resulting in the increase of syngas production(see Fig.9c).In addition,more lattice oxygen was released under a high temperature,not only improving the gasification process,but also producing lots of oxygen vacancies [40].During BCLG,the Ni3+ions was reduced to Ni2+as the reactions progressed and a charge compensation was required so as to achieve electroneutrality,which can be achieved by the formation of oxygen vacancies.The diffusion rate of lattice oxygen in the bulk phase of LN5F5 accompanied by the redox of transition metals were improved,because the oxygen consumed via surface reactions could be supplemented by the diffusion of oxygen in the bulk phase due to the exist of oxygen vacancies.The electron transfer process between reactants and OCs was accelerated,resulting in the high catalytic gasification performance of LN5F5.However,with the continuing increase of water injection rate,carbon conversion and gasification efficiency reached the peak at 850°C and then began to decrease at a higher temperature.This result could be contributed by a combination of factors:(1) Under high temperature,a large amount of water injected into the reactor vapored rapidly leading to a dramatic increase in gas velocity.Thus part of macromolecules in tar left the reactor without adequate contact with active sites of the perovskite.The unreacted tar molecules led to the decrease of carbon conversion and gasification efficiency since these two parameters were defined by the carbonaceous gases which could be directly detected by the gas analyzer.(2) The increase of steam may cover some active sites of the perovskite and poison the OCs,hindering the catalytic reactions taking place on the surface of OCs [41].(3)A portion of heat may be taken away through water evaporation,lowering the in-situ gasification temperature and thus reducing carbon conversion and gasification efficiency.Therefore,there were optimal values for gasification temperature and water injection rate that allowed the syngas production to be maximized.

Fig.10.Cyclic performance over 10 cycles of LN5F5 for algae gasification.(a)Gas product fractions and total gas yield,(b)syngas yield,(c)carbon conversion and gasification efficiency and (d) phase evaluation of LN5F5 after 1-cycle,5-cycle and 10-cycle.(N2 flow rate:1 L.min -1,O/B=0.4,water injection rate:0.3 ml.min -1,gasification temperature:850 °C).

Fig.9d gave the results of H2/CO change as a function of gasification temperature and water injection rate.The H2/CO had a downward tendency as temperature increased and water injection rate decreased,because the lower temperature and higher water injection rate promoted the WGS equilibrium (R4) to the right.However,as mentioned,too much steam injected into the reactor would give rise to a sharp acceleration in gas velocity,leading to the inadequate contact between reactants and active sites of the OCs.Consequently,a slight decrease of H2/CO was observed when the steam volume fraction further increased.According to the results of Fig.9,the highest carbon conversion,gasification efficiency and a relatively high syngas yield of 0.833 m3.kg-1biomass can be achieved when the gasification temperature was 850°C and water injection rate was 0.3 ml.min-1.

3.5.Cyclic reduction and oxidation

Multiple cycles of algae gasification and LN5F5 oxidation experiments were conducted in the fixed bed under optimal conditions,i.e.N2flow rate of 1 L.min-1,O/B of 0.4,gasification temperature of 850 °C and water injection rate of 0.3 ml.min-1.As observed in Fig.10a and b,during the first 5 cycles,the fraction of syngas exhibited an increasing trend while the CO2concentration decreased as the cycle numbers increased.Both total gas yield and syngas yield rose and gave the highest values of 1.106 m3.kg-1biomass and 0.907 m3.kg-1biomass at 4th cycle,respectively.However,with the continuing increase of cycle numbers,the reactivity of LN5F5 began to decrease and maintained stable after 7th cycle.The carbon conversion and gasification efficiency also showed similar up and down trends,seen in Fig.10c.

For the purpose of deeply investigating the cause of gasification performance change for LN5F5 in different cycles,phase changes and morphological and structural features of OCs were characterized by XRD and SEM analyses,presented in Figs.10d and 11.It can be seen that LaPO4and NiO phases as well as perovskite were detected for LN5F5 after gasification reactions.And the peak intensity of LaPO4was relatively higher for the OC after 5 redox cycles.The formation of LaPO4may be related to the higher phosphorus content in algae.Algae itself contained high concentrations of phosphate compounds as algae needed phosphate to exist.During the gasification process under high temperature(850°C),40%–70%of phosphorus was released in the gaseous form[42,43],which further reacted with La2O3formed after perovskite release oxygen(R9) to form stable LaPO4(R10) [44].LaPO4with high thermal and chemical stability was proven to be effective to decompose macromolecules such as chlorobenzene,phenol and ethanol which may exist in tar,either as a catalyst or support material [45].As mentioned,NiO also had a certain catalytic effect on biomass gasification process.The accumulation of LaPO4and NiO on the OC led to the continuous improvement of syngas yield,carbon conversion and gasification efficiency in the first 5 cycles.Moreover,there was no obvious agglomeration and sintering on the surface of OC after 5 cycles,as seen in Fig.11b,and the OC maintained good reactivity.However,for LN5F5 after 10 cycles,a lower peak intensity of LaPO4phase was observed,either because LaPO4was highly dispersed on the surface of OC,or because part of phosphorus turned into amorphous forms.In addition,as can be seen from Fig.11c,the adhesion of small grains occurred after 10 cycles making the surface of OC become smooth and pore diameters smaller.And this result,to some extent,prevented either gaseous reactants or oxygen penetrating into the OC and hindered the gas–solid reactions and OC regeneration,resulting in the decrease in reaction activity of LN5F5.Despite the reduction of activity for LN5F5 after 10 cycles,it was noteworthy that the catalytic gasification performance for LN5F5 was still obviously improved compared to that raw algae gasification with silica sands as bed materials.Furthermore,the formation of LaPO4also provided a potential effective method for treating fuels with high phosphorus content,such as sewage sludge with La-Ni-Fe perovskite series as OCs in the chemical looping applications [46].

Fig.11.The SEM images of LN5F5 perovskite:(a) LN5F5 after 1 cycle (10,000× and 50,000×),(b) LN5F5 after 5 cycles (10,000× and 50,000×),(c) LN5F5 after 10 cycles(10,000× and 50,000×).

4.Conclusions

LaNixFe1-xO3(0 ≤x ≤1) perovskite OCs were prepared with a citric acid–nitrate sol–gel method and characterized on their crystalline structure and reactivity performance for the BCLG process.The addition of nickel into the perovskite structure may improve its catalytic performance but decease the stability of the perovskite phase under a high temperature.Transformation in crystalline structure from orthorhombic system(0 ≤x ≤0.5)to rhombohedra system(0.5 ≤x ≤1)was observed with the increasing introduction of nickel.Although LaNiO3showed a high catalytic performance in the algae gasification,it was unstable at the high temperature and was difficult to be regenerated.As a result,LaNi0.5Fe0.5O3(LN5F5)possessed activity as good as LaNiO3at the temperature and atmosphere typically encountered in the process of algae chemical looping gasification.The effect of operating conditions including O/B,gasification temperature and water injection rate were investigated.As high as 0.833 m3.kg-1biomass syngas yield was obtained under the conditions of temperature at 850°C,water injection rate of 0.3 ml.min-1and OC and biomass ratio of 0.4.Moreover,LN5F5 was found to be stable after 10 redox cycles.The LaPO4and NiO phases were present after gasification tests,explaining the improvement of OC during the first 5 cycles.Despite the slight reduction of activity for LN5F5 after 10 cycles,it should be noted that the catalytic gasification performance for LN5F5 was still obviously improved compared to raw algae direct gasification and no carbon deposition was observed.As a whole,the inhibition of carbon deposition and increase in syngas production illuminated the effectiveness of LN5F5 as a functional OC for the biomass chemical looping gasification process.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The authors gratefully acknowledge the support of this research work by the National Natural Science Foundation of China(51761135119)and the Scientific Research foundation of Graduate school of Southeast University (YBPY1906,YBJJ1606,YBJJ1703).

Nomenclature

a lattice parameter,(1?=0.1 nm)

B peak width at a particular value of 2θ,rad

b lattice parameter,?

Cirelative concentration of syngas,%

c lattice parameter,?

D grain size,nm

fivolume fractions of outlet gases,%(vol)

Gssyngas yield,m3.kg-1

K function of the crystallite shape

MCmolar weight of C,g.mol-1

mbiomass of biomass initially added in the reactor,g

nicumulative molar flow rates of outlet gases,mol.min-1

ni,outmolar flow rates of outlet gas component,mol.min-1

QLHVlower heating value of algae,kJ.kg-1

Qslower heating value of the syngas,kJ.m-3

rAcationic radius of element A in ABO3,nm

rBcationic radius of element B in ABO3,nm

rOcationic radius of O in ABO3,nm

T0room temperature,°C

t tolerance factor

XCcarbon conversion efficiency,%

Xi,outgas volume concentration in the outlet gas,%(vol)

xC,redinstantaneous carbon conversion rate,mol min-1

ηsgasification efficiency,%

θ diffraction angle,(°)

λ wavelength of an X-ray,nm

φC,biocarbon content of the biomass,% (mass)