Article pubs.acs.org/IECR
Cerium Oxide Promoted Iron-based Oxygen Carrier for Chemical Looping Combustion Fang Liu,† Liangyong Chen,† James K. Neathery,† Kozo Saito,‡ and Kunlei Liu*,† †
Center for Applied Energy Research, University of Kentucky, Lexington, Kentucky 40511, United States Department of Mechanical Engineering, University of Kentucky, Lexington, Kentucky 40506, United States
‡
ABSTRACT: Oxygen carrier (OC) development is an important topic in chemical looping combustion (CLC). Bimetal oxide OCs usually impart better performance than monometal oxide OCs; one example of which is the introduction of CeO2 as a partially reducible material capable of generating oxygen vacancies that lead to faster oxygen transfer inside OC particle. In this study, CeO2 was used as an additive to a Fe2O3-based OC and its effect on physical properties, such as morphology, surface area and crushing strength, was analyzed in detail. The reactivity of OCs during reduction and oxidation was studied using thermogravimetric analysis mass spectrometry and a bench scale CLC setup. The results showed that the reduction reaction at the OC surface was independent of whether CeO2 was present or not, but after the surface oxygen had been consumed during the oxidation of fuel, the OC with CeO2 additive provided faster oxygen transfer rates from the bulk to the surface to produce better average reaction rates. The OCs after reduction and oxidation cycles were characterized by using X-ray diffraction and Raman scattering techniques. The promotional role of the CeO2 additive is postulated that it enables the creation of oxygen vacancies in a solid solution. These vacancies were able to transfer oxygen from Fe2O3 quickly to the surface of the OC by vacancy diffusion or even through an oxygen tunnel formed by vacancies. The formation of a CeO2 and Fe2O3 solid solution provides the prerequisite for these short-range interactions. nontoxic8 and could become more advantageous than these other oxides for commercial scale use if its reactivity could be enhanced. CeO2 has a fluorite structure, with each Ce4+ surrounded by eight equivalent, nearest O2− ions that form the corners of a cube and coordinated to four Ce4+ ions.9 When the Ce4+ ions are replaced by lower valence cations, oxygen vacancies will be created.10 Furthermore, when two cerium ions are replaced by trivalent ions, for example, by two Ce3+ ions, an oxygen vacancy, or lattice defect, is created that can be the most reactive site.11 Both surface and bulk oxygen vacancies occur in CeO2 and are suitable sites for adsorption.12 Hence, besides being useful as a support,13−18 CeO2 also can actively participate in chemical reactions, and the CeO2-containing OCs have shown promising results in CLC applications,19−24 in CH4 reforming for H2 and CO production17,25−38 (chemical looping reforming) and in CO2 splitting.39,40 Bhavsar et al.21 found that Ni supported on reducible oxides CeO2, showed accelerated reduction kinetics and significantly enhanced oxygen utilization compared to Ni supported on conventional, nonreducible oxide supports (Al2O3, SiO2). Miller et al.19 and Siriwardane et al.22 added 5% CeO2 to hematite (Fe2O3) and found improved OC performance for CH4 oxidation. Li et al.35 used Fe2O3 and CeO2 materials during the conversion of CH4 to synthesis gas and reported that Fe2O3 and CeO2 formed a solid solution having higher activity and selectivity than hematite by itself. Galvita et al.39 investigated the effect of
1. INTRODUCTION Recent studies1,2 show that CO2 concentration reached a level of 400 ppm in 2013, or 40% above preindustrial levels. The contribution of CO2 from industrial activity to increased global CO2 concentrations is widely accepted and points to the need to reduce the emission of this greenhouse gas.3 One possible combustion technology that shows promise for reducing CO2 emissions is chemical looping combustion (CLC). It is an oxyfuel technology, but has the advantages of in situ oxygen separation, low NOx emissions and low cost of CO2 emission abatement. It entails the use of an oxygen carrier (OC), usually a metal oxide, to provide oxygen for combusting fuels in a fuel reactor (FR) and then the reoxidization of the OC in a separate air reactor (AR). Through this cycled reaction, the flue gas is separated into two streams composed of effluent from the AR, which contains mostly N2 with a certain excess O2, and the effluent from the FR, which is primarily water vapor and CO2. Because water vapor is easily condensed, it is possible to produce an exhaust gas of highly concentrated CO2, which can be readily compressed and stored. The development of OCs is one of the important requirements to realize the advantage of CLC on a commercial scale because the reactions within the FR are inherently slow. Much effort has been directed at developing suitable OCs4 and their reactivity has been the key focus in these evaluations. OCs with a high reactivity will reduce the reaction time, and decrease the OC inventory; thereby, the size of the reactors can be lowered significantly, and operating and capital costs would be decreased.5 Compared to NiO, CuO and Mn2O3-based OCs, Fe2O3based OCs are known to have a relatively lower reactivity.6,7 However, iron oxide is more abundant, cost-effective, is © 2014 American Chemical Society
Received: Revised: Accepted: Published: 16341
August 8, 2014 September 23, 2014 October 1, 2014 October 1, 2014 dx.doi.org/10.1021/ie503160b | Ind. Eng. Chem. Res. 2014, 53, 16341−16348
Industrial & Engineering Chemistry Research
Article
checked using the silica vibrational mode at 520.7 cm−1. The Xray diffraction (XRD) patterns of the OCs were acquired using Cu Kα irradiation and a Rigaku SmartLab system with a 2θ range of 20−90°. The morphology of the OC particles was examined using scanning electron microscopy (SEM, Hitachi S4800) with the accelerating voltage set to 15 kV. The crushing strength was measured by an averaging of 30 measurements from a Shimpo FGE-10X instrument using particles that had been randomly extracted from samples having a particle size distribution range of the reaction tested samples. The BET surface area was determined using a Micromeritics ASAP 2000 gas adsorption analyzer in which each sample was degassed overnight at 160 °C and then subjected to isothermal N2 adsorption−desorption measurements at 77 K. 2.3. Thermal Experimental Setup. Performance of OCs was studied in a thermogravimetric analysis mass spectrometry (TG-MS) system (Netzsch STA449C and Netzsch QMS 403C) and a bench scale CLC setup. The TG-MS system is shown in Figure 1. About 500 mg (±1 mg) of OC was used. And the reaction temperature was set to 950 °C. The OC was alternately oxidized or reduced using O2 or CO diluted in argon, respectively. Mass flow controllers (MFCs) were used to control the flow rate of the feed gases and a LabView program was designed and used to control and monitor the MFCs. The redox cycles followed a sequence of: (i) 5 min argon purge at 200 mL min−1; (ii) 20 min oxidation with 20% O2 balanced with argon, at 200 mL min−1; (iii) 5 min argon purge at 200 mL min−1; (iv) 30 min reduction with 20% CO balanced with argon, at 200 mL min−1. The exhaust gas from the TG was analyzed by mass spectrometry (MS) operated in a multiple ion detection (MID) mode. A bench scale reaction system, shown in Figure 2, was also used to test the OC performance. The reactor is a fluidized bed reactor, fabricated from 310 stainless steel pipe with an inner diameter of 5 cm and height of 75 cm. It was used to simulate the oxidation and reduction of a CLC system via a switching valve that alternately enabled either oxidizing or reducing gas into the bottom of the reactor. During each test, about 100 g of the OCs was placed onto a height-adjustable distributor at the bottom of reactor, and the bed height of the OCs was approximately 7.5 cm. A K-type thermocouple was immersed into the center of the OC bed. The operating temperature was 950 °C. The minimum fluidization velocity at the temperature of 950 °C was around 2.9 cm s−1, corresponding to a flow rate of 0.83 L min−1. In operational conditions, it was determined that a 5 L min−1 flow rate of feed gas was appropriate to obtain a desired fluidization state. A Rosemont X-stream infrared multichannel gas analyzer was used to measure the composition (O2, CO, CO2 and CH4) of off-gas from the reactor. The fuel used was a mixture of 23% CO balanced with N2 with a flow rate of 5 L min−1.
CeO2 upon Fe2O3 during H2−CO2 redox reactions for CO2 utilization and found enhanced reaction capacity and increased stability relative to Fe2O3 by itself. As for the enhance mechanism, Miller et al.19 theorized that CH4 first reacted with CeO2 lattice oxygen to form CO and H2, and then the reduced cerium oxide promoted CH4 decomposition to active intermediates such as C and H2 which, in turn, facilitated a greater use of oxygen from the natural ore hematite. Li et al.30,34,37 interpreted the mechanism as the Fe2O3 and CeO2 mixture could better disperse surface Fe2O3 and the solid solution could enhance the oxygen mobility. Galvita et al.,39 on the other hand, attributed the promotional role of CeO2 to that of suppressing the sintering of iron oxide. However, the deep understanding of mechanism of how CeO2 interacts with Fe2O3 and promotes Fe2O3-based OCs during CLC is still needed. In this study, a small amount of CeO2 was used as an additive to the Fe2O3-based OC rather than being one of major compounds as a support. The effects of CeO2 additive on particles’ physical properties (morphology, Brunauer−Emmett−Teller (BET) surface area, mechanical strength) were investigated in detail. The OCs were assessed using a TG-MS and a bench scale CLC experimental setup. The compositions and phases of the OCs were examined by X-ray diffraction (XRD) and Raman spectroscopy. From these data is analyzed the mechanism of the CeO2 promotional role within the Fe2O3based OC.
2. EXPERIMENTAL SECTION 2.1. Oxygen Carriers. Commercially available metal oxides were used for OC preparation. Aluminum oxide powder (Sigma-Aldrich, standard grade), iron oxide (Sigma-Aldrich, 99.0% purity) and cerium oxide (Strem Chemicals, 99.9%) were used to prepare OC materials by a freeze granulation method. The OC slurry was prepared by ball milling mixtures of metal oxides, dispersant (A40) and binder (PVA) in deionized water for 12 h. The mass-based ratio of metal oxides, deionized water, binder and dispersant was 1:1:0.1:0.01. The well-mixed slurry was sprayed through a nozzle into liquid nitrogen, leading to the formation of frozen spherical particles which were then dried in a freeze-dryer (Virtis Advantage Plus). The dried particles were calcined in air at 1400 °C for 6 h. Two types of OC, labeled as OC #1 and OC #2, were produced for this study, with the composition of 10% CeO2, 50% Fe2O3 and 40% Al2O3 for OC #1, 50% Fe2O3 and 50% Al2O3 for OC #2. OC particles were sieved in the range of 150−300 μm, and particle distributions are listed in Table 1. Table 1. Properties of OCs
OC #1 OC #2
median diameter (μm)
mean diameter (μm)
apparent density (g/cm3)
surface area (m2/g)
pore volume (cm3/g)
216.0
232.6
0.98
0.34