- Email: [email protected]
Oxygen sorption and desorption properties of Sr–Co–Fe oxide
Chemical Engineering Science 63 (2008) 2211 – 2218 www.elsevier.com/locate/ces
Oxygen sorption and desorption properties of Sr–Co–Fe oxide Qinghua Yin, Jay Kniep, Y.S. Lin ∗ Department of Chemical Engineering, Arizona State University, Tempe, AZ 85287-6006, USA Received 10 November 2007; received in revised form 7 January 2008; accepted 12 January 2008 Available online 20 January 2008
Abstract SrCoFeOx has been investigated as a new sorbent for air separation and oxygen removal at high temperatures. X-ray diffraction analysis of a SrCoFeOx sample prepared by liquid citrate method reveals that the sample contains an intergrowth (Sr 4 Fe6−x Cox O13± ), perovskite (SrFe1−x Cox O3− ), and spinel (Co3−x Fex O4 ) phase. Both oxygen vacancies (VO¨ ) and interstitial oxygen ions (Oi ) are involved in the oxygen adsorption and desorption process for SrCoFeOx . Compared with the perovskite-type oxide La0.1 Sr 0.9 Co0.9 Fe0.1 O3− , SrCoFeOx has stronger structure stability in a reducing environment and it also exhibits a larger oxygen sorption capacity at temperatures higher than 800 ◦ C. Meanwhile, unlike La0.1 Sr 0.9 Co0.9 Fe0.1 O3− which shows a fast adsorption rate and a slow desorption rate at 900 ◦ C, SrCoFeOx shows a fast desorption rate and slow adsorption rate at the same temperature. X-ray diffraction data reveals that SrCoFeOx samples sintered at 1140 ◦ C have a higher amount of the intergrowth phase than samples sintered at 950 ◦ C due to slow formation kinetics. X-ray diffraction and thermogravimetric analysis of SrCoFeOx samples prepared by the citrate and solid state method show that the synthesis method strongly influences the amount of the three phases in a sample. 䉷 2008 Elsevier Ltd. All rights reserved. Keywords: Air separation; Mixed-conducting oxides; Defect chemistry; Adsorption
1. Introduction Air separation or oxygen removal from oxygen containing streams at high temperatures is highly desirable for several processes. For example, oxycombustion is a carbon capture technology where the fuel is burned in an oxygen or oxygen enriched carbon dioxide stream rather than air, resulting in a highly pure carbon dioxide exhaust that requires little separation. Barriers for oxycombustion include the high cost associated with oxygen production through cryogenic air separation and energy loss due to large quantities of CO2 exhaust recycle to control combustion temperatures (Klara and Srivastava, 2002; Rodewald et al., 2005). If air can be separated at high temperatures, warm oxygen can be directly used for the oxycombustion providing substantial energy saving for air separation. Recently, Lin and co-workers reported an air separation process with a perovskite-type metal oxide sorbent (Lin et al., 2000; Yang et al., 2002). This process not only separates air at ∗ Corresponding author. Tel.: +1 480 965 7769; fax: +1 480 965 0037.
E-mail address: [email protected] (Y.S. Lin). 0009-2509/$ - see front matter 䉷 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.ces.2008.01.016
high temperatures but offers extremely high selectivity for oxygen. The new sorption separation process takes advantage of the unique properties of these perovskite-type oxides that can adsorb large quantities of oxygen, but not other gases, into the oxygen vacancy sites at high temperatures. The sorption and desorption mechanism is based on the reversible defect reaction involving the gas phase molecular oxygen, oxygen vacancy, lattice oxygen, and electronic hole (Mizusaki et al., 1984, 1989). Infinitely high selectivity for oxygen over nitrogen and high oxygen adsorption amounts are the major characteristics of this group of materials. A series of fundamental studies on perovskite-type oxides as sorbents were performed in our laboratory including: selection and syntheses of materials (Yang et al., 2002; Yang and Lin, 2002), oxygen sorption equilibrium, (Yang et al., 2002; Yang and Lin, 2002, 2003a), oxygen sorption thermal effects, (Yang and Lin, 2005), and oxygen sorption and desorption kinetics as well as fixed-bed process performance (Yang and Lin, 2003b). To be a promising sorbent material for air separation, the perovskite-type oxide must have a high oxygen sorption capacity as well as a fast sorption/desorption
2212
Q. Yin et al. / Chemical Engineering Science 63 (2008) 2211 – 2218
rate. It has been reported that the perovskite-type oxide, La0.1 Sr 0.9 Co0.9 Fe0.1 O3− (LSCF1991), can adsorb oxygen up to 0.6 mmol/g with no sorption for nitrogen. The process has a fast oxygen sorption rate but a rather slow oxygen desorption rate. The desorption rate needs to be increased in order to achieve a high O2 —product purity and to improve sorbent regeneration efficiency. Our group has recently reported several approaches to increase the oxygen desorption rate of LSCF1991 for air separation. One approach is to decrease the crystalline size of the sorbent by decreasing the powder sintering temperature (Yin et al., 2006). We also found that both the oxygen sorption capacity and desorption rate can be improved by taking advantage of the oxygen vacancy order–disorder phase transition for some lanthanum cobaltites (Yin and Lin, 2007). For some perovskitetype ceramics below the phase transition temperature, decreasing the oxygen partial pressure (oxygen desorption) is accompanied with a phase transition from the oxygen vacancy disordered to ordered structure. Such a phase transition significantly enhances the oxygen desorption rate and the reverse oxygen adsorption step yields much higher oxygen sorption capacity as compared to the process without the order–disorder phase transition. The third approach to improve the properties of perovskite-type metal oxides is to dope Ag+ /Ni+ ions in the perovskite-type structure (Yin and Lin, 2006). Doping Ag+ /Ni+ into the LSCF1991 lattice structure increases the tendency of order–disorder phase transition and enhances the surface catalytic properties for oxygen sorption and desorption. These efforts can improve the oxygen sorption properties of perovskite-type oxide LSCF1991 in the temperature range of 600–800 ◦ C. The approaches listed above are based on modifying or taking advantage of the unique properties of the existing perovskitetype metal oxides. As a result, there are limitations in the above approaches with respect to improving the sorption and desorption properties for this group of metal oxide sorbents. For example, enhancement in the desorption rate by the disorder to order phase transition will only take place below the phase transition temperature (typically below 800 ◦ C). Therefore, the logical step to identify sorbents with improved properties is to explore metal oxides of different structure or composition. Ma et al. (1996a, b) reported that SrFeCo0.5 Ox membranes exhibit high oxygen ion conductivity with a stable oxygen permeation flux of 10−6 mol cm−2 s−1 at 900 ◦ C. This material consists of an intergrowth phase (Sr 4 Fe6−x Cox O13± ), a perovskite phase (SrFe1−x Cox O3− ), and a spinel phase (Co3−x Fex O4 ) (Ma et al., 1998; Guggilla and Manthiram, 1997; Fjellvag et al., 1997). This paper reports a study on the oxygen adsorption and desorption of this group of materials with a focus on a new non-perovskite-type material SrCoFeOx (SCF111) that exhibits much improved properties as a sorbent for air separation at temperatures higher than 800 ◦ C. 2. Experimental SrCoz Fey Ox (SCF) with different values of z and y were prepared by the liquid citrate method. Stoichiometric amounts
of the corresponding metal nitrates Sr(NO3 )2 , Fe(NO3 )2 and Co(NO3 )2 were first dissolved by de-ionized water in a beaker and then citric acid (50% excess for the reaction) was added. The resulting solutions were under heating and stirring conditions during the polymerization and condensation reactions, which were carried out at 100–105 and 105–110 ◦ C, respectively. At the end of the condensation reaction, viscous gels were obtained. Self-ignition was performed at approximately 400 ◦ C in air to burn out the organics from the resulting gels after they had been dried at 110 ◦ C for 20 h. Finally, the powders were sintered at 1140 ◦ C (or 950 ◦ C for studying the sintering temperature effect) for 20 h with a ramping rate of 60 ◦ C h−1 . For comparison, perovskite-type oxide LSCF1991 was prepared by the same citrate method by the procedure detailed in a previous publication (Yin and Lin, 2007). SCF oxides were also prepared by the solid state reaction (SSR) method. In this method, stoichiometric amounts of SrCO3 , Fe2 O3 , and Co(NO3 )2 were dissolved into ethanol and ball-milled for 20 h. The obtained mixtures were calcined in air at 950 ◦ C for 16 h. Finally, the resulting calcined powder was ball-milled again for 8 h. The morphologies and particle sizes of the resulting samples were studied by using SEM (Hitachi, S-4000). X-ray diffraction (XRD) (Simens, D50, CuK1) was utilized to characterize the crystal structure of the samples. Oxygen sorption and desorption on the samples were measured by a simultaneous DSC/TGA instrument (TA Instruments, SDT 600). The oxygen sorption/desorption process starts when the sample experiences a sudden increase/decrease in the oxygen activity of the gas stream passing through the sample compartment at a given temperature. For a typical experimental measurement, about 50 mg of a powder sample was placed in the alumina sample holder. Then the sample compartment was heated to 110 ◦ C to remove the water in the sample. The weight of the sample was monitored until a stable value was achieved. The dry sample was then heated to the desired temperature (e.g., 900 ◦ C) in a He atmosphere (PO2 = 0.0001 atm, determined by the impurity of helium) and kept isothermally until a stable weight was observed from the thermogravimetric analysis (TGA) kinetic curves. The feed gas was subsequently quickly switched from He (P o2 = 0.0001 atm) to dry air (P o2 = 0.21 atm) to initiate the oxygen sorption process. The corresponding weight change and the heat effect were automatically recorded by TGA/DSC instrument. The oxygen sorption capacity of the metal oxide was determined from the following equation by TGA measurements: 1 w(air) − w(He) q= , (1) 2MO w(0) where w(air) and w(He) are the equilibrated sample weight at a particular temperature in air and He, respectively. w(0) is the initial oxygen nonstoichiometry and weight of the sample under reference conditions (room temperature and 1 atm air). Mo (=16) is atomic weight of oxygen. In addition, the kinetics and heat effect during the desorption process were also measured for the same sample subjected to a change of the surrounding gas from air to He.
Q. Yin et al. / Chemical Engineering Science 63 (2008) 2211 – 2218
2213
3. Results and discussion
Table 1 The comparisons of the parameters between SCF111 and SrFeCo0.5 Ox
3.1. Structure and oxygen sorption properties of SrCoFeOx
Sample
˚ a (A)
˚ b (A)
˚ c (A)
˚ 3) Cell volume (A
Five SCF samples of different compositions (SrCo0.5 Fe1.25 Ox , SrCo0.5 FeOx , SrCo0.5 Fe0.5 Ox , SrCo0.5 Fe0.75 Ox and SrCoFeOx ) were prepared by the citrate method and sintered at 1140 ◦ C. The oxygen sorption capacities for these samples at different temperatures are compared in Fig. 1. As shown, the oxides with different Fe and Co concentrations show slightly different oxygen sorption capacities below 700 ◦ C. However, at temperatures above 800 ◦ C, SrCoFeOx (SCF111) has the highest oxygen sorption capacities among all the oxides. Therefore, SCF111 oxide was chosen as the candidate material for further study. The XRD patterns of air-sintered SCF111 before and after annealing in helium are shown in Fig. 2. The XRD
SrFeCo0.5 Ox SCF111
11.017 11.13
19.03 19.04
5.5463 5.52
1162.8 1168
0.44
q(mmol/g)
q (mmol/g sorbent)
0.40 0.35 0.30 0.25 0.20
0.36 SrCo0.5Fe0.755Ox
0.32
SrCo0.5Fe0.5Ox
0.28
SrCo0.5FeOx
600
650
700 750 800 Temperatrue(°C) SCF111
850
900
LSCF1991
Fig. 3. Comparison of oxygen sorption capacities between SCF111 and LSCF1991 (solid curves are for visual guide).
0.24 SrCo0.5Fe1.25Ox
0.20
600
700
800 900 Temperature (°C)
1000
Fig. 1. Comparison of oxygen sorption capacities among samples with different Co and Fe concentrations.
20
0.45
SrCoFeOX
0.40
0.16
0.50
30
40
50
60
70
2θ Fig. 2. XRD patterns of air-sintered SCF111 before (top) and after (bottom) annealing in helium.
data was analyzed using Rietveld’s model. The results of analysis show that the sintered SCF111 is similar to SrFeCo0.5 Ox , the material reported by Ma et al. (1996, 1998) and Ma and Balachandran (1997, 1998a, b), in that three distinct phases are present in the sintered sample. The three phases consist of an intergrowth phase (Sr 4 Fe6−x Cox O13± ), a perovskite phase (SrFe1−x Cox O3− ), and a spinel phase (Co3−x Fex O4 ). The intergrowth phase (Sr 4 Fe6−x Cox O13± ) can be described as a layered structure containing alternating perovskitelike and metal oxide blocks (Ma et al., 1998; Guggilla and Manthiram, 1997). SrFeCo0.5 Ox was found to consist of ∼70% Sr 4 Fe6−x Cox O13± , ∼25% SrFe1−x Cox O3− , and ∼5% Co3−x Fex O4 (Ma et al., 1998). The lattice parameters of SCF111 obtained from the Rietveld model fitting are summarized in Table 1. For comparison, the lattice parameters of SrFeCo0.5 Ox oxide reported by Ma et al. (1998) are also listed. Both compositions are orthorhombic and have similar lattice parameters and cell volume. The annealed sample was obtained by quenching the sample after it was exposed to helium at 800 ◦ C for 5 h. As shown, for the helium annealed SCF111 sample, the main XRD peaks are exactly the same with those of the air-sintered sample before helium annealing. This indicates that SCF111 has strong structure stability in a reducing atmosphere at 800 ◦ C. Fig. 3 shows the comparison of the oxygen sorption capacities between SCF111 and La0.1 Sr 0.9 Co0.9 Fe0.1 O3− (LSCF1991), a reference perovskite-type oxide reported in our
2214
Q. Yin et al. / Chemical Engineering Science 63 (2008) 2211 – 2218
1
100.0 SCF111
99.5
-1
98.5 LSCF1991
98.0
Heat Flow (w/g)
Weight Change(%)
0 99.0
-2 97.5 97.0 400 450
500
550
600 650 700
750
800
850 900
-3 950
Temperatrue(°C) Fig. 4. TGA/DSC curves of SCF111 and LSCF1991 in He.
100
Normalized weight change
previous studies, at different temperatures when the flowing gas was switched between air and helium. As shown in the figure, LSCF1991 has a higher oxygen sorption capacity than SCF111 at temperatures below 700 ◦ C. However, at temperatures above 800 ◦ C, the oxygen sorption capacity of SCF111 is much higher than that of LSCF1991. This indicates that the optimum operating temperature of SCF111 is above 800 ◦ C. Fig. 3 also shows a sudden increase of the oxygen sorption capacity at 800 ◦ C for SCF111 samples. This jump in the oxygen sorption capacity is related to the unique structure of SCF111. As mentioned above, the intergrowth phase in SCF111 is defined by alternating perovskite-like and rock-salt-like layers. The rock-salt-like layers are able to adsorb oxygen at high oxygen partial pressures and release oxygen in a reducing environment. The rock-salt-like layers therefore allow oxygen transport by means of the interstitial mechanism (Guggilla and Manthiram, 1997; Ma and Balachandran, 1997). From the chemical reaction equilibrium data with respect to temperature and oxygen partial pressure (Ma et al., 1996a), with an oxygen partial pressure in the range of 0.21 atm to ∼ 0.0001 atm (air and helium), the reaction CO3 O4 ⇔ 3CoO + 21 O2 occurs only at temperatures above 760 ◦ C. Below 760 ◦ C, the rocksalt-like layers in the intergrowth phase cannot adsorb oxygen in air and release oxygen in helium. SCF111 is similar to the perovskite-type oxides at temperatures below 760 ◦ C in that the only mechanism for oxygen sorption is oxygen vacancies. The structural change of SCF111 in helium with increasing temperature was studied by DSC/TGA. The results are shown in Fig. 4. For comparison, the results of LSCF1991 are also presented in the same figure. The TGA results show an initial weight loss of about 1.6 and 2.5 wt% followed by a constant weight in 650–800 and 580–860 ◦ C for SCF111 and LSCF1991, respectively. Both SCF111 and LSCF1991 samples have an endothermic peak at 800 ◦ C and 860 ◦ C, respectively, indicating the phase transition from the ordered brownmillerite to the disordered perovskite structure. As described
SCF111
80
LSCF1991
60
40 700°C 900°C
20
0
0
2000
3000
4000 Time(s)
5000
6000
7000
Fig. 5. The comparison of oxygen sorption kinetics between LSCF1991 and SCF111 at 700 ◦ C and 900 ◦ C.
above, one of the phases of SCF111 in air is a perovskite (SrFe1−x Cox O3− ) phase. In reducing environments at lower temperatures, perovskite-type metal oxides exhibit a brownmillerite structure and transition to a disordered perovskite structure at higher temperatures (Ma and Balachandran, 1998). Figs. 5 and 6 show the comparison of the normalized unsteady state oxygen adsorption and desorption kinetic curves between LSCF1991 and SCF111 at two different temperatures with the surrounding gas switching between air and helium. The change in oxygen partial pressure in the atmosphere surrounding the sample took place in less than 10 s, which is much smaller than the sorption uptake time. Hence the gas phase mixing has a negligible effect on the re-equilibrium kinetics. As shown in Figs. 5 and 6, both SCF111 and LSCF1991 exhibit faster oxygen adsorption and desorption rates at 700 ◦ C
Q. Yin et al. / Chemical Engineering Science 63 (2008) 2211 – 2218
Normalized weight change(%)
100
2215
700°C 900°C
80 LSCF1991 60
40
SCF111
20
0
0
1000
2000 3000 Time(s)
4000
5000
Fig. 6. Oxygen desorption kinetics of LSCF1991 and SCF111 at 700 ◦ C and 900 ◦ C.
than at 900 ◦ C. As reported in our previous study (Yin and Lin, 2007), the oxygen sorption and desorption process on LSCF1991 at 700 ◦ C is companied with oxygen vacancy order–disorder phase transition, which enhances the sorption and desorption kinetics. At 900 ◦ C, LSCF1991 remains in the perovskite structure in air or helium. The perovskite phase in SCF111 might go through a similar order–disorder phase transition at the lower temperature with switching the surrounding gas between air and helium. The oxygen sorption capacity and kinetics of SCF111 are comparable with those of SrCo0.8 Fe0.2 O3− under the same operational conditions (Yin and Lin, 2007), indicating that the perovskite phase in SCF111 may be SrCo0.8 Fe0.2 O3− . The desorption rate for LSCF1991 is slower than the sorption rate. This is because the oxygen sorption/desorption rate constant decreases with decreasing oxygen partial pressure at the equilibrium stage for perovskite-type oxides (Yin et al., 2006). However, unlike the perovskite-type LSCF1991, SCF111 shows a faster desorption rate than adsorption rate at 900 ◦ C. For oxygen transport into or out of the small SCF111 particles, surface charge transfer reactions are more likely to be the rate-limiting step. For SCF111, the surface charge transfer reactions may be represented by the following two simplified reactions for the perovskite phase and rock-salt-like layers of the intergrowth phase, respectively: 1 2 O2 1 2 O2
+ VO¨ ⇔ OxO + 2h, ⇔ Oi + 2h.
(2) (3)
It is possible that the rock-salt-like layers in the intergrowth phase have a much faster surface charge transfer reaction rate in oxygen sorption and desorption steps. More studies are needed to understand the detailed mechanism of oxygen transport in SCF111.
Fig. 7. SEM images of Sample 1 (A), Sample 2 (B), and Sample 3 (C) with different preparation methods and sintering temperatures.
3.2. Effects of synthesis methods on oxygen sorption properties of SrCoFeOx The SCF111 powders obtained by the SSR method (with a sintering temperature of 950 ◦ C) and the liquid citrate method (with the sintering temperatures of 950 ◦ C and 1140 ◦ C) are named as Sample 1, Sample 2, and Sample 3, respectively.
2216
Q. Yin et al. / Chemical Engineering Science 63 (2008) 2211 – 2218
100
layered-perovskite Perovskite-SrCo Fe O (Co,Fe)O *
99
Sample 3
Intensity
Weight(wt%)
**
*
Sample 2
**
*
*
Sample 3 98
97 Sample 2 96 Sample 1
Sample 1
*
*
* 20
30
40
50
95 100
*
60
200
300
70
400 500 600 700 Temperature(°C)
800
900 1000
2θ
2.5 0.7
Sample 1 2.0
0.6 Sample 2 0.5
1.5
0.4 Sample 3
1.0
0.5 600
0.3
650
700
750 800 Temperature(°C)
850
900
0.2 950
Oxygen sorption capacity (mmol/g)
Sample 3 is the same as the sample described in Section 3.1. Sample 1 exhibits an irregular morphology and nearly uniform size distribution. Fig. 7 shows SEM micrographs of the three samples. The average aggregate size of Sample 1 is less than 500 nm as measured from the SEM image. Samples 2 and 3, prepared by using the same method but at different sintering temperatures, exhibit two different shapes (the roundshape and needle-shape). The average grain sizes of Samples 2 and 3 are 1 and 10 m, respectively. Energy-dispersive X-ray (EDX) analysis reveals the same composition for both Samples 2 and 3. The overall atomic ratio of the metal elements determined by EDX is consistent with that given in the chemical formula. The room temperature XRD patterns of Samples 2 and 3 are compared in Fig. 8 with that of Sample 1. It can be seen that, by changing the preparation method and sintering temperatures, SCF111 exhibits different fractions of the three phases. Sample 3 has a higher percentage of the intergrowth phase than either Sample 1 or 2. Xia et al. (2000) showed that the perovskite phase forms first and then transforms to the intergrowth phase in three phase materials but the kinetics are slower at lower temperatures. Since the samples were sintered at their respective temperatures for roughly the same amount of time, the intergrowth layer is more prominent in Sample 3 because it was able to form faster due to the higher sintering temperature. Samples 1 and 2 were both sintered at 950 ◦ C, but the XRD patterns are different which indicates that the preparation method also has an influence on the composition. For example, Sample 2 shows a smaller amount of the spinel and perovskite phase in the mixture when compared with Sample 1. Fig. 9 shows the TGA plots of the samples heating in a helium atmosphere. The weight losses for the three samples decrease in the order of Sample 1 > Sample 2 > Sample 3. The results reveal that the oxygen loss amount increases with an increasing perovskite phase fraction in the samples. The perovskite phase loses a substantially larger amount of oxygen at higher temperatures compared to the intergrowth phase
Fig. 9. TGA curves of Samples 1–3 heating in He with a ramping rate of 10 ◦ C/min.
Weight Change%
Fig. 8. XRD patterns of Samples 1–3 after sintering in air.
Fig. 10. Comparison of oxygen sorption capacities among Samples 1–3 at different temperatures.
(Xia et al., 2000). The TGA results are consistent to the phase relationships derived from the XRD results in Fig. 8. Fig. 10 shows the comparison of weight changes and oxygen sorption capacities among the three samples at different temperatures when the flowing gas was switched between air and helium in the TGA measurements. Sample 1 yields a larger weight change and oxygen sorption capacity than Samples 2 and 3 at temperatures above 700 ◦ C. For example, the weight change of Sample 1 in the oxygen sorption process is around 2.4%. This is much higher than those of Samples 2 and 3, which are 1.5% and 1.3%, respectively at the same operation conditions. Sample 1 exhibits a much higher oxygen sorption capacity than Samples 2 and 3 due to difference in the composition among the three samples. Armstrong et al. (2000) studied the oxygen permeation flux as a function of PO2 gradient across the sample for single-phase Sr 4 Fe4 Co2 O13+ (consisting only of the intergrowth phase) and three phase Sr 4 Fe4 Co2 O13+ membranes. They found the three phase Sr 4 Fe4 Co2 O13+ membranes permeate oxygen one order of magnitude higher
Q. Yin et al. / Chemical Engineering Science 63 (2008) 2211 – 2218
desorption kinetics, and faster adsorption kinetics in comparison with the material with only the layered structures. It is possible to obtain SCF111 with excellent oxygen sorption properties by controlling the perovskite phase concentration in the three phase mixture.
Normalized weight change(%)
100
80 Sample 1 60
Sample 2 Sample 3
Acknowledgment
40
The authors acknowledge the support of the NSF (CTS0132694) on this project.
Sorption-900°C Desorption-900°C
20
0
0
1000
2000
3000
2217
4000
5000
6000
Time(s)
Fig. 11. Adsorption and desorption kinetics among Samples 1–3.
than the single-phase Sr 4 Fe4 Co2 O13+ membranes. This is due to the presence of the perovskite phase permeating oxygen two orders of magnitude higher than the intergrowth phase. The same reasoning applies here. In Sample 1, it is the perovskite phase that is dominant in the adsorption process. However, in Samples 2 and 3, it is the intergrowth phase that dominates in the sorption process. Fig. 11 shows the adsorption (a) and desorption (b) kinetics of the three samples at 900 ◦ C. As shown from Fig. 11(a), the adsorption rate for the samples at 900 ◦ C decreases in the order: Sample 1 > Sample 2 > Sample 3. The same trend of adsorption kinetics was observed for these samples at 800 ◦ C. However, the desorption rate for the samples at 900 ◦ C decreases in the reverse order: Sample 3 > Sample 2 > Sample 1. 4. Conclusions SrCoFeOx is reported as an alternate sorbent to LSCF1991 for high temperature air separation at temperatures greater than 800 ◦ C. X-ray diffraction results reveal that SCF111 is a composition that contains an intergrowth, perovskite, and spinel phase and it has strong structure stability in a reducing atmosphere at high temperatures. Compared with LSCF1991, SCF111 exhibits a higher oxygen sorption capacity, faster oxygen desorption rate, and slower oxygen sorption rate at temperatures above 800 ◦ C. Both oxygen vacancies (VO¨ ) and interstitial oxygen ions (Oi ) are involved in the oxygen adsorption and desorption process for SrCoFeOx . Samples of SCF111 sintered at 1140 ◦ C had a higher percentage of the intergrowth phase than samples sintered at 950 ◦ C due to the slower formation of the intergrowth phase at lower temperatures. The synthesis method also has a strong influence on the amount of the three phases as samples prepared by the citrate method had a larger amount of the perovskite phase than samples prepared by the solid state method. The perovskite phase plays an important role in the oxygen sorption property. With a higher concentration of the perovskite phase in the mixture, the corresponding material shows a higher oxygen sorption capacity, slower
References Armstrong, T., Prado, F., Xia, Y., Manthiram, A., 2000. Role of perovskite phase on the oxygen permeation properties of the Sr 4 Fe6−x Cox O13+ system. Journal of Electrochemical Society 147 (2), 435–438. Fjellvag, H., Hauback, B.C., Bredesen, R., 1997. Crystal structure of the mixed conductor Sr 4 Fe4 Co2 O13 . Journal of Material Chemistry 7 (12), 2415–2419. Guggilla, S., Manthiram, A., 1997. Crystal chemical characterization of the mixed conductor Sr(Fe, Co)1.5 Oy exhibiting unusually high oxygen permeability. Journal of Electrochemical Society 144 (5), L120–L122. Klara, S.M., Srivastava, R.D., 2002. US DOE integrated collaborative technology development program for CO2 separation and capture. Environmental Progress 21 (4), 247–253. Lin, Y.S., Mclean, D.L., Zeng, Y., 2000. High temperature adsorption process, US patent, 6,059,858. Ma, B., Balachandran, U., 1997. Oxygen nonstoichiometry in mixedconducting SrFeCo0.5 Ox . Solid State Ionics 100, 53–62. Ma, B., Balachandran, U., 1998a. Phase stability of SrFeCo0.5 Ox in reducing environments. Materials Research Bulletin 33 (2), 223–236. Ma, B., Balachandran, U., 1998b. Transport properties of the mixedconducting SrFe1.2 Co0.3 Ox . Journal of Electroceramics 2 (2), 135–142. Ma, B., Balachandran, U., Park, J.H., 1996a. Electrical transport properties and defect structure of SrFeCo0.5 Ox . Journal of Electrochemical Society 143 (5), 1736–1744. Ma, B., Balachandran, U., Park, J.H., Segre, C.U., 1996b. Determination of chemical diffusion coefficient of SrFeCo0.5 Ox by the conductivity relaxation method. Solid State Ionics 83, 65–71. Ma, B., Hodges, J.P., Jorgensen, J.D., Miller, D.J., Richardson, J.W., Balachandran, U., 1998. Structure and properties relationship in mixedconducting Sr 4 (Fe1−x Cox )O13± materials. Journal of Solid State Chemistry 141, 576–586. Mizusaki, J., Yamauchi, S., Fueki, K., Ishikawa, A., 1984. Nonstoichiometry of perovskite-type oxides La1−x Sr x CoO3− . Solid State Ionics 12, 119–124. Mizusaki, J., Mima, Y., Yamauchi, S., Fueki, K., Tagawa, H., 1989. Nonstoichiometry of the perovskite-type oxides La1−x Sr x CoO3− . Journal of Solid State Chemistry 80, 102–111. Rodewald, A., Kather, A.A., Frie, S., 2005. Thermodynamic and economic aspects of the hard coal based oxyfuel cycle. International Journal of Green Energy 2 (2), 181–192. Xia, Y., Armstrong, T., Prado, F., Manthiram, A., 2000. Sol–gel synthesis, phase relationships, and oxygen permeation properties of Sr 4 Fe6−x Cox O13+ (0 x 3). Solid State Ionics 130, 81–90. Yang, Z.H., Lin, Y.S., 2002. A semi-empirical equation for oxygen nonstoichiometry of perovskite-type ceramics. Solid State Ionics 150, 245–254. Yang, Z.H., Lin, Y.S., 2003a. High-temperature oxygen sorption in a fixed bed packed with perovskite-type ceramic sorbents. Industrial and Engineering Chemistry Research 42, 4376–4381. Yang, Z.H., Lin, Y.S., 2003b. Equilibrium of oxygen sorption on perovskitetype lanthanum cobaltite sorbent. A.I.Ch.E. Journal 49, 793–798. Yang, Z.H., Lin, Y.S., 2005. Synergetic thermal effects for oxygen sorption and order–disorder transition on perovskite-type oxides. Solid State Ionics 176, 89–96.
2218
Q. Yin et al. / Chemical Engineering Science 63 (2008) 2211 – 2218
Yang, Z.H., Lin, Y.S., Zeng, Y., 2002. High temperature sorption process for air separation and oxygen removal. Industrial and Engineering Chemistry Research 41, 2775–2784. Yin, Q.H., Yang, Z.H., Lin, Y.S., 2006. Effects of microstructure on oxygen transport in perovskite-type oxides. Journal of Material Science 41, 4865–4870.
Yin, Q.H., Lin, Y.S., 2006. Effect of dopant addition on oxygen sorption properties of La–Sr–Co–Fe–O perovskite type oxide. Adsorption 12, 329–338. Yin, Q.H., Lin, Y.S., 2007. Beneficial effect of order–disorder phase transition on oxygen sorption properties of perovskite-type oxides. Solid State Ionics 178, 83–89.
Oxygen sorption and desorption properties of Sr–Co–Fe oxide Qinghua Yin, Jay Kniep, Y.S. Lin ∗ Department of Chemical Engineering, Arizona State University, Tempe, AZ 85287-6006, USA Received 10 November 2007; received in revised form 7 January 2008; accepted 12 January 2008 Available online 20 January 2008
Abstract SrCoFeOx has been investigated as a new sorbent for air separation and oxygen removal at high temperatures. X-ray diffraction analysis of a SrCoFeOx sample prepared by liquid citrate method reveals that the sample contains an intergrowth (Sr 4 Fe6−x Cox O13± ), perovskite (SrFe1−x Cox O3− ), and spinel (Co3−x Fex O4 ) phase. Both oxygen vacancies (VO¨ ) and interstitial oxygen ions (Oi ) are involved in the oxygen adsorption and desorption process for SrCoFeOx . Compared with the perovskite-type oxide La0.1 Sr 0.9 Co0.9 Fe0.1 O3− , SrCoFeOx has stronger structure stability in a reducing environment and it also exhibits a larger oxygen sorption capacity at temperatures higher than 800 ◦ C. Meanwhile, unlike La0.1 Sr 0.9 Co0.9 Fe0.1 O3− which shows a fast adsorption rate and a slow desorption rate at 900 ◦ C, SrCoFeOx shows a fast desorption rate and slow adsorption rate at the same temperature. X-ray diffraction data reveals that SrCoFeOx samples sintered at 1140 ◦ C have a higher amount of the intergrowth phase than samples sintered at 950 ◦ C due to slow formation kinetics. X-ray diffraction and thermogravimetric analysis of SrCoFeOx samples prepared by the citrate and solid state method show that the synthesis method strongly influences the amount of the three phases in a sample. 䉷 2008 Elsevier Ltd. All rights reserved. Keywords: Air separation; Mixed-conducting oxides; Defect chemistry; Adsorption
1. Introduction Air separation or oxygen removal from oxygen containing streams at high temperatures is highly desirable for several processes. For example, oxycombustion is a carbon capture technology where the fuel is burned in an oxygen or oxygen enriched carbon dioxide stream rather than air, resulting in a highly pure carbon dioxide exhaust that requires little separation. Barriers for oxycombustion include the high cost associated with oxygen production through cryogenic air separation and energy loss due to large quantities of CO2 exhaust recycle to control combustion temperatures (Klara and Srivastava, 2002; Rodewald et al., 2005). If air can be separated at high temperatures, warm oxygen can be directly used for the oxycombustion providing substantial energy saving for air separation. Recently, Lin and co-workers reported an air separation process with a perovskite-type metal oxide sorbent (Lin et al., 2000; Yang et al., 2002). This process not only separates air at ∗ Corresponding author. Tel.: +1 480 965 7769; fax: +1 480 965 0037.
E-mail address: [email protected] (Y.S. Lin). 0009-2509/$ - see front matter 䉷 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.ces.2008.01.016
high temperatures but offers extremely high selectivity for oxygen. The new sorption separation process takes advantage of the unique properties of these perovskite-type oxides that can adsorb large quantities of oxygen, but not other gases, into the oxygen vacancy sites at high temperatures. The sorption and desorption mechanism is based on the reversible defect reaction involving the gas phase molecular oxygen, oxygen vacancy, lattice oxygen, and electronic hole (Mizusaki et al., 1984, 1989). Infinitely high selectivity for oxygen over nitrogen and high oxygen adsorption amounts are the major characteristics of this group of materials. A series of fundamental studies on perovskite-type oxides as sorbents were performed in our laboratory including: selection and syntheses of materials (Yang et al., 2002; Yang and Lin, 2002), oxygen sorption equilibrium, (Yang et al., 2002; Yang and Lin, 2002, 2003a), oxygen sorption thermal effects, (Yang and Lin, 2005), and oxygen sorption and desorption kinetics as well as fixed-bed process performance (Yang and Lin, 2003b). To be a promising sorbent material for air separation, the perovskite-type oxide must have a high oxygen sorption capacity as well as a fast sorption/desorption
2212
Q. Yin et al. / Chemical Engineering Science 63 (2008) 2211 – 2218
rate. It has been reported that the perovskite-type oxide, La0.1 Sr 0.9 Co0.9 Fe0.1 O3− (LSCF1991), can adsorb oxygen up to 0.6 mmol/g with no sorption for nitrogen. The process has a fast oxygen sorption rate but a rather slow oxygen desorption rate. The desorption rate needs to be increased in order to achieve a high O2 —product purity and to improve sorbent regeneration efficiency. Our group has recently reported several approaches to increase the oxygen desorption rate of LSCF1991 for air separation. One approach is to decrease the crystalline size of the sorbent by decreasing the powder sintering temperature (Yin et al., 2006). We also found that both the oxygen sorption capacity and desorption rate can be improved by taking advantage of the oxygen vacancy order–disorder phase transition for some lanthanum cobaltites (Yin and Lin, 2007). For some perovskitetype ceramics below the phase transition temperature, decreasing the oxygen partial pressure (oxygen desorption) is accompanied with a phase transition from the oxygen vacancy disordered to ordered structure. Such a phase transition significantly enhances the oxygen desorption rate and the reverse oxygen adsorption step yields much higher oxygen sorption capacity as compared to the process without the order–disorder phase transition. The third approach to improve the properties of perovskite-type metal oxides is to dope Ag+ /Ni+ ions in the perovskite-type structure (Yin and Lin, 2006). Doping Ag+ /Ni+ into the LSCF1991 lattice structure increases the tendency of order–disorder phase transition and enhances the surface catalytic properties for oxygen sorption and desorption. These efforts can improve the oxygen sorption properties of perovskite-type oxide LSCF1991 in the temperature range of 600–800 ◦ C. The approaches listed above are based on modifying or taking advantage of the unique properties of the existing perovskitetype metal oxides. As a result, there are limitations in the above approaches with respect to improving the sorption and desorption properties for this group of metal oxide sorbents. For example, enhancement in the desorption rate by the disorder to order phase transition will only take place below the phase transition temperature (typically below 800 ◦ C). Therefore, the logical step to identify sorbents with improved properties is to explore metal oxides of different structure or composition. Ma et al. (1996a, b) reported that SrFeCo0.5 Ox membranes exhibit high oxygen ion conductivity with a stable oxygen permeation flux of 10−6 mol cm−2 s−1 at 900 ◦ C. This material consists of an intergrowth phase (Sr 4 Fe6−x Cox O13± ), a perovskite phase (SrFe1−x Cox O3− ), and a spinel phase (Co3−x Fex O4 ) (Ma et al., 1998; Guggilla and Manthiram, 1997; Fjellvag et al., 1997). This paper reports a study on the oxygen adsorption and desorption of this group of materials with a focus on a new non-perovskite-type material SrCoFeOx (SCF111) that exhibits much improved properties as a sorbent for air separation at temperatures higher than 800 ◦ C. 2. Experimental SrCoz Fey Ox (SCF) with different values of z and y were prepared by the liquid citrate method. Stoichiometric amounts
of the corresponding metal nitrates Sr(NO3 )2 , Fe(NO3 )2 and Co(NO3 )2 were first dissolved by de-ionized water in a beaker and then citric acid (50% excess for the reaction) was added. The resulting solutions were under heating and stirring conditions during the polymerization and condensation reactions, which were carried out at 100–105 and 105–110 ◦ C, respectively. At the end of the condensation reaction, viscous gels were obtained. Self-ignition was performed at approximately 400 ◦ C in air to burn out the organics from the resulting gels after they had been dried at 110 ◦ C for 20 h. Finally, the powders were sintered at 1140 ◦ C (or 950 ◦ C for studying the sintering temperature effect) for 20 h with a ramping rate of 60 ◦ C h−1 . For comparison, perovskite-type oxide LSCF1991 was prepared by the same citrate method by the procedure detailed in a previous publication (Yin and Lin, 2007). SCF oxides were also prepared by the solid state reaction (SSR) method. In this method, stoichiometric amounts of SrCO3 , Fe2 O3 , and Co(NO3 )2 were dissolved into ethanol and ball-milled for 20 h. The obtained mixtures were calcined in air at 950 ◦ C for 16 h. Finally, the resulting calcined powder was ball-milled again for 8 h. The morphologies and particle sizes of the resulting samples were studied by using SEM (Hitachi, S-4000). X-ray diffraction (XRD) (Simens, D50, CuK1) was utilized to characterize the crystal structure of the samples. Oxygen sorption and desorption on the samples were measured by a simultaneous DSC/TGA instrument (TA Instruments, SDT 600). The oxygen sorption/desorption process starts when the sample experiences a sudden increase/decrease in the oxygen activity of the gas stream passing through the sample compartment at a given temperature. For a typical experimental measurement, about 50 mg of a powder sample was placed in the alumina sample holder. Then the sample compartment was heated to 110 ◦ C to remove the water in the sample. The weight of the sample was monitored until a stable value was achieved. The dry sample was then heated to the desired temperature (e.g., 900 ◦ C) in a He atmosphere (PO2 = 0.0001 atm, determined by the impurity of helium) and kept isothermally until a stable weight was observed from the thermogravimetric analysis (TGA) kinetic curves. The feed gas was subsequently quickly switched from He (P o2 = 0.0001 atm) to dry air (P o2 = 0.21 atm) to initiate the oxygen sorption process. The corresponding weight change and the heat effect were automatically recorded by TGA/DSC instrument. The oxygen sorption capacity of the metal oxide was determined from the following equation by TGA measurements: 1 w(air) − w(He) q= , (1) 2MO w(0) where w(air) and w(He) are the equilibrated sample weight at a particular temperature in air and He, respectively. w(0) is the initial oxygen nonstoichiometry and weight of the sample under reference conditions (room temperature and 1 atm air). Mo (=16) is atomic weight of oxygen. In addition, the kinetics and heat effect during the desorption process were also measured for the same sample subjected to a change of the surrounding gas from air to He.
Q. Yin et al. / Chemical Engineering Science 63 (2008) 2211 – 2218
2213
3. Results and discussion
Table 1 The comparisons of the parameters between SCF111 and SrFeCo0.5 Ox
3.1. Structure and oxygen sorption properties of SrCoFeOx
Sample
˚ a (A)
˚ b (A)
˚ c (A)
˚ 3) Cell volume (A
Five SCF samples of different compositions (SrCo0.5 Fe1.25 Ox , SrCo0.5 FeOx , SrCo0.5 Fe0.5 Ox , SrCo0.5 Fe0.75 Ox and SrCoFeOx ) were prepared by the citrate method and sintered at 1140 ◦ C. The oxygen sorption capacities for these samples at different temperatures are compared in Fig. 1. As shown, the oxides with different Fe and Co concentrations show slightly different oxygen sorption capacities below 700 ◦ C. However, at temperatures above 800 ◦ C, SrCoFeOx (SCF111) has the highest oxygen sorption capacities among all the oxides. Therefore, SCF111 oxide was chosen as the candidate material for further study. The XRD patterns of air-sintered SCF111 before and after annealing in helium are shown in Fig. 2. The XRD
SrFeCo0.5 Ox SCF111
11.017 11.13
19.03 19.04
5.5463 5.52
1162.8 1168
0.44
q(mmol/g)
q (mmol/g sorbent)
0.40 0.35 0.30 0.25 0.20
0.36 SrCo0.5Fe0.755Ox
0.32
SrCo0.5Fe0.5Ox
0.28
SrCo0.5FeOx
600
650
700 750 800 Temperatrue(°C) SCF111
850
900
LSCF1991
Fig. 3. Comparison of oxygen sorption capacities between SCF111 and LSCF1991 (solid curves are for visual guide).
0.24 SrCo0.5Fe1.25Ox
0.20
600
700
800 900 Temperature (°C)
1000
Fig. 1. Comparison of oxygen sorption capacities among samples with different Co and Fe concentrations.
20
0.45
SrCoFeOX
0.40
0.16
0.50
30
40
50
60
70
2θ Fig. 2. XRD patterns of air-sintered SCF111 before (top) and after (bottom) annealing in helium.
data was analyzed using Rietveld’s model. The results of analysis show that the sintered SCF111 is similar to SrFeCo0.5 Ox , the material reported by Ma et al. (1996, 1998) and Ma and Balachandran (1997, 1998a, b), in that three distinct phases are present in the sintered sample. The three phases consist of an intergrowth phase (Sr 4 Fe6−x Cox O13± ), a perovskite phase (SrFe1−x Cox O3− ), and a spinel phase (Co3−x Fex O4 ). The intergrowth phase (Sr 4 Fe6−x Cox O13± ) can be described as a layered structure containing alternating perovskitelike and metal oxide blocks (Ma et al., 1998; Guggilla and Manthiram, 1997). SrFeCo0.5 Ox was found to consist of ∼70% Sr 4 Fe6−x Cox O13± , ∼25% SrFe1−x Cox O3− , and ∼5% Co3−x Fex O4 (Ma et al., 1998). The lattice parameters of SCF111 obtained from the Rietveld model fitting are summarized in Table 1. For comparison, the lattice parameters of SrFeCo0.5 Ox oxide reported by Ma et al. (1998) are also listed. Both compositions are orthorhombic and have similar lattice parameters and cell volume. The annealed sample was obtained by quenching the sample after it was exposed to helium at 800 ◦ C for 5 h. As shown, for the helium annealed SCF111 sample, the main XRD peaks are exactly the same with those of the air-sintered sample before helium annealing. This indicates that SCF111 has strong structure stability in a reducing atmosphere at 800 ◦ C. Fig. 3 shows the comparison of the oxygen sorption capacities between SCF111 and La0.1 Sr 0.9 Co0.9 Fe0.1 O3− (LSCF1991), a reference perovskite-type oxide reported in our
2214
Q. Yin et al. / Chemical Engineering Science 63 (2008) 2211 – 2218
1
100.0 SCF111
99.5
-1
98.5 LSCF1991
98.0
Heat Flow (w/g)
Weight Change(%)
0 99.0
-2 97.5 97.0 400 450
500
550
600 650 700
750
800
850 900
-3 950
Temperatrue(°C) Fig. 4. TGA/DSC curves of SCF111 and LSCF1991 in He.
100
Normalized weight change
previous studies, at different temperatures when the flowing gas was switched between air and helium. As shown in the figure, LSCF1991 has a higher oxygen sorption capacity than SCF111 at temperatures below 700 ◦ C. However, at temperatures above 800 ◦ C, the oxygen sorption capacity of SCF111 is much higher than that of LSCF1991. This indicates that the optimum operating temperature of SCF111 is above 800 ◦ C. Fig. 3 also shows a sudden increase of the oxygen sorption capacity at 800 ◦ C for SCF111 samples. This jump in the oxygen sorption capacity is related to the unique structure of SCF111. As mentioned above, the intergrowth phase in SCF111 is defined by alternating perovskite-like and rock-salt-like layers. The rock-salt-like layers are able to adsorb oxygen at high oxygen partial pressures and release oxygen in a reducing environment. The rock-salt-like layers therefore allow oxygen transport by means of the interstitial mechanism (Guggilla and Manthiram, 1997; Ma and Balachandran, 1997). From the chemical reaction equilibrium data with respect to temperature and oxygen partial pressure (Ma et al., 1996a), with an oxygen partial pressure in the range of 0.21 atm to ∼ 0.0001 atm (air and helium), the reaction CO3 O4 ⇔ 3CoO + 21 O2 occurs only at temperatures above 760 ◦ C. Below 760 ◦ C, the rocksalt-like layers in the intergrowth phase cannot adsorb oxygen in air and release oxygen in helium. SCF111 is similar to the perovskite-type oxides at temperatures below 760 ◦ C in that the only mechanism for oxygen sorption is oxygen vacancies. The structural change of SCF111 in helium with increasing temperature was studied by DSC/TGA. The results are shown in Fig. 4. For comparison, the results of LSCF1991 are also presented in the same figure. The TGA results show an initial weight loss of about 1.6 and 2.5 wt% followed by a constant weight in 650–800 and 580–860 ◦ C for SCF111 and LSCF1991, respectively. Both SCF111 and LSCF1991 samples have an endothermic peak at 800 ◦ C and 860 ◦ C, respectively, indicating the phase transition from the ordered brownmillerite to the disordered perovskite structure. As described
SCF111
80
LSCF1991
60
40 700°C 900°C
20
0
0
2000
3000
4000 Time(s)
5000
6000
7000
Fig. 5. The comparison of oxygen sorption kinetics between LSCF1991 and SCF111 at 700 ◦ C and 900 ◦ C.
above, one of the phases of SCF111 in air is a perovskite (SrFe1−x Cox O3− ) phase. In reducing environments at lower temperatures, perovskite-type metal oxides exhibit a brownmillerite structure and transition to a disordered perovskite structure at higher temperatures (Ma and Balachandran, 1998). Figs. 5 and 6 show the comparison of the normalized unsteady state oxygen adsorption and desorption kinetic curves between LSCF1991 and SCF111 at two different temperatures with the surrounding gas switching between air and helium. The change in oxygen partial pressure in the atmosphere surrounding the sample took place in less than 10 s, which is much smaller than the sorption uptake time. Hence the gas phase mixing has a negligible effect on the re-equilibrium kinetics. As shown in Figs. 5 and 6, both SCF111 and LSCF1991 exhibit faster oxygen adsorption and desorption rates at 700 ◦ C
Q. Yin et al. / Chemical Engineering Science 63 (2008) 2211 – 2218
Normalized weight change(%)
100
2215
700°C 900°C
80 LSCF1991 60
40
SCF111
20
0
0
1000
2000 3000 Time(s)
4000
5000
Fig. 6. Oxygen desorption kinetics of LSCF1991 and SCF111 at 700 ◦ C and 900 ◦ C.
than at 900 ◦ C. As reported in our previous study (Yin and Lin, 2007), the oxygen sorption and desorption process on LSCF1991 at 700 ◦ C is companied with oxygen vacancy order–disorder phase transition, which enhances the sorption and desorption kinetics. At 900 ◦ C, LSCF1991 remains in the perovskite structure in air or helium. The perovskite phase in SCF111 might go through a similar order–disorder phase transition at the lower temperature with switching the surrounding gas between air and helium. The oxygen sorption capacity and kinetics of SCF111 are comparable with those of SrCo0.8 Fe0.2 O3− under the same operational conditions (Yin and Lin, 2007), indicating that the perovskite phase in SCF111 may be SrCo0.8 Fe0.2 O3− . The desorption rate for LSCF1991 is slower than the sorption rate. This is because the oxygen sorption/desorption rate constant decreases with decreasing oxygen partial pressure at the equilibrium stage for perovskite-type oxides (Yin et al., 2006). However, unlike the perovskite-type LSCF1991, SCF111 shows a faster desorption rate than adsorption rate at 900 ◦ C. For oxygen transport into or out of the small SCF111 particles, surface charge transfer reactions are more likely to be the rate-limiting step. For SCF111, the surface charge transfer reactions may be represented by the following two simplified reactions for the perovskite phase and rock-salt-like layers of the intergrowth phase, respectively: 1 2 O2 1 2 O2
+ VO¨ ⇔ OxO + 2h, ⇔ Oi + 2h.
(2) (3)
It is possible that the rock-salt-like layers in the intergrowth phase have a much faster surface charge transfer reaction rate in oxygen sorption and desorption steps. More studies are needed to understand the detailed mechanism of oxygen transport in SCF111.
Fig. 7. SEM images of Sample 1 (A), Sample 2 (B), and Sample 3 (C) with different preparation methods and sintering temperatures.
3.2. Effects of synthesis methods on oxygen sorption properties of SrCoFeOx The SCF111 powders obtained by the SSR method (with a sintering temperature of 950 ◦ C) and the liquid citrate method (with the sintering temperatures of 950 ◦ C and 1140 ◦ C) are named as Sample 1, Sample 2, and Sample 3, respectively.
2216
Q. Yin et al. / Chemical Engineering Science 63 (2008) 2211 – 2218
100
layered-perovskite Perovskite-SrCo Fe O (Co,Fe)O *
99
Sample 3
Intensity
Weight(wt%)
**
*
Sample 2
**
*
*
Sample 3 98
97 Sample 2 96 Sample 1
Sample 1
*
*
* 20
30
40
50
95 100
*
60
200
300
70
400 500 600 700 Temperature(°C)
800
900 1000
2θ
2.5 0.7
Sample 1 2.0
0.6 Sample 2 0.5
1.5
0.4 Sample 3
1.0
0.5 600
0.3
650
700
750 800 Temperature(°C)
850
900
0.2 950
Oxygen sorption capacity (mmol/g)
Sample 3 is the same as the sample described in Section 3.1. Sample 1 exhibits an irregular morphology and nearly uniform size distribution. Fig. 7 shows SEM micrographs of the three samples. The average aggregate size of Sample 1 is less than 500 nm as measured from the SEM image. Samples 2 and 3, prepared by using the same method but at different sintering temperatures, exhibit two different shapes (the roundshape and needle-shape). The average grain sizes of Samples 2 and 3 are 1 and 10 m, respectively. Energy-dispersive X-ray (EDX) analysis reveals the same composition for both Samples 2 and 3. The overall atomic ratio of the metal elements determined by EDX is consistent with that given in the chemical formula. The room temperature XRD patterns of Samples 2 and 3 are compared in Fig. 8 with that of Sample 1. It can be seen that, by changing the preparation method and sintering temperatures, SCF111 exhibits different fractions of the three phases. Sample 3 has a higher percentage of the intergrowth phase than either Sample 1 or 2. Xia et al. (2000) showed that the perovskite phase forms first and then transforms to the intergrowth phase in three phase materials but the kinetics are slower at lower temperatures. Since the samples were sintered at their respective temperatures for roughly the same amount of time, the intergrowth layer is more prominent in Sample 3 because it was able to form faster due to the higher sintering temperature. Samples 1 and 2 were both sintered at 950 ◦ C, but the XRD patterns are different which indicates that the preparation method also has an influence on the composition. For example, Sample 2 shows a smaller amount of the spinel and perovskite phase in the mixture when compared with Sample 1. Fig. 9 shows the TGA plots of the samples heating in a helium atmosphere. The weight losses for the three samples decrease in the order of Sample 1 > Sample 2 > Sample 3. The results reveal that the oxygen loss amount increases with an increasing perovskite phase fraction in the samples. The perovskite phase loses a substantially larger amount of oxygen at higher temperatures compared to the intergrowth phase
Fig. 9. TGA curves of Samples 1–3 heating in He with a ramping rate of 10 ◦ C/min.
Weight Change%
Fig. 8. XRD patterns of Samples 1–3 after sintering in air.
Fig. 10. Comparison of oxygen sorption capacities among Samples 1–3 at different temperatures.
(Xia et al., 2000). The TGA results are consistent to the phase relationships derived from the XRD results in Fig. 8. Fig. 10 shows the comparison of weight changes and oxygen sorption capacities among the three samples at different temperatures when the flowing gas was switched between air and helium in the TGA measurements. Sample 1 yields a larger weight change and oxygen sorption capacity than Samples 2 and 3 at temperatures above 700 ◦ C. For example, the weight change of Sample 1 in the oxygen sorption process is around 2.4%. This is much higher than those of Samples 2 and 3, which are 1.5% and 1.3%, respectively at the same operation conditions. Sample 1 exhibits a much higher oxygen sorption capacity than Samples 2 and 3 due to difference in the composition among the three samples. Armstrong et al. (2000) studied the oxygen permeation flux as a function of PO2 gradient across the sample for single-phase Sr 4 Fe4 Co2 O13+ (consisting only of the intergrowth phase) and three phase Sr 4 Fe4 Co2 O13+ membranes. They found the three phase Sr 4 Fe4 Co2 O13+ membranes permeate oxygen one order of magnitude higher
Q. Yin et al. / Chemical Engineering Science 63 (2008) 2211 – 2218
desorption kinetics, and faster adsorption kinetics in comparison with the material with only the layered structures. It is possible to obtain SCF111 with excellent oxygen sorption properties by controlling the perovskite phase concentration in the three phase mixture.
Normalized weight change(%)
100
80 Sample 1 60
Sample 2 Sample 3
Acknowledgment
40
The authors acknowledge the support of the NSF (CTS0132694) on this project.
Sorption-900°C Desorption-900°C
20
0
0
1000
2000
3000
2217
4000
5000
6000
Time(s)
Fig. 11. Adsorption and desorption kinetics among Samples 1–3.
than the single-phase Sr 4 Fe4 Co2 O13+ membranes. This is due to the presence of the perovskite phase permeating oxygen two orders of magnitude higher than the intergrowth phase. The same reasoning applies here. In Sample 1, it is the perovskite phase that is dominant in the adsorption process. However, in Samples 2 and 3, it is the intergrowth phase that dominates in the sorption process. Fig. 11 shows the adsorption (a) and desorption (b) kinetics of the three samples at 900 ◦ C. As shown from Fig. 11(a), the adsorption rate for the samples at 900 ◦ C decreases in the order: Sample 1 > Sample 2 > Sample 3. The same trend of adsorption kinetics was observed for these samples at 800 ◦ C. However, the desorption rate for the samples at 900 ◦ C decreases in the reverse order: Sample 3 > Sample 2 > Sample 1. 4. Conclusions SrCoFeOx is reported as an alternate sorbent to LSCF1991 for high temperature air separation at temperatures greater than 800 ◦ C. X-ray diffraction results reveal that SCF111 is a composition that contains an intergrowth, perovskite, and spinel phase and it has strong structure stability in a reducing atmosphere at high temperatures. Compared with LSCF1991, SCF111 exhibits a higher oxygen sorption capacity, faster oxygen desorption rate, and slower oxygen sorption rate at temperatures above 800 ◦ C. Both oxygen vacancies (VO¨ ) and interstitial oxygen ions (Oi ) are involved in the oxygen adsorption and desorption process for SrCoFeOx . Samples of SCF111 sintered at 1140 ◦ C had a higher percentage of the intergrowth phase than samples sintered at 950 ◦ C due to the slower formation of the intergrowth phase at lower temperatures. The synthesis method also has a strong influence on the amount of the three phases as samples prepared by the citrate method had a larger amount of the perovskite phase than samples prepared by the solid state method. The perovskite phase plays an important role in the oxygen sorption property. With a higher concentration of the perovskite phase in the mixture, the corresponding material shows a higher oxygen sorption capacity, slower
References Armstrong, T., Prado, F., Xia, Y., Manthiram, A., 2000. Role of perovskite phase on the oxygen permeation properties of the Sr 4 Fe6−x Cox O13+ system. Journal of Electrochemical Society 147 (2), 435–438. Fjellvag, H., Hauback, B.C., Bredesen, R., 1997. Crystal structure of the mixed conductor Sr 4 Fe4 Co2 O13 . Journal of Material Chemistry 7 (12), 2415–2419. Guggilla, S., Manthiram, A., 1997. Crystal chemical characterization of the mixed conductor Sr(Fe, Co)1.5 Oy exhibiting unusually high oxygen permeability. Journal of Electrochemical Society 144 (5), L120–L122. Klara, S.M., Srivastava, R.D., 2002. US DOE integrated collaborative technology development program for CO2 separation and capture. Environmental Progress 21 (4), 247–253. Lin, Y.S., Mclean, D.L., Zeng, Y., 2000. High temperature adsorption process, US patent, 6,059,858. Ma, B., Balachandran, U., 1997. Oxygen nonstoichiometry in mixedconducting SrFeCo0.5 Ox . Solid State Ionics 100, 53–62. Ma, B., Balachandran, U., 1998a. Phase stability of SrFeCo0.5 Ox in reducing environments. Materials Research Bulletin 33 (2), 223–236. Ma, B., Balachandran, U., 1998b. Transport properties of the mixedconducting SrFe1.2 Co0.3 Ox . Journal of Electroceramics 2 (2), 135–142. Ma, B., Balachandran, U., Park, J.H., 1996a. Electrical transport properties and defect structure of SrFeCo0.5 Ox . Journal of Electrochemical Society 143 (5), 1736–1744. Ma, B., Balachandran, U., Park, J.H., Segre, C.U., 1996b. Determination of chemical diffusion coefficient of SrFeCo0.5 Ox by the conductivity relaxation method. Solid State Ionics 83, 65–71. Ma, B., Hodges, J.P., Jorgensen, J.D., Miller, D.J., Richardson, J.W., Balachandran, U., 1998. Structure and properties relationship in mixedconducting Sr 4 (Fe1−x Cox )O13± materials. Journal of Solid State Chemistry 141, 576–586. Mizusaki, J., Yamauchi, S., Fueki, K., Ishikawa, A., 1984. Nonstoichiometry of perovskite-type oxides La1−x Sr x CoO3− . Solid State Ionics 12, 119–124. Mizusaki, J., Mima, Y., Yamauchi, S., Fueki, K., Tagawa, H., 1989. Nonstoichiometry of the perovskite-type oxides La1−x Sr x CoO3− . Journal of Solid State Chemistry 80, 102–111. Rodewald, A., Kather, A.A., Frie, S., 2005. Thermodynamic and economic aspects of the hard coal based oxyfuel cycle. International Journal of Green Energy 2 (2), 181–192. Xia, Y., Armstrong, T., Prado, F., Manthiram, A., 2000. Sol–gel synthesis, phase relationships, and oxygen permeation properties of Sr 4 Fe6−x Cox O13+ (0 x 3). Solid State Ionics 130, 81–90. Yang, Z.H., Lin, Y.S., 2002. A semi-empirical equation for oxygen nonstoichiometry of perovskite-type ceramics. Solid State Ionics 150, 245–254. Yang, Z.H., Lin, Y.S., 2003a. High-temperature oxygen sorption in a fixed bed packed with perovskite-type ceramic sorbents. Industrial and Engineering Chemistry Research 42, 4376–4381. Yang, Z.H., Lin, Y.S., 2003b. Equilibrium of oxygen sorption on perovskitetype lanthanum cobaltite sorbent. A.I.Ch.E. Journal 49, 793–798. Yang, Z.H., Lin, Y.S., 2005. Synergetic thermal effects for oxygen sorption and order–disorder transition on perovskite-type oxides. Solid State Ionics 176, 89–96.
2218
Q. Yin et al. / Chemical Engineering Science 63 (2008) 2211 – 2218
Yang, Z.H., Lin, Y.S., Zeng, Y., 2002. High temperature sorption process for air separation and oxygen removal. Industrial and Engineering Chemistry Research 41, 2775–2784. Yin, Q.H., Yang, Z.H., Lin, Y.S., 2006. Effects of microstructure on oxygen transport in perovskite-type oxides. Journal of Material Science 41, 4865–4870.
Yin, Q.H., Lin, Y.S., 2006. Effect of dopant addition on oxygen sorption properties of La–Sr–Co–Fe–O perovskite type oxide. Adsorption 12, 329–338. Yin, Q.H., Lin, Y.S., 2007. Beneficial effect of order–disorder phase transition on oxygen sorption properties of perovskite-type oxides. Solid State Ionics 178, 83–89.