Cao Kuang, Shuzhong Wang,*, Ming Luo, Jun Zhao,*

1 Key Laboratory of Thermo-Fluid Science and Engineering of MOE, School of Energy and Power Engineering, Xi’an Jiaotong University, Xi’an 710049, China

2 School of Energy and Power Engineering, Jiangsu University, Zhenjiang 212013, China

Keywords:Reaction kinetics Chemical looping with oxygen uncoupling(CLOU)Sintering Natural ore CO2 capture

ABSTRACT In the chemical looping with oxygen uncoupling (CLOU)process, CuO is a promising material due to the high oxygen carrier capacity and exothermic reaction in fuel reactor but limited by the low melting point.The combustion rate of carbon is faster than the decoupling rate of oxygen carrier(OC).Hence,high temperature tolerance and rapid oxygen release rate of CuO modified by three different ores were investigated in this study.The kinetics analysis of oxygen decoupling with Cu-based oxygen carriers was also evaluated.Results showed that CuO modified by chrysolite had faster oxygen release rate than that of CuO.Limestone showed obvious positive effect on the oxidization process.The selected OCs could keep stable in at least 20 cycles,for about 1200 min.Shrinking core model(SCM)fitted well for the decoupling process in the temperature range of 1123-1223 K.Reduction rate kinetic information may aid in the development of chemical looping with oxygen uncoupling (CLOU) technologies during reactor design and process modeling.Ternary doped copper oxide with chrysolite and limestone could improve the reactivity of CuO in decoupling and coupling process and also improve the high temperature tolerance.

1.Introduction

Chemical looping combustion(CLC)combining fuel combustion and pure CO2production without extra energy supplemented has attracted many people’s attention.This technology uses oxygen carrier (OC) particles to transfer oxygen from the air to the fuel.Fuel is oxidized by the OC particles to produce CO2and H2O in a fuel reactor.High concentration of CO2stream could be obtained after condensation of the exhaust gas.This process is flameless and no direct contact between fuel and air.

The CLC process has been demonstrated for gaseous fuel combustion, such as natural gas and syngas, using oxygen materials based on Ni [1], Cu [2,3]or Fe [4-7].Recently, the CLC process using solid fuels [8-12]has attracted great attention due to abundant reserves and lower price than natural gas.But the low reaction rate of solid-solid reaction between char and OCs limited the application of solid fuel CLC.Chemical looping with oxygen uncoupling (CLOU) process, proposed by Mattisson et al.[13], is used to solid fuel CLC process which oxygen carrier releases gaseous oxygen at high temperature to oxidize coal char, converting solid-solid reaction into gas-solid reaction and obtaining high conversion rate [14].Thus, compared to OCs in CLC, the OCs used in CLOU should release oxygen in fuel reactor and then be regenerated in the air reactor.CuO,Mn2O3,and Co3O4are the three classic metal oxides that can meet those conditions [13].

Cu-based oxygen carrier is the best choice with high oxygen carrier capacity and exothermic reaction in fuel reactor [9].But agglomeration is a potential issue due to the low melting of Cu(1358 K).In order to improve the mechanical stability and reactivity of CuO, low cost OC materials, especially natural materials are more favorable for the combustion of solid fuel.There are two kinds of supports often used to modify CuO.The one has high mechanical stability, such as Al2O3[15,16], Fe2O3[17], SiO2[18].The second supports are alkali metals [19,20]and alkaline earth metals[21]which could improve coal char gasification rate(In the actually CLOU process, CO2and H2O are often used as gasification process.This means that the gasification also occurred).For the first kind of support, Fe2O3should be more popular because it not only has good mechanical stability, but also provides oxygen which Fe2O3is an oxygen carrier.Moreover, CuO added in iron ore could improve the fuel conversion of iron ore,and the presence of iron species significantly enhanced the high temperature stability of CuO to avoid sintering/agglomeration[22].For the modification of alkali metals and alkaline earth metals,results showed that the presence of Ca results in a more abundant pore structure of char [23]which is beneficial to the gasification of char.Naturally magnesium mineral has been extensively used in biomass gasification as a support or as a catalyst because of its low cost, high mechanical strength and high reactivity toward tar destruction[24,25].Moreover, natural mineral contains Al2O3, SiO2which could also improve the mechanical stability of OCs at high temperature.In our previous study [26], the addition of chrysolite (Ch),iron ore(Fe)and limestone(L)showed good performance with biomass char.Hence, chrysolite, iron ore and limestone were used as supports in this study to improve the reactivity of CuO at high temperature.

Besides,the oxygen release rate had great influence on the char combustion.Rapid oxygen release rate could increase the contact time between fuel and gaseous oxygen thus the combustion will be more sufficient.Eyring et al.[27]found that CLOU process provides rapid combustion of the char with char burnout time lower than the decomposition time of OC.Rapid oxygen release rate is essential for coal CLOU process.In our previous study [26], iron ore and limestone could react with CuO to produce a new substance while chrysolite did not react with CuO.This means that Cu-Fe, Cu-L and Cu-Ch could show some differences in the oxygen release process.Hence, it is necessary to investigate the effect of ores on oxygen release process and obtain proper OCs with stable high reactivity at high temperature.

Furthermore, it is critical for proper design of the fuel reactor that decoupling rates are well understood [28].Chemical kinetics parameters of reduction process of OCs decided the inventory in the reactors and the circulation rate between air and FRs [29].Reduction rate data and kinetic information are of great help to the design and operation of CLOU reactors [29].Table 1 provides a brief review of research activities on CuO-based supported OCs for the CLOU process.It is obvious that there are little literatures devoted on the oxygen release process of CuO modified by iron,limestone and chrysolite.Hence, in this study, iron ore, chrysolite and limestone were used as supports to improve the high temperature tolerance of CuO.Proper proportions of CuO to ores were evaluated to find a suitable oxygen carrier for the CLOU process with high mechanical strength and temperature tolerance in order to reduce attrition rates.Moreover,the apparent activation energy(Ea), and the reaction model function f(x) were also estimated.

2.Experimental

2.1.Oxygen-carrier material

The CuO-based oxygen carriers were prepared by mechanical mixing method.CuO, natural iron ore, chrysolite and limestone were used as raw materials with particle size smaller than 50 μm.The mass ratios of CuO: ore in the oxygen carriers were 40:60,60:40 and 80:20,respectively.The detailed preparation process was presented in other literatures [26,38].The composition analysis of these ores are listed in Table 2.

Table 1 Summary of utilization of CuO-based OCs for the CLOU process [29]

Table 2 Elemental composition (%, mass) of the ore samples

Table 3 Reaction mechanisms and the solid-state reaction rate equations [16,22,43]

Here, Cu denotes CuO and the digit followed Cu represents the proportion of CuO in the sample.The iron ore,chrysolite and limestone are depicted by Fe, Ch and L, respectively.For example,Cu60Ch indicates a sample composed of 60% CuO and 40% (mass)chrysolite.All the OCs were calcined at 1223 K for 8 h.The calcined samples were ground and sieved to the size range of 75-100 μm.

2.2.Experimental apparatus and procedure

2.2.1.Thermo-gravimetrical analysis (TGA) experiments

A TGA experiments were conducted in a TGA system(WCT-2C)produced by Beijing Optical Instrument Factory.A detailed description could be found elsewhere [38-40].For each test, approximately 35-40 mg sample was placed in a crucible.During oxygen carriers decoupling and coupling process, the reactant in the crucible was heated in an air atmosphere at a rate of 20 K·min-1from ambient to the certain value to avoid the reduction reaction occurring during heating process.Then the gas switched to N2.After thorough reduction, air was used to oxidize the sample.The flow rate was set to be 60 ml·min-1.Different reaction temperatures were conducted to analyze the decoupling process from 973 to 1223 K with a temperature difference of 50 K.After determining the proper reaction temperature, 20 cycles of reduction and oxidation processes were carried out to investigate the high temperature reactivity and stability.The reaction time for reduction and oxidation process were set to be 30 min.The purity of the inlet N2gas was higher than 99.9%.

2.2.2.Characterization of the oxygen carriers

The surface morphology of both fresh and used oxygen carrier particles were analyzed by a scanning electron microscopy (SEM,FESEM, JSM7800F, JEOL, Japan).The X-ray diffraction (XRD) patterns for crystal structure of fresh and used particles were conducted on a diffractometer (PANalytical, the Netherlands).The BET surface area and pore volume were characterized by surface area and porosity analyzer (Quantachrome Austosob-iQ).

2.3.Data evaluation

2.3.1.OC decoupling and coupling process

Oxygen carrier conversion rate was calculated as

where m is the instantaneous mass of the sample,moxis the mass of the sample fully oxidized, mredis the mass of the sample fully reduced.

Additionally, the average conversion rate of OC achieving 90%conversion during the oxygen uncoupling was often used to investigate the cyclic redox reactivity of the OC sample [15,41].

where t0is the initial time of the conversion and t0.9is the time when the OC achievers 90%conversion.rave,redand rave,oxirepresent average rate in decoupling and coupling process, respectively.

2.3.2.Kinetic models of decoupling process

The reaction model can give insight into the reaction mechanism of the oxygen release process.In this study, the data tested were analyzed by the gas-solid reaction models that best fitted the experimental data.Around 10 mg oxygen carrier was used in every experiment to study the kinetic parameters of decoupling process.The reaction rate of oxygen release process is described by the equation:

where X is the conversion of oxygen carrier,t is the reaction time,f(X) is the mechanism function, k(T) is the temperature dependent reaction rate constant, following the Arrhenius relationship, which can be generally described by:

where A is the pre-exponential factor, Eais the activation energy, T is the absolute temperature and R is the universal gas constant.Eq.(4) could also be written by integrating the reaction rate

Hancock and Sharp proposed a convenient method to decide the most suitable model for the experimental data [42]:

where n is the slope and lnβ is the ordinate at the origin of the curve of ln[-ln(1-X)]vs.lnt.The value of n can be used to determine the mechanism model.For a particular gas-solid reaction or decomposition reaction, models can be used to describe the process, presented in Table 3.After determining the appropriate model, the activation energy (Ea) and pre-exponential factor (A) of oxygen release process can be calculated through Eq.(5).

Table 4 BET surface area, pore volume and XRD analysis of fresh OCs

3.Results and Discussion

3.1.Oxygen release of CuO in TGA.

3.1.1.Effect of the temperature

Based on the thermodynamic results in Fig.1, the equilibrium oxygen concentration was about 0 when the temperature was below 1073 K.This indicated that it was difficult for CuO to release oxygen at lower temperature.The conversion rate X of CuO during oxygen release process is shown in Fig.2.The conversion rate in Fig.2 also confirmed that there was nearly no oxygen release at 1023 K.With temperature rising,gaseous oxygen could be released and the reaction rate also increased.It took less time to complete the oxygen release process at higher temperature.A short response time at higher temperature meant a rapid decoupling rate which was preferred in a CLOU.The phenomenon was due to the fact that the increased equilibrium oxygen partial pressure(PO2,eq)at higher temperature led to an increase of driving force(PO2,eq-PO2)which promoted the decoupling process (In N2atmosphere, the oxygen concentrations at surface pO2is equal to zero).Higher temperature promoted O2formation in the surface which was beneficial for O2desorption, leading to a larger kinetic driving force.Zhang et al.[44]found that the O2formation and desorption in the surface were the rate limitation steps for oxygen release.Hence,it was better to perform an oxygen decoupling process at higher temperature.The oxygen release rate increased at the initial stage and then decreased as time went on.This was contributed to the fact that the content of O in the OC decreased with time and affected the generation and desorption of O2, leading to a decreased reaction rate.

Fig.1.The value of ΔG and equilibrium oxygen concentration of (R1) at different temperature.

Fig.2.The reduction conversion of CuO under different temperature.

3.1.2.High temperature stability

As shown in Fig.2, CuO showed the fastest oxygen release rate at 950 °C.Hence, 20 continuous cycles were conducted in TGA at 950°C to investigate the stability of CuO.Fig.3 shows the average conversion rate of CuO in N2/air atmosphere and the time consumed.It is obvious that the reaction rate of both oxygen decoupling and coupling process were decreased with time, especially for oxidization process.

As shown in Fig.3, the oxidization rate was higher than that of CuO decomposition.For the first cycle, rave,oxiof CuO was around 1.5 times larger than rave,red.The reason could be that the driving force in the decoupling process was around 0.015-0.1 between 1173-1273 K while the value was around 0.1-0.195 in the oxidization process.But the reactivity of oxidization dropped significantly.rave,oxiin the 20th cycle was reduced by around 80%.For the reduction process, the reaction rate was stable in the first 8 cycles.Decoupling rate decreased from the 9th cycle and showed a relatively gentle downward trend.The inactivation of CuO may be due to the agglomeration and sintering of CuO at high temperature.Decrease in porosity could lead to a decrease in diffusion capacity.But the negative effect of high temperature on the oxidation process was much greater than that on the reduction process.

Fig.3.20 cycles of the reduction-oxidation test in TGA at 1223 K.

3.2.Effect of the support

3.2.1.Effect of doping ratio on the oxygen release performance

In order to maintain the stability of copper oxide at high temperature, natural ores were used as support to improve the anti-agglomeration of CuO.The decoupling processes of OCs are exhibited in Fig.4.It is obvious that Cu-Ch showed faster releasing rate than CuO (The slopes of Cu-Ch(dX/dt) were higher than that of CuO) and the increased proportion of chrysolite reduced the reaction rate.But it was worth noting that CuO did not react with chrysolite, shown in Fig.5.The active component in OCs was still CuO.Chrysolite was only used as a support in the OCs.The increased oxygen release rate could be due to the larger surface area, shown in Table 4, which would be discussed later.For the oxygen release of Cu-Fe, Cu80Fe showed the best performance,while Cu40Fe could not finished the decoupling process within 30 min.An explanation to this could be that, the main component in Cu40Fe was CuFe2O4, which was spinel structure with chemical stability and firm structure.

Fig.4.Effect of proportion of chrysolite, limestone and iron ore during oxygen releasing process at 1223 K in N2 atmosphere.

Fig.5.XRD patterns of fresh (F) and used (U) Cu60ore.

Thereby, the good thermal stability of spinel structure [45]led to a slow decoupling rate.When the amount of CuO increased,the main components in the mixture changed from CuFe2O4to CuO and CuFe2O4.Hence, Cu80Fe and Cu60Fe completed the oxygen release process within 30 min, but both slower than that of CuO.Zhang et al.[44]found that the energy barriers of migration of oxygen atoms in the spinel crystal was higher than that of CuO.The stable crystal structure of spinel and the dense arrangement of internal atoms led to the migration of O protons become difficult.That explained why Cu40Fe has the slowest reaction rate of all OCs.For limestone support, there was no motivation on release oxygen process.The decoupling rates of Cu-L were lower than CuO.The reason could be that CaO reacted with CuO,producing the new substances CaCu2O3and Ca2CuO3, which affected the oxygen release process.This also indicated that CaCu2O3and Ca2CuO3had lower oxygen release rate than CuO.The large molecular structure led to the decrease migration rate of lattice oxygen in the crystal,which affected the oxygen release process.The optimum loading of limestone was 60%which showed the fastest oxygen release rate among Cu-L.This was perhaps due to the lower surface area and pore volume of Cu80L, shown in Table 4.Moreover, agglomeration and sintering of the Cu80L was found after 20 cycles which concretion occurred between the sample and the crucible.Higher loading rate of CuO could permit higher oxygen carrying capacity while may easily suffer from the agglomeration of the CuO clusters.Supports reacted with CuO producing macromolecular substances will decrease the decoupling rate and reduce the reactivity of OCs in CLOU process.Proper proportion of support are necessary to improve the reactivity and stability of CuO at high temperature.

The experiments were conducted in N2atmosphere until it finished the oxygen releasing process.This may be slower than the actual rate in a real chemical-looping with oxygen uncoupling(CLOU) process [46].Because the released oxygen would be consumed by the fuel.Thereby, the actual rate of oxygen release of CuO in the atmosphere surrounding with the fuels could be faster than the experimental results.

Fig.6.Average conversion rate of CuO-ore during 20 cycles of the reductionoxidation test in TGA at 1223 K: Cu-Ch (a), Cu-Fe (b), and Cu-L (c).

3.2.2.Effect of the number of cycles

20 cycles were conducted in TGA to investigate the stability of Cu-based oxygen carriers.rave,redand rave,oxiof CuO-ore under N2/air atmosphere at 950 °C are shown in Fig.6.According to Fig.6, it is obvious that rave,oxiwas larger than rave,red.The average reaction rate was greatly affected by the explosion time of OCs at 950 °C, especially for Cu80ore (only the first ten cycles of Cu80L is shown in the Fig.6 due to the severe agglomeration and sintering phenomenon).For oxygen release process, CuO modified by chrysolite showed positive effect on this stage.Cu80Ch had the highest average oxygen released rate.But rave,oxidecreased a lot and was even lower than Cu60Ch after the 16th cycle.rave,oxialso showed an obvious decline trend with cycles.The average oxygen release rate of Cu60Ch and Cu40Ch kept stable in the 20 cycles.The reactivity was not affected by the exposure time, indicating the better anti-agglomeration and sintering performance of Cu60Ch and Cu40Ch.The increased reaction rate, shown in Fig.4 and the stable reactivity in 20 cycles of Cu60Ch and Cu40Ch suggested that appropriate proportion of chrysolite in CuO not only promoted the stability of OC at high temperature, but also improved the decoupling process.However, the average oxidization rate of Cu60Ch showed a decreased trend, showing in Fig.6, while Cu40Ch kept stable in 20 cycles.Considering the reaction rate of fuel combustion, higher oxygen release rate and oxygen carrying capacity made Cu60Ch more competitive in CLC process for stable operation in a long time.In fact, the temperature of air reactor was lower than the fuel temperature which agglomeration and sintering of OCs would be weakened or not occurred.

CuO modified by iron ore showed more stable performance in 20 cycles.Although the average reaction rate of Cu-Fe was slower than that of CuO, the stable reactivity during 20 cycles also made iron ore a suitable support for CuO.Notably, rave,redand rave,oxiof Cu-Fe increased after the first cycle.That could be due to the activation of iron ore.Chen et al.[47]conducted about 5 redox cycles in a fixed reactor to activate the iron ore before experiments.The addition of iron ore also improved the oxidization process of CuO a lot.The production of spinel structure (CuFe2O4) reduced the reactivity of OC in the oxygen release process but helped CuO keep stability at high temperature.However, although the average oxidation rate of Cu-Fe remained stable in 20 cycles,rave,redof Cu80Fe decreased from the 17th cycle and continued the downward tendency.Cu40Fe showed a stable decoupling and coupling process.The lower content of CuO in Cu40Fe could lead to higher OC/fuel ratio.Hence, Cu60Fe is the most competitive OC in Cu-Fe OCs.

Loading content of CuO on a support material could affect the chemical stability of the Cu-based oxygen carriers [48].Cu80L could not complete 20 cycles due to the severe agglomeration.But it was remarkable that the modification of limestone showed obvious positive effect on the regeneration process.Cu40L and Cu60L had the highest average oxidization rate among all OCs.The promotion of limestone helped CuO keep stable and improved the oxidization reactivity at 1223 K.For the decoupling process,Cu40L and Cu60L also maintained stable oxygen release rate in 20 cycles.But Cu60L had higher oxygen release rate than that of Cu40L.This could be the reason that the main component of Cu60L was still CuO while the proportion of macromolecular substances Ca2CuO3and CaCu2O3increased in Cu40L, shown in Supplementary Material Fig.S1, which affected the oxygen release rate.Considering the active content in OCs, Cu60L could be more appropriate in CLOU due to its stability and higher oxygen carrying capacity.

Except Cu80L, all samples could help CuO keep stable which demonstrated that three types of ores could enhance the high temperature resistance of CuO.At the same doping ratio, Cu-Ch showed the highest oxygen release rate, followed by Cu-L and Cu-Fe.Cu60Fe and Cu60L also had positive effect on the stability of CuO at high temperature.But the coupling rate showed the opposite result.Cu-L had the highest oxidization rate, followed by Cu-Fe and Cu-Ch.The modification of chrysolite promoted the oxygen releaser rate and limestone on CuO showed obvious positive effect on the regeneration process.In fact, Cu80L and Cu40Fe could also complete the decoupling and coupling process if the reaction time is long enough.The modification of ores could not influence the oxygen carrying capacity of OCs.

3.2.3.Kinetics analysis of oxygen uncoupling process

Base on the analysis in Section 3.2.2, Cu60ore were selected as proper candidates in CLOU.Models for the reduction of Cu-based OCs are calculated and determined according to gas-solid reaction mechanisms listed in Table 3 (Cu60Fe after 5 cycles was used to investigate the kinetics analysis at 1123-1223 K).The value n of Cu60Ch, Cu60Fe and Cu60L at 1123-1223 K are between 1 and 2.Suitable kinetic model could be reaction-order model, shrinking core model (2D,3D) and nucleation model (2D).The results calculated by the four models are compared with data from TGA and found that the shrinking core model (2D) is more appropriate to predict the experimental data for the uncoupling process.The rate constants and the regression coefficients of each metal oxide at different reduction temperatures are summarized in Table 5.

Table 5 The parameters for model fitting in the uncoupling process

The curves in Fig.7 represent the simulated data using shrinking core model (2D) of different OCs during decoupling process.The simulated data and experimental data fits well for Cu60ore.An activation energy (Ea) and pre-exponential factor (A) for the reduction are also calculated in this study from the linear fit to temperature-dependent Arrhenius plots, shown in Fig.8.In order to illustrate better, the oxygen release experiment was carried out at five different temperatures of 1153, 1176, 1203, 1228 and 1253 K, respectively.The activation energies (Ea) and preexponential factors for Cu60Ch were 196.02 kJ·mol-1and 6.67 × 107min-1, for the Cu60Fe were 269.41 kJ·mol-1and 1.20 × 1011min-1, and for Cu60L were 232.93 kJ·mol-1and 2.29 × 109min-1, respectively.

It should be noted that Cu60Ch had the fastest oxygen release rate and lowest activation energy among Cu60ore.Chrysolite couldnot react with CuO while iron ore and limestone consumed CuO to produce new substances,especially for iron ore which CuFe2O4has stable spinel structure.Cu60Fe also showed the highest activation energy among Cu60ore.Even though CuFe2O4, CuCa2O3and Ca2-CuO3could release oxygen, the large molecular structure affected the oxygen transfer in the lattice, thus resulting higher activation energy.It is suggested that the support should better not react with CuO, but only as a carrier to improve the stability of CuO.

Fig.7.Comparison of simulated conversion curves with the experimental data:(a)Cu60Ch,(b)Cu60Fe,and(c)Cu60L,solid lines represents experimental data;dashed lines represents simulated data.

Besides,the main purpose of the kinetic study is to calculate the solids inventory in the system and circulation rate of OCs which these parameters are related with the reactivity and oxygen transport capacity.The circulation rate of solid per MWth, is calculated as [32]:

where mOis the mass of oxygen required per kg of solid fuel to full combustion, LHV is the lower heating value of the solid fuel, ROCis the oxygen transport capacity of OCs.

According to the study [32], the maximum circulation rate is calculated as the minimum value of ΔXOCis 0.1.The minimum mass of solids in the fuel and air reactors per MWth of fuel is calculated as:

Assuming the particles in the reaction are well mixed,the average reaction rate was calculated as [32]:

where tcis the time required to complete reaction.

The expression that describes the oxygen release process is determined by Eq.(4).The pre-exponential factor A, activation energy Eaand the function f(X)are all obtained in our study.Hence,the solids inventory and solids circulation rate could be calculated from Eqs.(9) and (10), which is useful to guide the operating conditions and reactor design.But detailed analysis and calculation should be further studied, where the types of coal (LHV), size parameters of reactor, heat and material balance of reactor need to be considered.

Fig.8.Arrhenius plot for the oxygen uncoupling reaction of Cu60ore.

Fig.9.SEM images of all fresh (F) and used (U) samples after 20 cycles at 950 °C.

3.3.Characterization of particles

XRD and BET analysis are shown in Table 4.The main components of Cu-Ch were CuO and Mg2SiO4.CuO did not react with chrysolite while the new phase in Cu-Fe and Cu-L were found.Macromolecular structure with a larger specific molar volume blocked the transfer of oxygen in lattice and leaving the matrix[46], leading to lower oxygen release rate.The main active phase of Cu40Fe was CuFe2O4(Detailed XRD patterns are shown in Supplementary Materials Fig.S1).This explained why Cu40Fe had the lowest oxygen release rate.For limestone modification, the increased amount of limestone in CuO also caused an increase XRD diffraction intensity of Ca2CuO3, leading to lower oxygen release rate.Moreover, the surface area and pore volume reduced with the increased proportion of iron ore and limestone which were also responsible for the reduced oxygen release rate.Cu80L with the lower oxygen release rate had lower surface area and pore volume.For chrysolite modification, Cu-Ch had high BET surface area and pore volume.CuO did not react with chrysolite.Small granules and grains on the surface of Cu-Ch, shown in Fig.9,increased the surface area of OCs which also explained why Cu-Ch showed the high oxygen release rate.XRD analysis of Cu60ore after 20 cycles were conducted to investigate the phase variation after 20 cycles, shown in Fig.5.Results showed that no obvious phase variation occurred before and after 20 cycles.Stable phase of Cu60ore could be obtained which also suggested that chrysolite,iron ore and limestone could be proper supports to improve the mechanical stability at high temperature.

SEM images of fresh and used oxygen carriers of Cu-ore are presented in Fig.9.As shown in Fig.9(a)-(c),small granules and grains on particle surface could be observed in fresh Cu-Ch.The surface morphology of the particles kept stable after 20 cycles.No agglomeration and sintering phenomenon was observed on the surface.The surface topography of Cu-Fe also showed little difference between fresh and used oxygen carriers, indicating the stable structure of Cu-Fe in oxygen releasing process.Cu40L and Cu60L showed stable structure after 20 cycles.The presence of CaCu2O3in Cu60L and Cu40L may have better high-temperature resistance than CuO, leading to an enhancement of stability of OC samples.But severe sintering and agglomeration occurred on the surface of Cu80L.Formation of agglomerates and sintering could affect the fluid-dynamic properties of OC, possibly leading to defluidization phenomenon in a CLC and contributing to the failure in operation.The modification of CuO with appropriate amount of ore helped CuO keep stability at high temperature condition.

4.Conclusions

The decoupling process of CuO modified by three different ores and the activation energy were investigated in this study.The optimum amount of ores addition into copper were obtained in the TGA experiments.Results showed that higher loading content of CuO could permit higher oxygen carrying capacity while may easily suffer from the agglomeration of the CuO clusters.Cu80ore showed high reaction rate but the reactivity reduced a lot after 20 cycles, especially for Cu80L.The production of CuFe2O4,Ca2CuO3,CaCu2O3reduced oxygen release rate but had better high temperature tolerance than CuO.Supports reacted with CuO producing macromolecular substances could decrease the decoupling rate and reduce the reactivity of OCs in CLOU process.Hence,support should better not react with CuO, but only as a carrier to improve the stability of CuO.Cu60ore showed the best performance in CLOU process.After undergoing 20 redox cycles at atmospheric pressure, Cu60ore maintained high reaction rate and stability.The modification of chrysolite improved the oxygen release process and limestone promoted the oxidization rate.XRD patterns and SEM images of fresh and used Cu60ore also validated the improvement of chrysolite, iron ore and limestone on the reactivity of CuO at high temperature.Shrinking core model(SCM) fitted well for the decoupling process in the temperature range of 1123-1223 K.The values of activation energy (Ea) for Cu60Ch, Cu60Fe and Cu60L were 196.02, 269.41, 232.93 kJ·mol-1,respectively.

Overall,this paper studied the improvement of tolerance of CuO supported by three ores at 1223 K and the reactivity of oxygen release process.These preliminary results suggest that CuObased support by ores are potential to be used as an oxygen carrier.Chrysolite could be used as a support to enhance the oxygen release rate while limestone could be employed to improve the oxidization rate.Further study should be conducted to study the ternary doped copper oxide with chrysolite and limestone.

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 financial support by the Fundamental Research Funds for the Central Universities(xjh012019019) and the National Natural Science Foundation of China (51606087).

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.radonc.2018.09.004.