Lei Liu,Zhenshan Li,Ye Li,Ningsheng Cai

Key Laboratory for Thermal Science and Power Engineering of Ministry of Education,Department of Energy and Power Engineering,Tsinghua University,Beijing 100084,China

Keywords:CO2 capture Oxygen carrier Oxygen uncoupling Fluidized-bed Thermogravimetric analysis Agglomeration

ABSTRACT Oxygen uncoupling characteristics of a natural manganese ore and a perovskite-type oxide CaMn0.5Ti0.375Fe0.125O3 were studied by using a microfluidized bed thermogravimetric analysis(MFB-TGA) technology which is based on a real-time mass measurement of fluidizing particles inside a bubbling bed reactor.The chemical stability,kinetics of the oxygen release and uptake reactions and fluidization property were investigated and the experimental data measured by MFB-TGA were compared with the results in a regular TGA instrument (TGA Q500).The regular TGA Q500 results show the reactivity of both the manganese ore and perovskite oxide are stable for multi cycles,and the oxygen uncoupling capacity of the manganese ore is～1.2%(mass)which is～2 times higher than that of the perovskite oxide.However,the experimental results from the MFB-TGA indicated that there is a serious agglomeration for the manganese ore.A very important finding is that the reaction rate of oxygen release and oxygen uptake of the perovskite oxide measured by the MFB-TGA are～2 and～4 times faster than that of testedby the TGA Q500.We can conclude that MFB-TGA is a very useful tool to measure the reactivity stability and kinetics of oxygen carriers in high-throughput analysis instead of the regular TGA.

1.Introduction

Chemical looping combustion(CLC)is an alternative technology for reducing the economic and energy costs associated with the CO2capture from the fossil fuel conversion processes [1–3].A CLC unit consists of two interconnected fluidized bed reactors,an air reactor and a fuel reactor.A metal oxide,named oxygen carrier,transports the active lattice oxygen from the air reactor to the fuel reactor,to convert the volatiles and syngas (CO and H2) from the pyrolysis and gasification of solid fuel,such as coal,petroleum coke,biomass,etc.[3,4].Then,the reduced oxygen carrier is carried back to the air reactor again and oxidized by air.Chemical looping with oxygen uncoupling (CLOU) takes advantages of gaseous oxygen released from some kinds of oxygen carrier materials,such as Cu-based and Mn-based synthetic oxides or ores,to burn the solid fuel in the fuel reactor[5–7].Combustion with gas phase of oxygen has been proven to be efficient and fast in terms of converting solid fuel char[5,8].To further improve the oxygen uncoupling capacity,some bimetallic and polymetallic oxide oxygen carriers have also been developed,such as Ca-Mn-based and Ca-Mn-Fe-based perovskite oxides,due to more suitable thermodynamics for O2release at in a lower oxygen partial pressure [3,9,10–12].

Oxygen release and uptake of an oxygen carrier in CLOU plays an important role in converting the syngas and char,thermogravimetric analysis(TGA)is a widely used experimental tool to test the stability and kinetics of oxygen carriers by recording the real-time mass signal of the sample,including redox reactivity and oxygen coupling capacity [13,14].However,in TGA,poor mass and heat transfer could significantly underestimate the reactions kinetics,since the sample particles are located in a packed state inside the crucible,which leads to a significant gas diffusion resistance [15].Furthermore,the packed state of sample in TGA cannot reflect the actual situation of fluidization in the common CLC/CLOU reactors[16].The fluidized bed reactor is another method used for the reactivity and kinetic study of oxygen carriers,where the sample particles are well-mixed and mass and heat transfers are enhanced[17].Based on the concentration signal of the tail gas species,the fluidized bed reactor method faces the major challenges of imperfect plug flow,axial mixing effect and the delay of gas signals analysis as well,which leads to a large error for the gas concentration signal,especially when a reaction is very fast within several seconds,such as the oxidization reaction of Cu-based and Mn-based oxygen carriers [18,19].In this situation,deconvolution method was proposed to correct the concentration signal,as demonstrated by the oxidation of Cu2O[17,18].However,it is hard for the deconvolution method to eliminate the signals delay caused by the effect of mixing in the reactor and gas sampling system on the fast oxidation reaction [18].To conclude,although the chemical stability of oxygen carrier can be studied by TGA and fluidized bed reactor,the kinetic measurement for redox reaction and oxygen uncoupling is still a big challenge in the field of CLC or CLOU.Therefore,it is very necessary to develop a method that is able to measure effectively the fast redox kinetics of oxygen carriers.

In this work,the micro-fluidized bed thermogravimetric analysis (MFB-TGA) method is used to evaluate the CLOU characteristic of a manganese ore and a CaMn0.5Ti0.375Fe0.125O3-based perovskitetype oxide.The novel point of MFB-TGA is the online measurement of fluidizing particles mass in a bubbling bed reactor.The kinetics of oxygen release and uptake were studied and compared with the results measured in a regular TGA (TGA Q500).

2.Experiment

2.1.Oxygen carrier materials

In this work,a natural manganese ore and a CaMn0.5Ti0.375-Fe0.125O3-based perovskite-type oxide were used as oxygen carrier,and their CLOU properties were investigated and compared in regular thermogravimetric analysis and micro-fluidized bed thermogravimetric analysis.Mn-containing oxides and perovskite oxides are two of oxygen carriers possessing the CLOU property [20].The natural manganese ore has attracted great attention by Technology University of Chalmers,due to its low cost,good reactivity,and CLOU property as well [7].Recently,the perovskite oxide,CaMnO3-δ,has been developed as an oxygen carrier,considering this material has been considered as a promising oxygen carrier for CLOU application [21].Among them,the CaMn1-x-yTixFeyO3-δseries of oxygen carrier materials were developed and investigated[10].The raw manganese ore,from Fujian,China,is very fine.Therefore,powders of raw manganese ore were treated by rolling granulation in a micro-rotary furnace with a heater.The granulation method was described in previous published work in detail[19].Powders were uniformly stirred in the rotating furnace for 1 h.After that,a dextrin solution was injected and many spherical particles gradually formed.Next,the obtained spherical particles were calcined in air at 1100 °C for 8 h then cooled to 850 °C and calcined in air for 4 h to stabilize the structure.Finally,the particles were crushed and sieved to 180–250 μm.The materials were mainly composed of 49.20% (mass) MnO,14.42% (mass) Fe2O3,11.38% (mass) K2O,7.76% (mass) SiO2,3.42% (mass) Al2O3,1.25%(mass)CaO,etc.,determined by XRF.The perovskite oxygen carrier from SINTEF Industry,Norway,were prepared by means of spray drying prior to calcination at 1350°C for 12 h The perovskite materials were composed of 40.52% (mass) CaO,26.38% (mass) TiO2,22.94% (mass) MnO and 7.07% (mass) Fe2O3determined by XRF.Its particles were sieved between 180–250 μm for test.More properties of the oxygen carriers are listed in Table 1.

Table 1 Physical properties of the oxygen carrier samples

2.2.Regular thermogravimetric analysis (TGA)

Oxygen uptake and oxygen release cycles were test in a thermogravimetric analysis(TGA)instrument using a TGA Q500 system as a representative of regular TGA instruments.TGA is a widely used experimental method for oxygen carrier materials study.The TGA Q500 is composed of a mass transducer with a measurement accuracy of 0.001 mg,a furnace chamber,a sample pan (alumina crucible),etc.,see Fig.1.For acquiring the reactivity and stability of the tested oxygen carriers,the reaction temperature and sample weight were continuously recorded by a data acquisition system connected to the TGA.The test was carried out at 900°C by changing the gas atmosphere between 21%(vol)O2in 79%(vol)N2for 30 min and 100%(vol)N2for 60 min.The gas flowrate was controlled at 100 ml min-1.For each gas switching,such as switching the gas agent from 21% (vol) O2in 79% (vol) N2to 100% (vol) N2,a N2gas stream is used to purge the space in furnace chamber.For regular TGA instruments,the particles are packed on the pan.To minimize the effect of mass transfer occurring inside the packed particle layer on the kinetics,about 15 mg of oxygen carrier sample was placed into the crucible for test.

2.3.Micro-fluidized bed thermogravimetric analysis (MFB-TGA)

A micro-fluidized bed thermogravimetric analysis apparatus(MFB-TGA) was applied to investigate the CLOU properties of the oxygen carriers under fluidizing state.The apparatus is mainly composed of an electric furnace,a bubbling fluidized bed reactor,a gas supply system and a measuring system,see Fig.2.Gas flowrates were controlled by calibrated mass flowmeters.The gas was injected into the bottom of the reactor through switching wellconnected magnetic valves.The bubbling bed reactor was placed on the mass transducer with a measurement accuracy of 1 mg.The bed temperature was measured by a K-type thermocouple enclosed in the reactor above the gas distributor.A differential pressure sensor was installed to measure the pressure difference between the gas inlet and outlet.The details can be found elsewhere[22,23].The MFB-TGA with fast mass and heat transfer provides an option for approaching to a more real kinetics than conventional TGA instruments.For instance,the gasification kinetics of a lignite char measured in MFB-TGA are 2 to 4 times faster than that measured in a thermogravimetric analysis instrument(TGA) [23].

During a typical test,about 16,000 mg of silica sand was added into the reactor as the bed material about 3000 mg fullyoxidized samples were added from the top of the reactor before heating up to 900 °C under a 21% (vol) O2fluidizing agent.When the mass and temperature signals were stable,20 cycles of isothermal oxygen uptake and oxygen release were conducted by alternating the fluidizing agent between 21% (vol) O2for 60 seconds and 100% (vol) N2for 300 seconds.The fluidizing gas flowrates were set to 1200 ml﹒min-1to achieve the superficial velocity of about 4 Umf.Previous work has shown that the stability of the MFB-TGA can be guaranteed at the superficial velocity of 0–6 Umf[23].To eliminate the noise during gas switch between 21% (vol) O2and 100 vol% N2,3.50% (vol) He mixed with 75.50% (vol) N2were serviced as the balance gas in the 21% (vol) O2agent.In the result,the 21% (vol) O2agent keeps similar densities with 100 vol% N2.Finally,the effect of gas switch on mass signals could be ignored,see Fig.2(b).The mass signals were not affected during gas switching and the experimental error was ± 1 mg [23].The response time was within 1 second,which cause that the MFB-TGA could take advantage of the quick gas switch and purge.Therefore,multicycle test was conducted to study oxygen uptake for 30 seconds and oxygen release for 60 seconds,respectively.In addition,while performing kinetic studies,the MFB-TGA can also detect and judge the fluidization performance of oxygen carriers,as discussed below.

Fig.1.Schematic diagram of the TGA Q500.

Fig.2.(a)Schematic diagram of the MFB-TGA;(b)Mass change between the gases switch at 900 °C (Only 16000 mg silica sand was filled).

2.4.Data evaluation

In order to compare the experimental data acquired in the TGA Q500 and MFB-TGA,the mass change versus time was recorded in real time.The mass change of the tested oxygen carriers during the oxygen uncoupling process,as well as oxidization processes with O2is calculated by the following equation,

where mois the initial mass of added oxygen carrier samples in the TGA Q500 or MFB-TGA;mtrepresents the mass measured by the mass transducer at time t.

3.Results and Discussion

3.1.Multicycles of oxygen carrier in TGA Q500

The multicycles of oxygen uptake and oxygen release processes were carried out in the TGA Q500 under 900 °C,and mass change versus time of the manganese ore and perovskite were shown in Figs.3 and 4.For each cycle,both the manganese ore and perovskite lasted a long time for the oxygen uncoupling and oxidation reactions.The CLOU characteristics of the manganese ore could be obtained and there was a slight decay in the capacity of oxygen uncoupling for 20 cycles,see Fig.3.The mass loss is decreased from～1.8%to～1.2%(mass)during the oxygen uncoupling processes.The CLOU property of the manganese ore can be stabilized at～1.2%(mass) after the 16th cycles.Based on the result in the TGA Q500,although there is a decay in the reactivity,the oxygen uncoupling capacity is still～1.2% (mass) of the lattice oxygen in the manganese ore oxygen carrier,and this fraction of lattice oxygen can be released in the form of gaseous oxygen in the fuel reactor.Fig.4 shows the CLOU property of CaMn0.5Ti0.375Fe0.125O3oxygen carrier.It can be seen that the capacity of the oxygen uncoupling was stable with the increasing cycle numbers and there is no decay for the 20 cycles.According to the result in the TGA Q500,the mass loss of CaMn0.5Ti0.375Fe0.125O3was～0.65%(mass)during the oxygen uncoupling processes.The experimental results in Fig.3 show that the manganese ore undergoes a mass loss during the multicyclic test.That is to say,the mass of the manganese ore oxygen carrier continuously decreases.In contrast,the mass of perovskite oxide can be kept stable,as shown in Fig.4.The manganese ore oxygen carrier contains 11.38% (mass) K2O,and the melting point of K2O (～770 °C) is lower than that of other metal oxides.It is well known that the low-melting K2O is prone to be evaporated at a high temperature.Therefore,there would be slight evaporation of K2O at the experimental temperature,which results in a decrease in the mass of the manganese ore oxygen carrier.A non-isothermal experiment was conducted,see Fig.S1 in the Supplementary Material.The results show that the mass loss of the manganese ore oxygen carrier occurred once the operating temperature approached to the melting point of K2O.The K2O content was decreased by 2.7% (mass).The TGA Q500 results shows the manganese ore has a lager potential than the perovskite oxide as oxygen carrier for the CLOU application,considering～1.2% (mass)oxygen uncoupling capacity.

Fig.3.The multicycle results of the manganese ore in TGA Q500.

Fig.4.The multicycle results of CaMn0.5Ti0.375Fe0.125O3 in TGA Q500.

Fig.5.The multicycle results of CaMn0.5Ti0.375Fe0.125O3 in MFB-TGA (oxygen uptake:60 seconds;oxygen release:300 seconds).

3.2.Multicycles of oxygen carrier in MFB-TGA

The oxygen uptake and release processes of CaMn0.5Ti0.375-Fe0.125O3oxygen carrier at 900 °C are conducted in MFB-TGA.The oxygen uncoupling results of the manganese ore are not shown here because of defluidization problem occurring in the bubbling bed reactor,as discussed below.Fig.5 shows the reactivity and stability of CaMn0.5Ti0.375Fe0.125O3oxygen carrier in MFBTGA for CLOU process.4th,11th and 18th cycles,as representatives,show the kinetics of the oxygen uncoupling and oxidation reactions are repeatable.The mass loss of oxygen release is maintained at～0.6%(mass)and there is no decay in the capacity of oxygen uncoupling processes,as same as the obtained result in TGA Q500.However,the reaction times for the oxygen uptake/release processes are 60/300 seconds respectively,which are far shorter than that of investigated in TGA Q500.Furthermore,another 20 cycles were carried out in MFB-TGA and the times for the oxygen uptake/release processes are shortened to 30/60 seconds,as shown in Fig.6.Similarly,taking the 4th,11th and 18th cycles as representatives,the experimental results are repeatable.It can be observed that the CLOU property of CaMn0.5Ti0.375Fe0.125O3oxygen carrier is stable and the mass loss of oxygen uncoupling is～0.4%(mass).Experimental data shown in Figs.5 and 6 indicate that MFB-TGA can not only test the stability and kinetics of oxygen carriers like regular TGA,but also has the characteristics of highthroughput analysis of oxygen carriers.The oxygen uptake and release processes can be achieved in short time,due to very quick gas switch and purge,as well as very low limitation of gas diffusion in fluidization bed.Moreover,it can be observed that the reproducibility of MFB-TGA result is sufficient because a relatively larger amount of oxygen carrier sample was used (3 g CaMn0.5Ti0.375-Fe0.125O3particles).It should be noted that the mass loss of CaMn0.5Ti0.375Fe0.125O3during the multicyclic test in Fig.5 might be caused by the attrition of silica sand,since the oxygen uncoupling capacity is stable at～0.6% (mass) and the experiments are reproducible with cycle number.

3.3.Comparison between MFB-TGA and TGA Q500

Comparing the experimental results of oxygen uncoupling processes of CaMn0.5Ti0.375Fe0.125O3between MFB-TGA and TGA Q500,it can be further proven that MFB-TGA enables stable and continuous testing of oxygen carriers,as seen in Fig.7.It should be noted that the mass losses obtained by TGA Q500 and MFB-TGA are different,mainly caused by the time for oxygen coupling in MFB-TGA is far shorter than that in TGA Q500.

Fig.6.The multicycle results of CaMn0.5Ti0.375Fe0.125O3 in MFB-TGA (oxygen uptake:30 seconds;oxygen release:60 seconds).

Fig.7.Comparison of stability of oxygen uncoupling processes of CaMn0.5Ti0.375Fe0.125O3 between MFB-TGA and TGA Q500.

Fig.8.Comparison the kinetics of (a) oxygen release and (b) oxygen uptake of CaMn0.5Ti0.375Fe0.125O3 between MFB-TGA and TGA Q500.

From Fig.8,the reaction rates of oxygen uncoupling and oxidation of the perovskite oxide between MFB-TGA and TGA Q500 were compared.The comparison clearly illustrates the difference of oxygen release and oxygen uptake between MFB-TGA and TGA Q500.The reproducibility results of MFB-TGA has been showed before.Fig.8(a) and (b) shows that both the oxygen release and oxygen uptake rates in MFB-TGA were obviously much faster than that in TGA Q500 under the same conditions.It can be seen that the time to reach a mass loss of 0.2% in TGA Q500 was almost double than required in MFB-TGA during the oxygen uncoupling processes.The gas diffusion and mass transfer effect could be responsible for the difference.In MFB-TGA,the quick gas purge using a big gas flowrate(1.2 L﹒min-1)was achieved,while a small gas flowrate(100 ml﹒min-1) in TGA Q500 was used.This means that the O2in the reaction zone could be purged by N2in MFB-TGA in a very short time.In addition,the released oxygen has almost no effect on the gas phase oxygen concentration because of the large amount of purging gas used by the MFB-TGA.For the oxygen uptake process,the full oxidation of the perovskite can be finished within～7.5 s in MFB-TGA which is～4 times faster than that in TGA Q500.It is commonly acknowledged that the mass and heat transfer limitations occurring inside regular TGA will probably underestimate the reaction rate [13].Therefore,the oxygen uncoupling rate obtained by MFB-TGA are more accurate and reliable for the evaluation of CLOU property of oxygen carrier materials.

3.4.Fluidization of oxygen carriers in MFB-TGA

Fig.9.The pressure difference curve of MFB-TGA when testing CaMn0.5Ti0.375Fe0.125O3.

Fig.10.(a)The pressure difference curve of MFB-TGA when testing manganese ore;(b) The photo and schematic diagram of the defluidized manganese ore in MFBTGA.

During measuring the CLOU property,the fluidization property of the tested oxygen carriers can be detected in MFB-TGA simultaneously.The gas flowrate of the reactor was controlled at 1200 ml﹒min-1before heating the reactor from room temperature to 900 °C.The bed materials were in incipient fluidization at the initial state.During the rising temperature and stability test processes,the pressure difference signals between the inlet and outlet of the bubbling bed reactor were collected.The fluidization property can be judged by the fluctuation of pressure difference signals.Fig.9 shows the pressure difference versus time when using CaMn0.5Ti0.375Fe0.125O3as oxygen carrier in the MFB-TGA.The pressure difference would be rising with the increasing temperature during the heating stage,due to the increasing gas velocity at an unchanged gas flowrate,until the reactor temperature was stabilized at 900°C.There was a corresponding transition of bed materials from incipient fluidization to bubbling fluidization.From the fluctuation of pressure difference curve,agglomeration of oxygen carrier particles did not occur at 900 °C.Therefore,based on the experimental data in MFB-TGA,the CaMn0.5Ti0.375Fe0.125O3oxygen carrier has a good performance in terms of fluidization.

However,for the manganese ore,once heated up to～750°C,the pressure difference curve had a drastic change,the fluctuation of pressure drops became to a smooth line.This pressure drop signals change means the oxygen carrier agglomerated and defluidization occurred in MFB-TGA.The bed materials had a transition from incipient fluidization to agglomeration in MFB-TGA,as demonstrated in Fig.10.The sintering of the manganese ore is due to its low melting temperature caused by K2O [24].There is 11.38%(mass) K2O in the manganese ore,the melting point of K2O is～770 °C.In the test,the defluidization temperature is clear to the melting point of K2O.Once the K2O phase melted on the surface of the manganese ores,the particles became sticky and agglomeration occurred.Therefore,according to the experimental results,it can be judged that the manganese ore is not possible as oxygen carrier for CLOU application,which is contrary to the result obtained by TGA Q500.It can be concluded that MFB-TGA has the function to detect the fluidization properties of oxygen carrier materials during the reactivity and kinetics test,compared with the regular TGA.

4.Conclusions

The micro-fluidized bed thermogravimetric analysis(MFB-TGA)method is used to investigate the CLOU property of oxygen carriers,and the obtained results are compared with a regular thermogravimetric analysis (TGA) instrument,TGA Q500.A natural manganese ore and a perovskite-type CaMn0.5Ti0.375Fe0.125O3were selected as oxygen carrier materials,and the experimental results in MFB-TGA and TGA Q500 were analyzed to demonstrate the advantage of using MFB-TGA to test oxygen carriers.The TGA Q500 results show that the amounts of oxygen uncoupling of the manganese ore and perovskite are stabilized at～1.2% and 0.65%(mass),respectively,and the manganese ore appears to have more potential as an oxygen carrier than the perovskite oxide for CLOU application.However,the results from MFB-TGA shows agglomeration and defluidization would occur in fluidization bed when using the manganese ore as oxygen carrier.Comparing the experimental results of the perovskite oxide in MFB-TGA with that in TGA Q500,the oxygen uptake and oxygen release are stable and there is no decay in the capacity of oxygen uncoupling.However,the kinetics of oxygen release and oxygen uptake reaction are～2 and～4 times faster than that of tested in TGA Q500,respectively,due to the strong mass and heat transfer in fluidized bed.Furthermore,the pressure drop curve shows agglomeration did not occur for CaMn0.5Ti0.375Fe0.125O3.It can be concluded that MFB-TGA can take advantage of high-throughput analysis of oxygen carriers in terms of the reactivity and kinetics to replace the regular TGA.Besides,the fluidization property could be investigated with the help of MFB-TGA.

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 research has received funding from the National Natural Science Foundation of China (51976102),the National Key Research and Development Plan of China (2016YFB0600802-A and No.2017YFE0112500).

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2020.11.023.

Nomenclature

moinitial mass of added oxygen carrier samples,mg

mtreal-time mass at time t,mg

t time,s

UmfMinimum fluidization velocity,m﹒s-1