Application of Ba_0.5Sr_0.5Co_0.8Fe_0.2O_3−δ membranes in an oxy-fuel combustion reactor

Highlights

• Long-term performance and stability of BSCF oxygen transport membrane in an oxy-fuel combustion reactor has been evaluated.

• Effect of fuel flow rate on BSCF membrane stability.

• Continuous and stable oxy-fuel combustion using BSCF membrane when 15% excess oxygen is present at the permeate side.

• Highest flux of oxygen using BSCF membranes for applications in oxy-fuel combustion.

ABSTRACT

Oxygen separation from air through ceramic ion transport membranes is anticipated to be one of the technologies that can make fossil fuel power plants to operate without carbon dioxide (CO₂) emission. Consequently, the level of carbon emission in the atmosphere, which causes global warming, could be contained. The most promising material for such ion transport membranes is the Ba_0.5Sr_0.5Co_0.8Fe_0.2O_3−δ (BSCF), which has been reported to have the highest oxygen permeation flux. However, lack of chemical stability of these membranes in CO₂ rich environment is recognized as a major problem. In this study, the performance and stability of BSCF membranes in an oxy-fuel combustor have been evaluated using methane gas. The flow rate of the methane (CH₄) gas has been optimized to the value of 0.65 ml/min at 920 °C. Under this flow rate, the BSCF stability has also been evaluated. The study has shown that the BSCF stability depends on the amount of excess-oxygen in the reactor. If all oxygen, generated by the membrane, is consumed during the combustion reaction then BSCF starts to deteriorate and its permeability decreases with time. On the other hand, when the combustion reaction does not consume all oxygen, and a minimum excess-oxygen is retained, the BSCF remains stable with constant oxygen flux. This fact has been established using two BSCF membranes with different thicknesses and same exposed area of 113 mm². The first membrane has 1.4 mm thickness and its initial oxygen permeability under combustion reaction is 0.83 µmol cm^-2 s⁻¹ with an excess of 7.5% oxygen. After 190 h of combustion, the oxygen permeability has dropped to 0.76 µmol cm⁻² s⁻¹ (7.9% reduction). The second BSCF membrane has a thickness of 1.0 mm produces 0.91 µmol cm⁻² s⁻¹ with a 15% excess oxygen during the combustion reaction. This membrane has shown excellent stability for more than 200 h of complete methane combustion. This is a remarkable result which shows that a carefully controlled fuel volume introduced in a BSCF membrane-reactor provides high oxygen output and excellent stability for a long period of time.

Introduction

Global warming is one of the major concerns in environmental management and protection [1]. The emission of CO₂ gas is the main contributor to the global warming as this gas absorbs the infrared radiations that lead to the increase in the overall temperature of the earth. Therefore, efforts are increasingly directed towards reducing of the CO₂ emission. One of the major contributors to the CO₂ emissions (40%) is Power plants, that operate with fossil fuel [2]. There are three possible technical solutions for CO₂ capture in fossil-fuel power plants: post-combustion, pre-combustion, and oxy-fuel combustion [3], [4]. In the oxy-fuel combustion fuel is burnt in a pure oxygen atmosphere and produces high concentration of CO₂ and water steam, which can be easily condensed. Pure oxygen environment, inside the reactor, can be generated by either external oxygen supply or internal oxygen extraction using oxygen-ion transport membranes (OTMs). It has been validated [5], [6] that OTMs are excellent candidates for oxy-fuel combustion because of their low energy consumption and their easy integration into membrane reactors. Dense solid (non-porous) ceramic membranes that exhibit high ionic and electronic conductivity are a great choice for producing pure oxygen from air at high temperatures. The oxygen production by these membranes is quite economical, clean and efficient. It has been demonstrated that the efficiency of oxy-fuel power plants based on OTMs is higher than those operating with cryogenic distillation [7]. Kunze and Spliethoff [8] evaluated the efficiency of the three carbon capture routes (post-combustion, pre-combustion and oxy-fuel combustion) in Integrated Gasification Combined Cycle (IGCC). They concluded that the process using OTMs revealed better net efficiency and had excellent long term potential.

Oxygen-ion Transport Membranes are crystalline ceramic membranes, which contain oxygen-ion vacancies. When these membranes are activated by heat, oxygen-ions can travel through the membrane via vacancy diffusion. Mixed conducting perovskite-type oxides represent one type of the most promising materials satisfying such purposes [9]. When the A-site of perovskite is doped with lower valence state metal ions, oxygen vacancies as well as a change in the valence state of the B ions in the lattice will occur in order to maintain the electrical neutrality [10]. Some of the perovskites have considerably high oxygen ionic conductivity with overwhelming electronic conductivity at elevated temperatures, and usually high permeation fluxes are found for such materials.

Since Teraoka et al. [11] first reported the remarkable high oxygen permeation flux through the ceramic discs based on the La_1−xSr_xCo_1−yFe_yO_3−δ (LSCF) perovskite oxides in the 1980s, cobalt-containing perovskite membranes have been widely investigated. Although the cobalt-based materials are generally recognized to have low chemical stability, long-term operation stability of such membrane reactors have been reported in the literature illustrating that cobalt based mixed conducting membranes are able to operate stably under the stringent partial oxidation condition [12], [13]. Among the various cobalt-based oxygen separation membrane compositions, Ba_0.5Sr_0.5Co_0.8Fe_0.2O_3−δ (BSCF), as first reported by Shao et al., [10] have attracted much interest due to its high oxygen permeability. Many researchers have tried to develop new cobalt-free perovskite materials with intrinsically improved chemical stability, such as La–Sr–Ga–Fe [14], Be–Ce–Fe [15] and Ba–Sr–Zn–Fe [16]. However, thus far, none of the oxygen permeation fluxes could reach the level of cobalt-contained materials.

In a power plant environment, the flue gas stream volume consists of about 25–30% water vapor, 70–75% CO₂, 1–3% O_2, and 400 ppm SO₂ concentration [17]. BSCF membranes, with the highest oxygen permeability [4], are one of the best membrane candidates for oxygen separation in power plant reactors. However, recent studies have shown that BSCF is not an ideal material for long-term service in CO₂ containing environment. This is due to its high reactivity and the formation of (Ba, Sr)CO₃ compounds at the operating temperatures [18], [19]. These instability issues lead to severe degradation in the performance of the BSCF membranes [20], [21].

Many researchers have tried to investigate the performance of the OTMs by measuring oxygen permeability while using CO₂ (pure or mixture) as a sweep gas [17], [18], [22]. The results show that as soon as the inert carrier gas is switched to CO₂ at the sweep side, the permeability of BSCF membranes deteriorate greatly. Other researchers have exposed the BSCF membranes to CO₂ rich environment and used various characterization techniques like XRD, TG/DSC, SEM/EDS etc. to determine the reason for this deterioration [19], [21], [23]. A reaction zone was observed to be formed on the surface of BSCF membranes when annealed in CO₂ containing atmosphere [17]. This layer was determined to be a mixed barium and strontium carbonate of the form (Ba_xSr_1−x)CO₃. It was found that this carbonate layer covers the complete surface of the BSCF pallet when annealed at 700 °C and 800 °C, but for 600 °C only few spots of the surface were covered.

In order to solve the stability issue with BSCF, researchers have tried to modify the membrane materials by doping or replacing the A or the B site of the perovskite structure with other elements. For example, Kim et al. [24] showed that Lanthanum doped BSCF showed good stability under air, CO₂ containing atmosphere, and low oxygen partial pressure. Wang et al. [25] showed that Nb doping could improve the stability of BSCF against CO₂. Efimov et al. [26] considered different calcium containing perovskite materials and concluded that La_0.6Ca_0.4Co_0.8Fe_0.2O_3−δ (LCCF) was one material which, not only retained perovskite structure at high temperatures but also showed constant oxygen flux in CO₂ rich environment without showing any formation of carbonates. Another material investigated for its use as OTM in oxy-fuel combustion was La_0.6Sr_0.4Ti_0.3Fe_0.7O_3–δ (LSTF) which showed strong resistance and durability for CO₂ environment but its oxygen flux was found to be relatively low [27]. A cobalt free composition BaFe_0.55Nb_0.45O_3−δ was tested for oxygen permeability by Yi et al. [28] with CO₂ as sweep gas. It did not show any significant degradation. Similarly, Engels et al. [17] showed that the compositions Sr_0.5Ca_0.5Mn_0.8Fe_0.2O_3−δ (SCMF) and La₂NiO_4+δ (LNO) had high chemical stability in the presence of CO₂ but they do display instability when 360 ppm SO₂ was introduced in the sweep. All of these materials, discussed so far, have better chemical stability than BSCF but considerably low oxygen permeability. Chen et al. [29] measured the oxygen permeability of perovskite-type oxide SrCo_0.8Fe_0.2O_3−δ (SCF). They showed that using pure CO₂ as sweep gas, the membrane starts to degrade immediately. A summary of chemical stability of OTM materials is presented in Table 1, which confirms that most OTMs with high oxygen permeability flux show relatively poor stability in CO₂ environment; whereas, all stable OTMs have low permeation flux.

OTM based oxy-fuel power plants could be operated with either a four-end or a three-end design [30]. In a four-end design an oxygen consuming sweep (syngas or methane) is delivered to the permeate (low pressure) side of the membrane. In this way the partial pressure difference across the membrane is maintained, hence continuous oxygen supply through the membrane [30]. The membrane, in this case, is in a state of a direct contact with the flue-gas; therefore, the material has to resist all components of the flue-gas while producing an acceptable flux of oxygen. Stadler et al. have shown that a power plant based on the four-end concept has better efficiency than that of the three-end concept. On the other hand, in a three-end concept, the production of oxygen is sustained by either applying higher pressure on the feed side or by removing the permeated oxygen using vacuum [31]. The membrane, in this case interacts only with air, can be made from many different materials of high oxygen flux. For this reason, a three-end design concept is considered to be more suitable due to the absence of rigorous chemical stability of the membrane [7]. Castillo [32] has studied BSCF material in three-end membrane concept and has shown that the overall plant efficiency could be 4% higher than that based on cryogenic technology. Gupta et al. [33] have used a tubular OTM device consisting of three layers (oxygen incorporation, oxygen separation, and fuel oxidation layers). The performance of these OTM tubes has been tested in a coal derived syngas fuel. The material used in this process was 40 vol% (La_0.825Sr_0.175)_0.94Cr_0.72Mn_0.26V_0.02O_3−δ and 60 vol% Zr_0.802Sc_0.18Y_0.018O_1.901. The oxygen flux of the membrane was relatively low but stable for 80 h of continuous operation. Another case of CO₂ capture via direct oxy-fuel combustion was reported using methane [34]. The membranes were made in the form of U-shaped hollow fiber of (Pr_0.9La_0.1)₂ (Ni_0.74Cu_0.21Ga_0.05) O_4+δ (PLNCG). These membranes were operating for 450 h with good stability but the oxygen flux values were much lower than that of BSCF membranes.

The approach that has been taken in evaluating stability of OTMs with high permeability, in the reported literature, is to test them in CO₂ rich environment. This approach is good for quick membrane screening for OTM stability; however, it over estimates the real operating conditions of oxy-fuel reactors. Therefore, this approach might eliminate good candidates that are relatively stable in a less harsh environment. In this study, a new approach has been taken, where the OTM is being subjected to actual reactor condition of pure fuel-gas flow. BSCF membranes have been evaluated for oxygen separation using methane gas as fuel on the permeate side. The volume of the methane gas has been optimized to the value where all methane and oxygen gases are consumed by the combustion reaction. Quantitative analysis of all gases (CH₄, O₂, CO, CO₂, N₂) have been studied using an online gas chromatograph. The membrane permeability and stability have been evaluated for long runs, up to 200 h.