Preparation and Characterization of Lanthanum-Promoted Copper-based Oxygen Carriers for Chemical Looping Combustion Process

Chemical looping combustion (CLC) is an oxygen combustion process that results in the higher CO2 concentrations in flue gases. Copper-based oxygen carriers have high reactivity, resulting in high purity CO2 generation, but face the challenges of low copper melting point and potential thermal sintering under high-temperature cyclic operation. A temperature-programmed reduction and oxidation (TPR-TPO) technique (Redox) was used to simulate of the cyclic operation of copper-based oxygen carriers in the CLC process. Multiple material characterization methods were applied to the fresh and used oxygen carriers, including the determination of pore structure and surface area by Brunauer-EmmettTelle (BET), crystal structure by X-ray diffraction (XRD) and surface phase diagram analysis by scanning electron microscopy and energy-dispersive X-ray spectroscopy (SEM-EDX). Thermally stable copper-based oxygen carriers have been successfully prepared using commercially-available Al2O3 as the supporting material and selected additives. A substantial resistance to agglomeration and durability in over 50 cyclic Redox tests under high temperature conditions (oxidation at 800°C and reduction at 800°C) has been confirmed for the Sample labeled as CLCA. The oxygen transfer capability was almost stable, and the CuO content ranged from 20 wt% to 23 wt% for the CLCA sample. The supporting material (Al2O3) significantly impacted the oxygen releasing performance and thermal stability of the copper-based oxygen carriers, due to the formation of a new crystal phase (CuAl2O4) which cross-links the CuO and the γ-Al2O3. Lanthanum accelerated the cross-linking between CuO and Al2O3.


INTRODUCTION
In conventional combustion systems, fuel is directly mixed with air for burning.This results in low partial pressure CO 2 in its flue gases because of dilution by nitrogen from the air.The low concentration of CO 2 will create a significant energy penalty (Herzog et al., 2001) when it needs to be separated.For a coal-fired power plant, roughly one-fifth of the electricity produced would be lost to CO 2 separation and compression.Its incremental cost would be 80% over combustion without carbon capture based on DOE estimates, if 90% carbon capture is required (DOE/NETL, 2010).High costs of carbon capturing have been major hurdles for the promotion of carbon capture technologies into commercial application.Oxy-firing combustion, using a mixture of oxygen and recycled CO 2 to replace air, represents a new direction allowing for conventional boilers to be employed with modifications without the penalties by post-combustion CO 2 capture (Buhre et al., 2005).However, oxygen production is expensive and energy intensive.Oxy-firing further faces challenges to improve boiler efficiency and to enhance heat transfer.Among all available or proposed technologies (Lyngflet et al., 1998) involving CO 2 capture during combustion process, chemical looping combustion (CLC) is a new class of processes involving indirect combustion of fuels, without air-firing or oxygen-firing, and simultaneously resulting in high concentration of CO 2 for its sequestration.Relative to a pulverized bituminous coal fired power plant without CO 2 capture, the increase in the electricity generation cost for a CLC plant was about 12-22%.The incremental electricity cost was lower than other CO 2 capture technologies.The estimated cost of the capture per tonne of CO 2 avoided was estimated to be 6-13 Euro for CLC.Similar evaluations concluded that the cost was 18-37 Euro for a pre-combustion technology using IGCC, and 13-30 Euro for an oxy-fuel process (Petrakopoulou et al., 2010;Adanez et al., 2012).Chemical looping combustion significantly improves the energy conversion efficiency, in terms of electricity generation (Ekstrom et al., 2008), because it improves the reversibility of the fuel combustion process through two linked parallel processes (Anheden and Svedberg, 1998;McGlashan, 2008;Sempuga et al., 2010Sempuga et al., , 2011)).
In chemical looping combustion process, an oxygen carrier consisting of metal (Me) or the reduced metal oxide (Me x O y-1 ) associated with its oxidized form (Me x O y ), is circulated between two inter-connected reactors, i.e.,air reactor and fuel reactor.In the fuel reactor, the metal oxide (Me x O y ) reacts with the fuels to produce CO 2 , H 2 O and metal (Me) or the reduced metal oxide (Me x O y-1 ).Pure CO 2 is ready for subsequent sequestration after H 2 O is condensed.In the air reactor, the metal (Me) or the reduced metal oxide (Me x O y-1 ) reacts with air to form metal oxide (Me x O y ).The advantages of CLC compared to normal combustion processes include: no energy is needed for separation, because CO 2 is not mixed and diluted with nitrogen.And the control of other gaseous pollutants in this process becomes more efficient, because of their higher concentrations in a lower mass flow of nitrogen-free flue gas.Also, the high thermal conversion efficiency and flame-free combustion leading to the near elimination of thermal NO x emission (Cao et al., 2004;Cao and Pan, 2005).
The critical role of durable oxygen carriers has been well addressed in the development of this novel chemical looping combustion process, especially its performance under the harsh combustion environment (high temperatures and cyclic operation).Currently, a number of metals, such as iron (Fe), nickel (Ni), copper (Cu), Cobalt (Co), manganese (Mn), Chromium (Cr), Tungsten (W), Strontium (Sr), and Barium (Ba) have been the major ingredients of prepared oxygen carriers (Ishida and Jin, 1996;Jin et al., 1998Jin et al., , 1999;;Lyngfelt et al., 2001;Mattisson et al., 2001;Ishida et al., 2002;Cao et al., 2004;Cho et al., 2004;de Diego et al., 2004;Johansson et al., 2004;Cho et al., 2005;Cao and Pan, 2006;Corbella et al., 2006;de Diego et al., 2007;Tian et al., 2008;Li et al., 2009;Mattisson et al., 2009;Siriwardane et al., 2009;Dennis and Scott, 2010;Forero et al., 2010;Li et al., 2011;Galinsky et al., 2013).One exception to using metal-based oxygen carriers is the cyclic operation between calcium sulfate (CaSO 4 ) and calcium sulfide (CaS) by Alstom Power, Inc and others (Song et al., 2004;Thibeault et al., 2005;Andrus et al., 2009).The kinetics of reactions varies widely, depending upon the types of metal oxides, particle sizes (70 µm to 2 mm), properties of fuels (H 2 , CO and natural gas, oils and coals) and temperatures (800-1200°C).If impregnated on some inert support materials (Al 2 O 3 , SiO 2 , ZrO 2 , TiO 2 , bentonite, sepiolite and yttrium-stabilized zirconia (YSZ)), attrition-resistance and durability of oxygen carriers could be further enhanced.Almost full conversion of fuels could be achieved in a few minutes depending on the metals impregnated in the oxygen carriers.Other preparation methods included sol-gel method (Tian et al., 2008;Siriwardane et al., 2009) and freeze granulation (Johansson et al., 2004(Johansson et al., , 2006)).Compared with Fe under similar operational temperatures, Ni, Co, Cu and their oxides generally showed higher activities (Siriwardane et al., 2009).Iron-based carriers are abundant and lower cost than other metals.The use of coal as fuel for the CLC process aroused much attention recently.Generally, the coal is firstly gasified in situ using CO 2 in the presence of oxygen carriers.The choice of certain oxides to effect exothermic reactions in the fuel reactor is preferable to facilitate an elevated temperature for the promotion of higher reaction kinetics of endothermic gasification of solid fuels.Copper is a such oxide (Cao and Pan, 2006).Generally, a circulating fluidized bed, which is preferable in the chemical looping process, has huge thermal inertia.It is possible to reach high temperatures in the fuel reactor by a higher solids circulation.On the other hand, if selected oxygen carriers could directly release free oxygen atom, its reaction kinetics is expected to be much higher than indirect transfer of oxygen.Copper oxide (CuO) has properties of direct oxygen releasing (Mattisson et al., 2009).However, the key question for the copperbased oxygen carriers is whether or not it can withstand gasification temperatures of above 800°C, particularly from the point of view of avoiding sintering and deactivation of copper oxide in its reduced form.This is the main reason adduced in the literature to reject the Cu-based oxygen carriers for the CLC process.
Dennis and Scott (2010) demonstrated that gasification rates of Hambach lignite and its char are significant at 900°C when copper-based oxygen carrier is used, because the CuO could produce gas-phase O 2 , which promoted oxidation of syngas and char.During the course of the experiments, the BET surface area of the support fell from 60 m 2 /g, just after preparation, to around 6 m 2 /g after 20 cycles.However, this drop did not appear to affect the overall capacity of the oxygen carrier to react with fuels, and the reaction kinetics with CO.The drop of BET surface area of Cu-based carriers likely caused sintering of Cu under elevated temperatures; however the sintering process did not result in reaction kinetics drop because the overall performance of the looping scheme is dominated by the much slower kinetics of the gasification reaction of solid fuels.de Diego (2005), Gayán (2010), andForero (2010) and Corbella (2010) systematically studied the impacts of supporting materials on the performance of copper-based oxygen carriers, such as mechanical mixing, co-precipitation, impregnation using alumina as support.The contents of CuO of these carriers were between 10 and 26 wt%.The particles were calcined at different temperatures in the range of 550-950°C and further tested over 100 redox cycles in a fluidized bed facility.It was observed that the CuO content in the oxygen carrier, the calcinations temperature used in the preparation, and the conversion reached by the oxygen carrier during the reduction period affected the agglomeration process.They manipulated preparation conditions to produce Cu-based oxygen carriers with high reactivity and small attrition rates, so that the agglomeration of these oxygen carriers in the fluidized reactor was avoided.
Based on previous studies on reaction kinetics, oxygen transfer capability, energy balance between two interconnected reactors and CO 2 purity (Cao and Pan, 2006), copper (Cu) was selected as the metal in our oxygen carrier study.However, the low melting point, the accumulative thermal sintering under high-temperature cyclic operation and redistribution on the supporting materials are major challenges to preparing a thermally stable copper-based oxygen carrier (Adanez et al., 2006;Corbella et al., 2006;de Diego et al., 2007).
In this study, the thermal stability of copper-based oxygen carriers promoted by chemical additives was investigated.A temperature-programmed reduction and oxidation (TPR-TPO) technique (Redox) was used to simulate of the cyclic operation of chemical looping combustion process using Cu-based oxygen carriers prepared in our laboratories.Multiple material characterization methods were applied to both the freshly prepared and used CuO oxygen carriers.This included the determination of pore structure and surface area by Brunauer-Emmett-Telle (BET) method, the crystal structure by X-ray diffraction (XRD) and surface phase diagram analysis by scanning electron microscopy and energy-dispersive X-ray spectroscopy (SEM-EDX).

Preparation of Copper-based Oxygen Carriers
To minimize mass transfer limitations and enhance the strength of our oxygen carrier, a commercially available supporting material (Al 2 O 3 ) was used in the preparation of oxygen carriers.It was purchased from SASOL Germany GmbH.The average size is 250 µm and its purity is high based on SEM-EDX analysis.
It was found that less than 5 wt% of CuO was loaded on the supporting material (Al 2 O 3 ) using the traditional "wet" impregnation method.It was found that it was difficult for copper materials penetrating into the moisture-filled pore structures of the supporting material.After several trials, a "dry" impregnation method was developed.The major difference between the "wet" and "dry" methods was the removal of moisture occupying the interior pore structure of supporting materials prior to doping procedures.The CuO loading was increased to over 20 wt%.The "dry" impregnation method is listed below: 1) the supporting material was dried in an oven at 120°C overnight; 2) droplets of saturated copper nitrate (Cu(NO 3 ) 2 ) were titrated on this dried supporting material, until it was fully wetted; 3) a few droplets of ethanol was applied to the copper nitrate saturated oxygen carriers and then the prepared sample was dried in oven at 120°C again for another 2 hours.4) Procedure 1) through 3) will be repeated two times, 5) Samples were calcined in an oven at 500°C for 5 hours.
The reason to use 500°C in the calcination procedure is to control the reaction rates of the decomposition of impregnated Cu(NO 3 ) 2 and the vaporization of the moisture residues in a mild condition for preventions of strength loss of prepared oxygen carriers.More than 50 variations of chemical formulae of Cubased carrier were prepared by adding different kinds of additives, pursuant to the aforementioned procedures.This study presented three typical samples of copper oxide oxygen carrier (Sample), which differed significantly in their respective physical and chemical properties and were extensively investigated.Sample 1 is labeled as CA (C-Copper, A-Al 2 O 3 ), which was prepared according to the aforementioned standard procedure without using any additives.Sample 2 is labeled as CCA (C-Citric acid, C-Copper, A-Al 2 O 3 ), in which a dispersion reagent (citric acid) was applied simultaneously with copper nitrate.Sample 3 is labeled as CLCA (C-Citric acid, L-Lanthanum, C-Copper, A-Al 2 O 3 ), in which a dispersion-promotion reagent (citric acid) and thermal durability enhancing reagent, (Lanthanum nitrate) were applied simultaneously with copper nitrate.The ratio of Cu(NO 3 ) 2 :La(NO 3 ) 2 :citric-acid was kept constantly at 1:0.05:0.005by weight.

Tests Facilities of Non-isothermal and Isothermal Redox Cycles of Cu-based Oxygen Carriers
The copper loading amount and the durability of oxygen carrier Samples were evaluated under TPR (10%H 2 ) and TPO (air) (Temperature Programmed Reduction and Oxidation) cycles in a Micrometrics ChemiSorb 2720 instrument.The ChemSorb is a microreactor, with microgram amount of samples, and thus there is only one temperature measurement point inside it.25 mL/min of 10% hydrogen in argon was introduced into ChemiSorb system to reduce the prepared oxygen carriers under programmed temperature from room temperatures to 800°C at a temperature ramp rate of 20 °C/min.This was followed by dropping the temperatures to room temperature with argon.Then air was introduced into the instrument for oxidation at similar temperature ramp rate of 20 °C/min.An alternative method was to maintain a constant temperature at 800°C between reduction and oxidation cycles.In each stage of either oxidation or reduction, the duration was 5 mins; and the duration of the argon purge between oxidation and reduction stages was also 5 mins.Both Schemes of redox cycles are shown in Fig. 1.The test procedure of Scheme 1 was applied to samples CA and CCA, and both Scheme 1 and Scheme 2 were applied to sample CLCA.During the redox cycles, oxygen carriers were fully reduced and fully oxidized.Both procedures were repeated until the desired number of Redox cycles of oxygen carriers was reached.The use of ChemiSorb system only gave the evaluation of cyclic performance of oxygen carriers in view of their oxygen transfer capabilities in static conditions.Further studies of oxygen carriers in a CLC fluidized bed are strongly suggested to determine their kinetics and attrition-resistance durability, as was performed in our other works.
The peak intensity, corresponded to changes of thermal conductivity, was caused by the joint effects of the decrease of H 2 and increase of H 2 O in offgas.Thus, the peak areas corresponded to the total oxygen consumption of the tested oxygen carriers.The calibration curve of hydrogen intensity as a function of CuO concentration was established using different amounts of pure CuO in the system and also the calibration curve of hydrogen amount was setup to verify the hydrogen consumption.Thus the loading amount of CuO in the supporting materials could be determined using TPR tests.Because tests performed in ChemiSorb were under the static condition of oxygen carriers (a fixed-bed operational mode), the reported changes of the CuO content between different cycles should be an assessment of the oxygen transfer capability of oxygen carriers.

Instrumentation
BET Technique.The surface area and pore size distribution was analyzed by a Micromeritics ASAP 2020 (accelerated surface area and porosimetry analyzer, Micromeritics Instrument Corp).Specific surface area was based on the BET equation in the relative pressure range of 0.05-0.35P0 (P 0 is the critical pressure of nitrogen).The mesopore size distribution was based on the density functional theory (DFT).The adsorption isotherms were obtained using nitrogen gas as the adsorbate at 77 K.
XRD Technique.The XRD analysis on the prepared samples was made using a Scintag X'Tra (ThermoARL) X-ray diffraction meter equipped with a Peltier cooled Si solid detector.Cu Kα 1 (0.154 nm) was used as the radiation.Diffraction patterns were collected at 45 kV and 40 mA, at 0.01° step and a count time of 0.500 s over a range of 5.00-90.0°(2θ)with a step scan rate of 1.2 °/min.

SEM-EDX technique.
The morphology characterization of the sample was analyzed using a JEOL JSM 5400-LV scanning electron microscope.The sample was crushed and then mounted on aluminum stubs using double-sided carbon tape.It was then observed uncoated in low vacuum mode at a chamber pressure of 110 milli torr using a backscattering electron detector.Elemental analysis for the sample was obtained using a KEVEX Quantum detector with a new MOXTEK AP3.3 thin filmed window and an IXRF energy dispersive spectrometer.

Sample Tests in TPO-TPR
Fig. 2 and Fig. 3 present SEM images of fresh and used oxygen carrier samples.Copper (Cu) has larger atomic number than aluminum and thus is highlighted in contrast to the supporting materials (Al 2 O 3 ) in SEM images of samples.As indicated in the SEM images of three fresh oxygen carrier Samples (Figs. 2(a) and 2(c) for Sample CA and Sample CCA, and Fig. 3(a) for Sample CLCA), copper was evenly distributed on the surface of the supporting materials.EDX analysis also indicated the majority of copper accumulated on the outer layer of the supporting materials (Figs.2(a) and Fig. 3(a)).The inside of supporting material was less occupied by CuO.
After the TPR-TPO cyclic tests using 10% H 2 (in TPR) and air (in TPO), Sample CA and Sample CCA showed a major phase transformations and agglomerations.Fig. 2(b) shows an image of the Sample CA after only 7 cycles of TPR-TPO (Scheme 1).A significant melting and agglomeration of CuO-Cu was identified, compared to the fine crystal of CuO-Cu on the surface of the fresh Sample CA.A hard crest of CuO-Cu had formed on the surface of supporting materials, which demonstrated significant melting or agglomeration of Cu after cyclic Redox tests.Similar to Sample CA in Fig. 2(b), the application of dispersion reagents in CCA samples, which helped to homogeneously distribute CuO-Cu on the surface, did not significantly prevent the melting or agglomeration of CuO-Cu after 10 TPR-TPO cycles (Fig. 2(d Thus, the controlled test temperature of 800°C was likely just a "skin" or "surface" temperature of oxygen carriers, and actual "bulk" temperatures of oxygen carriers might be higher than 800°C and could approach the melting point of Cu, as a result of the exothermic reduction of CuO by H 2 .This had inspired us to further modify temperature measurement system in the current commercial ChemiSorb.Table 1 presents the oxygen transfer capability, in view of the measured CuO content of oxygen carriers using H 2 .For all three copper-based oxygen carriers, the oxygen transfer capabilities were almost stable during the cyclic operation.For CA sample, the CuO contents ranged between 16 wt% to 20 wt%, CCA from 22.9 wt% to 24.3 wt%, and CLCA from 20 wt% to 23 wt%.Investigation into the intensity curves of TPR under Scheme 1, as shown in Fig. 4, indicated there were two stages of reduction starting at lower temperatures.The first reduction peak was around 210°C and the second reduction peak was around 260°C.However, the peak of the intensity curve for pure CuO was identified to be 325°C.Different reduction peaks of CuO Samples in TPR tests indicated different physiochemical occurrence of CuO.Shifts of the peak temperatures of different status of CuO could be further indication of changes of reduction kinetics of CuO by H 2 .The lower peak temperatures of prepared CuO on supporting materials may indicate improved reduction kinetics of CuO by H 2 , i.e., finely-dispersed copper on supporting materials exhibited improved reduction kinetics by H 2 than pure CuO powders.Test using Scheme 1 procedure also showed that reduction reactions of both supported copper-based oxygen carrier Samples and the pure CuO powder could occur under much lower temperatures than 800°C, indicating highly active copper as oxygen carriers when H 2 was used as fuel in this study.The chemical looping system is a complex system, in which some areas out of the fuel reactor had lower temperatures.The kinetic studies presented the potential contributions of reduction of oxygen carrier in areas out of the reaction area of the traditional fuel reactor. The continuous exposure of oxygen carriers under increasing TPR-TPO cycles resulted in the increased area of the second reduction peak and it gradually became the dominant one.It indicated copper oxygen carriers are subject to the phase transformations with increasing TPR-TPO cycles.Shifts of the peak temperatures to higher temperatures indicated reduction kinetics decrease of oxygen carriers by H 2 .Referring to aforementioned SEM pictures (especially for Sample CA and Sample CCA), it seemed that there was more accessible CuO/Cu on the surface of tested oxygen carriers with increasing TPR-TPO cycles, naturally leading to decrease in the mass transfer limitation inside supporting material and increase kinetics.These two conflicting observations in TPR-TPO and SEM on reduction kinetics of oxygen carriers might imply the  formation of new crystal phases of either copper itself or between copper and supporting materials, which adversely affected the reduction kinetics of oxygen carriers with increasing TPR-TPO cycles.Under such low temperature ranges (200-300°C), the limitation of mass transfer inside the pores of supporting materials will be unlikely significant, but this is not true for the mass transfer of H 2 inside newly formed copper block or layer because of tendencies of melting, agglomeration or further new phase from reaction between CuO and Al 2 O 3 .

Sample Characterizations
To identify phase transformation of oxygen carriers during their TPR-TPO cycles, samples was subjected to analysis using XRD, as shown in Fig. 5.There were clear evidences of the formation of an integrated crystal phase between CuO and Al 2 O 3 , i.e., CuAl 2 O 4 spinel.This agreed with all previous studies regarding the use of copper as an ingredient of oxygen carriers (Adanez et al., 2006;Arjmand et al., 2011Arjmand et al., , 2012)).The formation of CuAl 2 O 4 spinel was found for all tested oxygen carriers in this study although to different degrees.This was a new solid phase formed in the cyclic TPR-TPO tests because the CuAl 2 O 4 spinel was not found in the fresh Samples (only CuO and Al 2 O 3 were presented in XRD patterns).However, not all the CuO/Cu was converted to CuAl 2 O 4 spinel, because the fresh crystal phases of CuO, or Cu and Al 2 O 3 could still be identified in the tested Samples.Combining the investigation of TPR curves, it may imply that the shift of peak temperature and the subsequent decrease of the CuO reduction kinetics may have resulted from the formation of CuAl 2 O 4 spinel between copper and supporting material.The gradual shift of peak area between the first peak and the second peak may imply the acceleration of transformation from CuO-Cu to CuAl 2 O 4 with increasing TPR-TPO cycles.It was noticed that the calculated loading amounts of CuO-Cu by the integration of the consumption peak areas of hydrogen in the TPR-TPO curves was almost constant during phase transformation under different cycles for both Scheme 1 and Scheme 2 operations.This may indicate that all loaded CuO-Cu on oxygen carriers was fully accessible.Reversibility of reduction and oxidization cycles of CuO-Cu was achievable even the CuO in porous Al 2 O 3 occurred as CuAl 2 O 4 .The perovskite structures of metal oxides as oxygen carrier candidates have been intensively discussed recently.It has been identified to provide a higher oxygen mobility in the bulk, and good ability to host large concentrations of vacancies in its structure at elevated temperatures (Mihai et al., 2010, Wang et al., 2011;Kallen et al., 2013).In the XRD analysis of this study, we did not actually find the formation of the perovskite structure among lanthanum (La), copper (Cu), aluminum (Al) and oxygen (O), possibly because La was in a lower molecule ratio.
BET pore structure analysis method was further applied to identify the transformation of the micro-structure of CuO-Cu on the surface of supporting material.Variations of pore size distribution of fresh and used oxygen carriers were presented in Fig. 6.There was no major change of the pore size distribution between fresh supporting material (Al 2 O 3 ) and its treated samples under higher temperatures (1150°C) within a range of 2 nm to 10 nm, but an increase of mesopore volume of treated supporting material in a range greater than 15 nm.In the same Figure, variations of pore volume distribution of fresh and used CuO samples of different methods were also presented.It indicated that loading CuO on supporting material without any additives decreased its pore volume (CA Samples).As expected, used CA Sample would lose smaller pores below 15 nm significantly after 7 TPR-TPO cycles.However, the volume of pore size greater than 15 nm was increased, which was similar to that of unloaded supporting material treated under similar temperature.This was also the case for CCA Sample although the dispersion agent was applied.Pore volumes of CCA sample between 2 and 15 nm had similarly decreased, but its pore volume of larger pores size (greater than 15 nm) had increased.Combining both evidence from analysis of BET and SEM, one may conclude that the reduced copper may be less mobile and confined inside smaller pores of supporting materials, resulting in decreases of pore volume of smaller pores because of agglomeration.The increase of pore sizes enhanced the mobility of melting copper within larger pores and partial copper movement toward the outside of supporting materials, thus resulting in the increase of pore volume of larger pores.The use of dispersion agent in Sample CCA seemed somehow controlling this unfavorable tendency.The pore size distribution of the fresh CLCA Sample followed the same trend of Sample CCA.It was noticed that there was still significant loss of smaller pores, from 3 to 12 nm for used Sample CLCA.This may imply CuO-Cu was immobile inside micro-pores of the supporting material due to the promotion of lanthanum in Sample CLCA, but a tendency of phase changes of CuO-Cu under even smaller dimension were still there.Compared to TPR-TPO tests where peaks of reduction in Scheme 1 occurred as low as 210-260°C, 800°C in Scheme 2 seemed much higher than necessary for the peak temperatures of redox reactions of oxygen carriers.Reasons for operating CLC at such high temperatures are solely dependent on the purity of CO 2 generated (thermodynamics), and prevention of sulfur deposits on oxygen carriers (thermodynamics), as well as consideration to enhance redox kinetics of CuAl 2 O 4 spinel formed during cyclic operation of oxygen carriers.Under such high operation temperatures, there were serious agglomeration tendency of the CA or CCA Samples.This was in contrast to substantial increases of the resistance to agglomeration of lanthanum-promoted copper-based oxygen carriers (CLCA) as seen in over 50 cyclic TPR-TPO cycles.However, further studies indicated that structural change had still occurred in nanometer dimension based on pore-structure analysis.This necessitates the continuous investigation into the interactions between molecules inside oxygen carriers before prepared oxygen carriers for use in the pilot-scale and commercial application.Furthermore, the new phase between CuO-Cu and Al 2 O 3 deserved to be further studied.If this new phase could be stable under higher temperatures (compared to definitely unstable tendency of CuO-Cu under similar temperatures), this may give us an opportunities to enhance its formation in order to obtain the thermal-stable Cu-based oxygen carriers.This has been applied in the continuation of this study in a 100 W and a 10 kW CLC facilities in ICSET, and also frequently found in tests of copper-based oxygen carriers in other research groups.

CONCLUSIONS
In this study, a temperature-programmed reduction and oxidation (TPR-TPO) technique (Redox) was used to simulate of the cyclic operation of copper-based oxygen carriers in the CLC process.Multiple material characterization methods were applied to the fresh and used oxygen carriers, including the BET, XRD and SEM-EDX.The formula of the copper-based oxygen carrier has been finally optimized to prevent its agglomeration, which is a major issue for copper based oxygen carriers.This thermally stable copper-based oxygen carrier was prepared using commercially-available Al 2 O 3 as the supporting material and promoted by Lanthanum.A substantial resistance to agglomeration and durability in over 50 cyclic Redox tests under high temperature conditions (oxidation and reduction at 800°C) has been confirmed.The supporting material (Al 2 O 3 ) significantly impacted the oxygen releasing performance and thermal stability.The formation of a new crystal phase (CuAl 2 O 4 ), cross-linking CuO and γ-Al 2 O 3 , was supposedly to contribute to the stability of copperbased oxygen carrier under severe condition (close to melting point of copper).Lanthanum contributed speeding the formation of the cross-linking between CuO and Al 2 O 3 .

ACKKOWLEDGEMENTS
We gratefully acknowledge mutual financial supports through projects by the United State Department of Energy (DE-FC26-FE0001808), Energy and Environmental Cabinet (Department for Energy Development and Independence Commonwealth of Kentucky) (PO2 127 0900025004 1) and Cenovus FCCL Ltd in Canada

Fig. 2 .
Fig. 2. SEM images of prepared oxygen carrier samples before and after TPR-TPO tests.Scheme 2. As indicated in Figs.3(b) and Fig.3(c), a comparison of fresh and tested Sample CLCA had shown that the melting or agglomeration was significantly reduced on the surface of Sample CLCA over an extended 50 TPR-TPO cycles even under the more severe conditions of Scheme 2. It was expected that Sample CLCA would experience some over-heating during TPR-TPO cyclic tests in Scheme 2 because of exothermic reduction reaction by copperbased oxygen carriers.Thus, the controlled test temperature of 800°C was likely just a "skin" or "surface" temperature of oxygen carriers, and actual "bulk" temperatures of oxygen carriers might be higher than 800°C and could approach the melting point of Cu, as a result of the exothermic reduction of CuO by H 2 .This had inspired us to further modify temperature measurement system in the current commercial ChemiSorb.Table1presents the oxygen transfer capability, in view of the measured CuO content of oxygen carriers using H 2 .For all three copper-based oxygen carriers, the oxygen transfer capabilities were almost stable during the cyclic operation.For CA sample, the CuO contents ranged between 16 wt% to 20 wt%, CCA from 22.9 wt% to 24.3 wt%, and CLCA from 20 wt% to 23 wt%.Investigation into the intensity curves of TPR under Scheme 1, as shown in Fig.4, indicated there were two stages of reduction starting at lower temperatures.The first reduction peak was around 210°C and the second reduction peak was around 260°C.However, the peak of the intensity curve for pure CuO was identified to be 325°C.Different reduction peaks of CuO Samples in TPR tests indicated different physiochemical occurrence of CuO.Shifts of the peak temperatures of Fig. 3. SEM-EDS analysis of CLCA samples before and after TPR-TPO tests.

Table 1 .
Measurement of CuO loading amount of prepared oxygen carriers in TPR using 10% H 2 .